AIDV Index
Roberto L López-Dávila2021, Artificial Intelligence and Democratic Values Index
visibility
…
description
624 pages
link
1 file
raft and research contributor to this report which is the premier publication of the Center for AI and Digital Policy. It assess the AI policies and practices of countries around the world. In this 2021 edition, the report expand the rated countries from 30to 50, based on a rigorous methodology and 12 metrics established to assess alignment with democratic values. The report update previous assessments to include new legal instruments, as the UNESCO Recommendation, against which evaluate countries’ efforts to develop and use AI in accordance with international human rights standards. The 2021 report is the result of the work of more than 100 AI policy experts in almost 40 countries.
This report is available online at:
https://caidp.org/reports/aidv-2021/
ISBN 979-8-9857883-0-3
© Center for AI and Digital Policy 2022
2
ARTIFICIAL
INTELLIGENCE
AND
DEMOCRATIC
VALUES
Center for AI and Digital Policy
CAIDP.ORG
.
Preface
The past year has produced rapid changes in the world of AI policies
and practices. National governments and international organizations are
moving quickly to create new frameworks in an attempt to maximize the
benefits and minimize the risks of Artificial Intelligence. At the same time,
private companies and government agencies are gathering vast repositories
of data, deploying new AI systems to manage every activity from
employment and transportation to education for children and care for the
elderly. At the outset, it is worth considering whether the gap between the
policies to govern the deployment of these new technologies and the actual
deployment is narrowing or growing wider, as this would be a critical
indicator of the respect for democratic values at the heart of a humancentered polity.
As we undertook the second survey of national AI policies and
practices we built on our earlier work, we identified new trends, and we
revisited the metrics we had established for trustworthy and human-centric
AI. Here are a few key findings from the past year:
The UNESCO Recommendation on the Ethics of AI, adopted by 193
countries in November 2021, was the single most significant AI policy
development of the past year. The UNESCO Recommendation speaks
directly to the widespread – and widely shared – aspiration of countries that
AI should benefit humanity. In a field that barely existed a few years ago,
the UNESCO AI Recommendation is a remarkably comprehensive AI
policy framework, touching upon established AI concerns, such as fairness,
accuracy, and transparency, and emerging AI issues, including gender
equity and sustainable development. UNESCO’s proposal for Ethical
Impact Assessment provides a powerful new tool to assess, in advance, the
consequences of the deployment of AI systems. Recognizing the
importance of the first global framework for AI ethics, we have this year
altered one of our metrics to take account of the significance of the
UNESCO Recommendation on AI. It is a development worth
acknowledging and celebrating. In future reports, we will likely add another
metric to assess the far more challenging issue of implementation.
Since publication of our last report, we also note the introduction of
the European Commission proposal for the regulation AI. The Commission
has set out a comprehensive, risk-based approach that could extend the
“Brussels Effect” to the global governance of AI. The European Parliament
has also signaled its intention to strengthen key provisions, and likely will
prohibit the use of AI techniques for remote biometric identification.
Meanwhile, the Council of the European Union, under the Presidency of
i
Slovenia and now France, have proposed additional texts that would, among
other changes, extend the prohibition on social scoring to private companies
as well as public agencies.
2021 also marked the adoption of Resolution 473 in Africa,
concerning the need to undertake a study on human and peoples’ rights and
artificial intelligence. The African Commission on Human and Peoples’
Rights called on State Parties “to ensure that the development and use of
AI, robotics and other new and emerging technologies is compatible with
the rights and duties in the African Charter and other regional and
international human rights instruments, in order to uphold human dignity,
privacy, equality, non-discrimination, inclusion, diversity, safety, fairness,
transparency, accountability and economic development as underlying
principles that guide the development and use of AI, robotics and other new
and emerging technologies.”1 It is a powerful statement.
China has also adopted sweeping new laws for both data protection
and the regulation of recommendation algorithms. Although the privacy
rules look very similar to the GDPR and the regulation for the governance
of recommendation algorithms share similar ambitions to proposals pending
in both the European Union and the US Congress, there are real concerns
about AI policies that are intended to favor a government in power. Against
the backdrop of democratic values, the goals of transparency and
accountability are offset by the inherent bias of such a legal structure.
We also noted this year the growing conflict over the deployment of
facial recognition for mass surveillance. While the European Parliament
voted to ban the use of AI technology for this purpose, many governments
and private companies pushed forward new systems for surveillance in
residential communities, inside school classrooms, and at public parks.
These are not the CCTV cameras of old, but sophisticated image processing
systems, designed specifically to identify individuals in public spaces by
name. In some countries, this system of unique identification is then tied to
elaborate government databases for scoring people based on their allegiance
to the government in power. It is a form of social control beyond the
imagination of even George Orwell.
We call attention also to the unfortunate failure of negotiators at the
UN conference in late December to make progress on a proposal to limit –
or better to prohibit – the use of lethal autonomous weapons. This occurred
in the same year that the United Nations was able to verify the use of
African Commission on Human and Peoples' Rights, 473 Resolution on the need to
undertake a Study on human and peoples’ rights and artificial intelligence (AI), robotics
and other new and emerging technologies in Africa - ACHPR/Res. 473 (EXT.OS/ XXXI)
2021 (Feb. 25, 2021),
1
ii
autonomous drone swarms to target and kill retreating military forces in the
civil war in Libya.
As the field of AI policy rapidly matures, we observe the growing
presence of judicial decisions, now shaping the laws of algorithms. In
several cases, including the secretive evaluation of employee performance,
courts have rejected opaque automated decisions. These judgements are
based on well-established legal frameworks, such as the GDPR, though we
see also legislative efforts to make automated decision-making with AI
techniques more accountable. We report these outcomes favorably as
algorithmic transparency remains one of our key metrics for the evaluation
of AI policies and practices.
In addressing the need to advance democratic values in the age of
AI, the ability of the European Union, the United States, and allies to work
in common purpose remains central. On that front, the past year provides
reason for both optimism and concern. The EU and the US launched a Trade
and Technology Council in 2021 that set out a common framework on AI
policy that could promote further transatlantic cooperation. The good news
is that “human rights” and “democratic values” undergird many of the
proposals. Top officials in the Biden Administration also expressed support
for the EU AI Act, a key legislative framework that will likely move
forward in 2022.
At the same time, the future of the EU AI Act is not certain, as some
politicians have made the mistake of assuming it is possible to trade the
protection of rights for innovation. Technologies that fail to protect rights
are not innovative, they are oppressive and stifling. On the US side, several
federal agencies have initiated AI-related “listening sessions,” but the
necessary work of establishing legal standards to protect democratic values
has yet to begin.
Still, our survey of national AI policies and practices also revealed
the hard work of many NGOs, advocates, academics, and government
officials, around the world, who have fully engaged the challenges that AI
poses and are prepared to stand on the front lines in defense of fundamental
rights. The remarkable progress made by the ReclaimYourFace campaign
in Europe, and similar campaigns in Africa, Asia, and Latin America speak
to a rapidly growing public recognition that not all technologically
transformative impacts should be welcome. There is a growing
understanding that “red lines” are necessary to safeguard fundamental
rights. And in that recognition may be found also the key to aligning AI
policies and practices, to narrowing the gap between the world of AI as it is
and the world of AI we wish to inhabit. If AI is to remain human-centric,
then we must determine the appropriate applications of AI.
iii
We cannot say enough to thank the extraordinary CAIDP team
members – the Global Academic Network, the Research Group, the
Regional Coordinators, the CAIDP Fellows, the law school externs, and
friends – who made possible this report. From an early project with a
handful of people, the current report reflects the efforts and dedication of
more than 100 experts in almost 40 countries. They did the hard work of
researching and writing, presenting for discussion, and then preparing the
assessments and ratings that make possible comparative analysis. We are
also grateful to the advisors and reviewers who provided comments on
earlier drafts of the report and direction for future editions. Our board
members and dedicated volunteers have kept us on course during this period
of remarkable growth. And we thank the supporters of CAIDP who have
helped establish one of the most influential organizations in the field of AI
policy. Together we share a commitment to a better society, more fair, more
just — a world where technology promotes broad social inclusion based on
fundamental rights, democratic institutions, and the rule of law.
To those in the AI policy field, whether advisors, decisionmakers,
heads of government, independent experts, or simply members of the public
who are interested in the growing impact of artificial intelligence on our
lives and our societies, we hope you will give this report your attention. The
rate of change is accelerating. We must act before it is too late.
As always, we welcome your advice, suggestions, revisions, and
updates. Please send editorial comments to [email protected] and visit our
website caidp.org to find more about our activities, recent policy
developments and how to get involved.
Marc Rotenberg
President and Founder
Merve Hickok
Research Director
Karine Caunes
Global Program Director
Center for AI Digital Policy
CAIDP.ORG
February 2022
iv
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Acknowledgements
Editorial Team: Marc Rotenberg, Merve Hickok, Karine Caunes
Communications Team: Cecilia Garibotti, Rachel Stockton.
2021 Contributors: Adeboye Adegoke, Temofe Akaba, Selim Alan, Zelal
Binici, Afi Blackshear, Bridget Boakye, Giuliano Borter, Stephanie Cairns,
Shangamitra Chakraborty, Lyantoniette Chua, Roberto López Dávila, Pam
Dixon, Divya Dwivedi, Elisa Elhadj, Pete Furlong, August Gweon, Lida
Haghnegahdar, Gokce Cobansoy Hizel, Jason K. Johnson, Eddan Katz,
Dongwoo Kim, Writankar Kundu, Thobekile Matimbe, Natalia Menendez,
Kushang Mishra, Charles Kajoloweka, Mélissa M’Raidi-Kechichian,
Marcos López Oneto, Alex Moltzau, Tamra Moore, Oarabile Mudongo,
Somaieh Nikpoor, Diana Kemunto Nyakundi, Ubongabasi Edidiong Obot,
Marcos López Oneto, Ananya Ramani, Rishi Ray, Rachel Stockton,
Anirbar Sen, Niovi Vavoula, Sherry Wu, Khatia Zukhubaia, Larissa Zutter.
2020 Contributors: Mariame Almoghanam, Francesca Bignami, Anne
Carblanc, Karine Caunes, Emilio De Capitani, Pam Dixon, Paula Soumaya
Domit, Giovanni De Gregorio, Douglas Frantz, Merve Hickok, Regina
Iminova, Lorraine Kisselburgh, Rebecca Leeper, Wonki Min, Pablo
Molina, Maria Helen Murphy, Stephanie Perrin, Bilyana Petkova, Oreste
Pollicino, Nguyen Dinh Quy, Cristos Velasco, Wendell Wallach, and
Larissa Zutter.
Reviewers: Dr. Elena Abrusci, Brunel University London; Isabelle Buscke,
Brussels Office, Federation of German Consumer Organisations; Dr.
Giovanni De Gregorio, University of Oxford; Dr. Gry Hasselbalch,
DataEthics.eu, InTouchAI.eu; Dr. Masao Horibe, Professor Emeritus at
Hitotsubashi University; Former Chairman of Personal Information
Protection Commission(PPC), Japan; Leyla Keser, IT Law Research
Center, Istanbul Bilgi University (with Işıl Selen Denemeç, Zümrüt
Müftüoğlu, and Dr. Atilla Aydin); Prof. Dr. Katja Langenbucher, Goethe
University; Prof. Gianclaudio Malgieri, EDHEC Business School; Prof.
Wonki Min, Technology Ambassador, Government of Korea; Prof.
Valsamis Mitsilegas, Queen Mary University of London; Dr. Pablo G.
Molina, Georgetown University; Dr. Irena Nesterova, Researcher at the
Institute of Legal Science, Faculty of Law, University of Latvia, member of
the UNESCO Ad Hoc Expert Group for the Recommendation on the Ethics
of Artificial Intelligence; Prof. Oreste Pollicino, Bocconi University; Prof.
v
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Viviana Polisena, Universidad Católica de Córdoba; Prof. Dr. Ingrid
Schneider, Universität Hamburg, Department of Informatics; Prof.
Giovanni Sartor, University of Bologna, European University of Florence;
Prof. Ben Shneiderman, University of Maryland; Dr. Cristos Velasco.
Please note also that the 2022 members of the CAIDP Research Group are
preparing updates that will appear in the next edition of Artificial
Intelligence and Democratic Values. Their names and affiliations will be
found at the CAIDP website under “Research Group 2022.”
CAIDP Board of Directors
•
•
•
•
•
•
•
•
•
Merve Hickok, Chair
Marc Rotenberg, President
Karine Caunes, Secretary
Pablo Molina, Treasurer
Doaa Abu-Alyounes
Len Kennedy
Lorraine Kisselburgh
Cristos Velasco
Larissa Zutter
CAIDP Global Academic Network
•
•
•
•
•
•
•
•
•
•
•
•
•
Karine Caunes, Director, CAIDP Global Academic Network (France)
Elena Abrusci, University of Brunel (United Kingdom)
Professor Alessandro Acquisti, Carnegie Mellon University (United
States)
Professor Francesca Bignami, George Washington University (United
States)
Professor Meredith Broussard, New York University (United States)
Professor Joseph David, Sapir College (Israel)
Professor Laura K. Donohue, Georgetown Law (United States)
Hlengiwe Dube, Center for Human Rights (South Africa)
Giovanni De Gregorio, Oxford University (Italy)
Gry Hasselbalch, Data Ethics EU and InTouchAI EU, (Denmark)
Professor Masao Horibe, Hitotsubashi University and Chuo University
(Japan)
Professor Meltem Ineli-Ciger, Suleyman Demirel University (Turkey)
Malavika Jayaram, Berkman-Klein Center (Singapore)
vi
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Professor Leyla Keser, Istanbul Bilgi University (Turkey)
Lorraine Kisselburgh, Purdue University (United States)
Professor Eleni Kosta, Tilburg University (the Netherlands)
Professor Katja Langenbucher, Goethe-University (Germany)
Professor Gianclaudio Malgieri, EDHEC Business School (France)
Professor Wonki Min, President SUNY Korea (Korea)
Professor Valsamis Mitsilegas, Queen Mary University of London
(Great Britain)
Professor Maria Murphy, National University of Ireland, Maynooth
(Ireland)
Professor Irena Nesterova, University of Latvia (Latvia)
Professor Bilyana Petkova, University of Graz (Austria)
Professor Viviana Polisena, Universidad Católica de Córdoba
(Argentina)
Professor Oreste Pollicino, Bocconi University (Italy)
Professor Stuart Russell, University of California at Berkeley (United
States)
Professor Idoia Salazar, San Pablo CEU University (Spain)
Professor Edward Santow, University of Technology Sydney
(Australia)
Professor Giovanni Sartor, University of Bologna / European University
Institute (Italy)
Professor Ingrid Schneider, Universität Hamburg (Germany)
Professor Lee Tiedrich, Duke University (United States)
Marc Rotenberg, President CAIDP (United States)
CAIDP Fellows
•
•
•
•
•
•
•
•
•
•
•
Temofe Akaba (2021)
Selim Alan (2021)
Bridget Boakye (2021)
Giuliano Borter (2020)
Lyantoniette Chua (2021)
Stephanie Cairns (2021)
Pete Furlong (2021)
August Gweon (2021)
JJ Johnson (2021)
Regina Kronan (2020)
Rebecca Leeper (2020)
vii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
•
•
•
•
•
•
•
Roberto L. López-Dávila (2021)
Natalia Menéndez González (2021)
Tamra Moore (2021)
Oarabile Mudongo (2021)
Somaieh Nikpoor (2021)
Ubongabasi Obot (2021)
Paula Soumaya Domit (2020)
Khatia Zukhubaia (2021)
Larissa Zutter (2020)
viii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Endorsements
“A very worthwhile analysis.”
- Stuart Russel, University of Berkeley, California
“An impressive body of work.”
- Alessandro Acquisti, Carnegie Mellon University
“An excellent report, a very useful tool for the community and a reference
to see nexus between AI and Democratic Values”
- Leyla Keser, Istanbul Bilgi University
"A very insightful report on AI laws and policies that addresses significant
international developments and provides a terrific comparison of AI laws
and policies across 50 countries.”
- Lee J. Tiedrich, Duke University
“The CAIDP's Report on AI and Democratic Value is an incredibly
comprehensive and brave collection of policy and legal developments on
the field of AI in the world. The AIDV report can be a point of reference
not only for scholars and NGOs, but also for policymakers across the
globe. The fundamental rights challenge of AI is too big to be ignored: the
regulatory urgency should be complemented by far-seeing wisdom and the
AIDV report can be a real help in this direction.”
- Gianclaudio Malgieri, EDHEC Business School
“This report is unique in that it compares AI policies and practices around
the world and should inspire progress toward trustworthy and humancentric AI. The report also documents the important role of digital rights
and consumer groups in this endeavor.”
- Ursula Pachl, European Consumer Association (BEUC)
“It's rare to read a document that has the potential to influence national
policies on artificial intelligence around the world. The country
descriptions and evaluations are thoughtfully organized, well-written, and
carefully documented. The country rankings enable readers to have a
broad understanding of who the leaders are and why, while pointing out
what still needs to be done. The AI and Democratic Values index gives
me hope that AI policy efforts can improve human rights, social justice,
and dignity.”
- Ben Shneiderman, author, Human-Centric AI (Oxford 2022)
ix
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
“In this historical moment of global interest negotiation and competition
on AI we are in right now, we urgently need a shared narrative serving the
human interest and democracy. With its comprehensive overview of key
global and local power actors and initiatives The Artificial Intelligence
and Democratic Values Index is an essential instrument for the
facilitation of this shared global dialogue on AI now and in many years to
come.”
- Gry Hasselbalch, DataEthics.eu, InTouchAI.eu
““A comprehensive and detailed overview of the global governance of AI
- an essential body of work as AI emerges from the scientific frontier to
center stage in society.”
- Joichi Ito, Director, Center for Radical Transformation, Chiba
Institute of Technology
“This is a very ambitious and important exercise. The outcome is
impressive- detailed and rigorous.”
- Valsamis Mitsilegas, Queen Mary University of London
“A unique comparative analysis of AI policy across the world. The
Artificial Intelligence and Democratic Values Index does not just
provide a comprehensive perspective on the evolution of AI policy but
also contribute to underlining the constitutional relationship between AI
and democracy.”
- Oreste Pollicino, Bocconi University
- Giovanni De Gregorio, University of Oxford
“This comparative study sheds critical light on the adoption of Artificial
Intelligence in democratic societies. AI and Democratic Values is an
indispensable reference source for regulators, reporters, academics, and
practitioners
- Dr. Pablo G. Molina, Georgetown University and founder,
International Applied Ethics and Technology Association
(iaeta.org)
x
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
TABLE OF CONTENTS
Preface.......................................................................................................... i
Acknowledgements ................................................................................. v
Endorsements ......................................................................................... ix
EXECUTIVE SUMMARY ........................................................................ 1
Purpose and Scope .................................................................................. 1
The 2022 Edition .................................................................................... 2
Findings .................................................................................................. 3
Recommendations ................................................................................... 4
New Recommendations (2022) .............................................................. 4
THE GLOBAL AI POLICY LANDSCAPE .............................................. 5
The Council of Europe ............................................................................ 5
CAHAI ................................................................................................ 6
Committee of Ministers ...................................................................... 7
European Committee on Crime Problems .......................................... 8
Parliamentary Assembly ..................................................................... 9
European Court of Human Rights..................................................... 10
Commissioner for Human Rights ..................................................... 10
Commission for the Efficiency of Justice ......................................... 11
The European Union ............................................................................. 11
The European Commission ............................................................... 11
The European Parliament .................................................................. 15
Committees – AIDA, IMCO, LIBE .................................................. 18
The Two Councils ............................................................................. 20
The Court of Justice of the European Union .................................... 22
The European Data Protection Board ............................................... 22
The European Data Protection Supervisor ........................................ 23
Fundamental Rights Agency ............................................................. 25
High-Level Expert Group on AI ....................................................... 26
G7.......................................................................................................... 27
G20........................................................................................................ 28
Global Privacy Assembly ..................................................................... 30
The OECD ............................................................................................ 31
Global Partnership on AI .................................................................. 32
OECD AI Observatory ...................................................................... 32
National Implementation .................................................................. 33
The OECD ONE PAI ........................................................................ 34
xi
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
United Nations ...................................................................................... 35
The Secretary General....................................................................... 35
UNESCO Recommendation on AI Ethics ........................................ 37
UN High Commissioner for Human Rights...................................... 39
International Telecommunications Union......................................... 40
UN Special Rapporteur ..................................................................... 40
UN and Lethal Autonomous Weapons ............................................. 41
The Vatican ........................................................................................... 42
Technical Societies ............................................................................... 43
Civil Society ......................................................................................... 44
Latin America ................................................................................... 44
Africa ................................................................................................ 46
Asia ................................................................................................... 48
Europe ............................................................................................... 50
United States ..................................................................................... 52
COUNTRY REPORTS............................................................................. 54
Argentina .............................................................................................. 54
National AI Strategy ......................................................................... 54
Regional/Provincial........................................................................... 55
Public Participation ........................................................................... 56
Privacy .............................................................................................. 56
Algorithmic Transparency ................................................................ 57
Human Rights ................................................................................... 58
OECD/G20 AI Principles ................................................................. 58
Facial Recognition ............................................................................ 59
Lethal Autonomous Weapon Systems .............................................. 60
Evaluation ......................................................................................... 60
Australia ................................................................................................ 62
National AI Strategy ......................................................................... 62
Public Participation ........................................................................... 63
Data Protection.................................................................................. 65
Global Partnership on AI .................................................................. 66
Algorithmic Transparency ................................................................ 66
OECD/G20 AI Principles ................................................................. 67
Human Rights ................................................................................... 67
Evaluation ......................................................................................... 68
Austria ................................................................................................... 69
National AI Strategy ......................................................................... 69
Public Participation ........................................................................... 70
Facial Recognition ............................................................................ 71
xii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Predictive Policing ............................................................................ 71
AI Oversight...................................................................................... 72
Algorithmic Transparency ................................................................ 72
OECD AI Principles ......................................................................... 73
Human Rights ................................................................................... 73
Lethal Autonomous Weapons ........................................................... 74
Evaluation ......................................................................................... 75
Bangladesh ............................................................................................ 76
National AI Strategy ......................................................................... 76
Public Participation ........................................................................... 76
Data Privacy Law .............................................................................. 77
Biometric Recognition ...................................................................... 78
Algorithmic Transparency ................................................................ 78
Human Rights ................................................................................... 79
Lethal Autonomous Weapons ........................................................... 79
Evaluation ......................................................................................... 80
Belgium ................................................................................................. 81
National AI Strategy ......................................................................... 81
Regional/Community Strategies ....................................................... 83
Public Opinion .................................................................................. 84
Independent AI oversight .................................................................. 85
Public Participation ........................................................................... 85
Facial Recognition ............................................................................ 86
Algorithmic Transparency ................................................................ 86
Lethal Autonomous Weapons Systems ............................................ 87
OECD/G20 AI Principles ................................................................. 87
Human Rights ................................................................................... 88
Evaluation ......................................................................................... 88
Brazil ..................................................................................................... 89
National AI Strategy ......................................................................... 89
Public Participation ........................................................................... 91
Research & Development ................................................................. 92
Privacy .............................................................................................. 93
Data Protection Authority ................................................................. 94
Medical Data ..................................................................................... 95
Algorithmic Transparency ................................................................ 95
AI and the Judiciary .......................................................................... 96
Facial recognition.............................................................................. 98
OECD/G20 AI Principles ................................................................. 99
Human Rights ................................................................................... 99
Lethal Autonomous Weapons ......................................................... 100
xiii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Evaluation ....................................................................................... 100
Canada ................................................................................................ 101
National AI Strategy ....................................................................... 101
Directive on Automated Decision-making ..................................... 102
Predicting Homelessness ................................................................ 103
Public Participation ......................................................................... 103
Data Protection................................................................................ 104
Algorithmic Transparency .............................................................. 105
Facial Recognition .......................................................................... 106
Networking and collaborative platforms ......... Error! Bookmark not
defined.
Global Partnership on AI ................................................................ 107
OECD/G20 AI Principles ............................................................... 107
Lethal Autonomous Weapons ......................................................... 108
Human Rights ................................................................................. 108
Evaluation ....................................................................................... 108
China ................................................................................................... 109
National AI Strategy ....................................................................... 109
AI Core Values ............................................................................... 110
Facial Recognition .......................................................................... 113
Medical AI ...................................................................................... 114
Use of AI in Covid-19 Response .................................................... 114
AI Ethics ......................................................................................... 115
AI and Surveillance......................................................................... 116
Public Opinion ................................................................................ 119
Data Protection................................................................................ 119
Fundamental Rights & OECD AI Principles .................................. 120
Evaluation ....................................................................................... 121
Colombia ............................................................................................. 123
National AI Strategy ....................................................................... 123
Ethical Framework for Artificial Intelligence................................. 125
Regulatory Sandboxes and Beaches ............................................... 126
Public Participation ......................................................................... 127
Research & Development ............................................................... 128
Privacy and Data Protection............................................................ 128
Data Infrastructure .......................................................................... 130
AI and the Judiciary ........................................................................ 132
Facial Recognition .......................................................................... 133
OECD/G20 AI Principles ............................................................... 135
Human Rights ................................................................................. 135
Lethal Autonomous Weapons ......................................................... 135
xiv
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Evaluation ....................................................................................... 136
Denmark.............................................................................................. 137
National AI Strategy ....................................................................... 137
The OECD AI Principles ................................................................ 138
Universal Declaration of Human Rights ......................................... 138
Public Participation ......................................................................... 140
AI Policies and Practices in the Public Sector ................................ 141
Oversight Mechanisms.................................................................... 142
Universal Guidelines for AI ............................................................ 142
Independent Council for Ethical use of Data .................................. 143
Evaluation ....................................................................................... 143
Dominican Republic ........................................................................... 144
National AI Strategy ....................................................................... 144
Privacy and Data Protection............................................................ 145
Human Rights ................................................................................. 145
Facial Recognition .......................................................................... 146
Lethal Autonomous Weapons ......................................................... 146
Evaluation ....................................................................................... 146
Egypt ................................................................................................... 147
National AI Strategy ....................................................................... 147
Public Participation ......................................................................... 148
Regional Leadership ....................................................................... 149
Responsible AI ................................................................................ 150
OECD/G20 AI Principles ............................................................... 150
Human Rights ................................................................................. 151
Biometric Recognition .................................................................... 151
Evaluation ....................................................................................... 152
Estonia ................................................................................................ 153
National AI Strategy ....................................................................... 154
OECD AI Principles ....................................................................... 156
Human Rights ................................................................................. 156
Algorithmic Transparency .............................................................. 157
Public Participation ......................................................................... 158
Evaluation ....................................................................................... 159
Finland ................................................................................................ 161
National Approach to Artificial Intelligence .................................. 161
Access to Data................................................................................. 164
Foreign Policy and AI ..................................................................... 164
Public Participation ......................................................................... 166
Facial Recognition .......................................................................... 166
Lethal Autonomous Weapons ......................................................... 166
xv
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD AI Principles ....................................................................... 167
Algorithmic Transparency .............................................................. 167
Human Rights ................................................................................. 168
Evaluation ....................................................................................... 169
France.................................................................................................. 170
National AI Strategy ....................................................................... 170
The Health Data Hub Controversy ................................................. 171
Launch of National AI Research Institutes ..................................... 173
AI Cloud.......................................................................................... 173
National Pilot Committee for Digital Ethics ................................... 174
Fundamental Rights ........................................................................ 175
Facial Recognition .......................................................................... 176
Consumer Perspective ..................................................................... 178
The Global Partnership on AI ......................................................... 178
Algorithmic Transparency .............................................................. 178
OECD/G20 AI Principles ............................................................... 179
Human Rights ................................................................................. 180
Lethal Autonomous Weapons ......................................................... 180
Evaluation ....................................................................................... 180
Germany.............................................................................................. 182
National AI Strategy ....................................................................... 182
Public Participation ......................................................................... 184
Data Ethics Commission ................................................................. 185
Facial Recognition .......................................................................... 186
Predictive Policing .......................................................................... 187
AI Oversight.................................................................................... 187
Algorithmic Transparency .............................................................. 189
OECD/G20 Principles and Global Partnership on AI..................... 190
Human Rights ................................................................................. 190
Lethal Autonomous Weapons ......................................................... 191
Evaluation ....................................................................................... 191
Hong Kong .......................................................................................... 193
National AI Strategy ....................................................................... 193
Fundamental Rights and OECD AI Principles ............................... 194
Public Participation ......................................................................... 195
Data Protection and Algorithmic Transparency ............................. 195
AI and Surveillance......................................................................... 197
Evaluation ....................................................................................... 198
India .................................................................................................... 200
National AI Strategies ..................................................................... 200
Quad Group ..................................................................................... 203
xvi
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI Policy Development and Oversight ........................................... 203
India’s AI Stack and Aadhaar ......................................................... 204
Public participation ......................................................................... 206
OECD/G20 AI Principles ............................................................... 207
Data Protection................................................................................ 207
Algorithmic Transparency .............................................................. 208
Human Rights ................................................................................. 209
Evaluation ....................................................................................... 209
Indonesia ............................................................................................. 210
National AI Strategy ....................................................................... 210
AI Initiatives ................................................................................... 211
Jakarta Smart City Initiative ........................................................... 212
AI Summit 2020 .............................................................................. 212
Pancasila Values ............................................................................. 212
AI Oversight.................................................................................... 213
Public Participation ......................................................................... 213
Data Privacy Law ............................................................................ 214
OECD/G20 AI Principles ............................................................... 215
Human Rights ................................................................................. 216
Evaluation ....................................................................................... 216
Iran ...................................................................................................... 217
National AI Strategy ....................................................................... 217
Freedom of Internet and Digital Rights .......................................... 218
Data Protection................................................................................ 218
Digital ID ........................................................................................ 219
Facial Recognition .......................................................................... 220
OECD/G20 Principles ..................................................................... 220
UNESCO AI Ethics Recommendation ........................................... 221
Human Rights ................................................................................. 221
Lethal Autonomous Weapons ......................................................... 221
Evaluation ....................................................................................... 222
Ireland ................................................................................................. 223
National AI Strategy ....................................................................... 223
Public Participation ......................................................................... 224
Public Trust in AI............................................................................ 225
Human Rights ................................................................................. 225
G20/OECD Principles ..................................................................... 226
Global Partnership on AI ................................................................ 226
Global Privacy Assembly ............................................................... 226
Algorithmic Transparency .............................................................. 226
AI Oversight.................................................................................... 227
xvii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Lethal Autonomous Weapons ......................................................... 228
Public Services Card Facial Recognition Controversy ................... 228
Evaluation ....................................................................................... 229
Israel.................................................................................................... 230
National AI Strategy ....................................................................... 230
National AI Program ....................................................................... 231
Ethical and Legal Aspect of AI ....................................................... 231
International Cooperation ............................................................... 233
Public Participation ......................................................................... 234
AI Week .......................................................................................... 234
Privacy and Data Protection............................................................ 234
Algorithmic Transparency .............................................................. 235
Medical Data ................................................................................... 236
Covid-19 Tracking Controversy ..................................................... 236
Social Ranking ................................................................................ 238
Facial recognition............................................................................ 238
OECD AI Principles ....................................................................... 239
Human Rights ................................................................................. 240
Autonomous Weapons .................................................................... 240
Evaluation ....................................................................................... 241
Italy ..................................................................................................... 242
National AI Strategy ....................................................................... 242
Strategic Programme on AI 2022-2024 .......................................... 244
National AI Ecosystem ................................................................... 245
Human capital ................................................................................. 246
The Italian Institute for Artificial Intelligence (I3A) ...................... 247
White Paper on Public Administration ........................................... 247
The Rome Call for Ethics ............................................................... 249
Public Participation and Access to Documents .............................. 249
Facial Recognition .......................................................................... 251
OECD/G20 AI Principles ............................................................... 252
Data Protection................................................................................ 253
Algorithmic Transparency .............................................................. 254
Human Rights ................................................................................. 254
Evaluation ....................................................................................... 254
Japan ................................................................................................... 255
National AI Strategy ....................................................................... 255
AI R&D Guidelines ........................................................................ 256
Social Principles of Human-Centric AI .......................................... 256
AI R&D Guidelines and AI Utilization Guidelines ........................ 257
Data Free Flows with Trust............................................................. 258
xviii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation ......................................................................... 259
Data Protection................................................................................ 260
OECD/G20 AI Principles ............................................................... 261
Algorithmic Transparency .............................................................. 261
Use of AI for policy decisions ........................................................ 261
Facial Recognition .......................................................................... 262
Human Rights ................................................................................. 263
Evaluation ....................................................................................... 263
Kazakhstan .......................................................................................... 264
Overview and National AI Strategy ................................................ 264
AI Core Values ............................................................................... 265
Facial Recognition and Smart Cities............................................... 266
Medical AI ...................................................................................... 267
AI Ethics ......................................................................................... 267
Data Protection................................................................................ 268
Algorithmic Transparency .............................................................. 269
OECD AI Principles ....................................................................... 269
Human Rights ................................................................................. 269
Evaluation ....................................................................................... 270
Kenya .................................................................................................. 271
National AI Strategy ....................................................................... 271
Public Participation ......................................................................... 273
International Partnership on AI ....................................................... 274
Data Protection................................................................................ 274
Human Rights ................................................................................. 275
Evaluation ....................................................................................... 275
Korea ................................................................................................... 277
National AI Strategy ....................................................................... 277
AI R&D Strategy ............................................................................ 278
Amendments to the Three Major Data Privacy Laws ..................... 279
Personal Information Protection Commission (PIPC) .................... 279
Global Partnership – OECD, G20, GPAI, and UNESCO............... 281
AI Ethics ......................................................................................... 282
Algorithmic Transparency .............................................................. 283
Improvement of Policies and Laws for the Era of AI ..................... 284
Human Rights Advocacy ................................................................ 284
Evaluation ....................................................................................... 285
Malaysia (new) ................................................................................... 286
National AI Strategy Overview ...................................................... 286
Background and Related National Policies..................................... 286
AI-Rmap ......................................................................................... 287
xix
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD/G20 AI Principles ............................................................... 288
Human Rights ................................................................................. 288
AI in healthcare ............................................................................... 288
AI System for Surveillance ............................................................. 290
Data Protection Laws ...................................................................... 290
Autonomous vehicles ...................................................................... 291
Evaluation ....................................................................................... 292
National AI Strategy ....................................................................... 293
Implementation of the National AI Strategy ................................... 293
Public Participation ......................................................................... 294
Global Partnerships: OECD, G20, GPAI, and COMEST ............... 295
Data Protection................................................................................ 296
Algorithmic Transparency .............................................................. 297
Human Rights ................................................................................. 297
Evaluation ....................................................................................... 298
Netherlands ......................................................................................... 299
Strategic Action Plan for AI ........................................................... 299
Predictive Policing .......................................................................... 301
Automated Profiling Fiasco ............................................................ 302
AI Registry ...................................................................................... 303
Public Participation ......................................................................... 303
Fundamental Rights and OECD AI Principles ............................... 304
Algorithmic Transparency .............................................................. 305
Evaluation ....................................................................................... 307
Nigeria ................................................................................................ 308
National AI Strategy ....................................................................... 308
Public Participation ......................................................................... 309
Research and Development............................................................. 310
Privacy ............................................................................................ 310
Digital Rights in Nigeria ................................................................. 311
COVID-19 Pandemic ...................................................................... 312
Facial Recognition .......................................................................... 313
Selected AI initiatives in Nigeria .................................................... 313
Evaluation ....................................................................................... 314
Norway................................................................................................ 316
National AI Strategy ....................................................................... 316
Public Participation (A) .................................................................. 317
Global Partnerships ........................................................................ 318
OECD AI Principles ....................................................................... 319
Human Rights.................................................................................. 319
Oversight: Data Protection Authority ............................................ 320
xx
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection Laws and Policies ................................................ 323
Evaluation ....................................................................................... 323
Philippines .......................................................................................... 325
National AI Strategy ....................................................................... 325
Public Participation ......................................................................... 326
OECD/G20 Principles ..................................................................... 326
Universal Declaration of Human Rights ......................................... 327
Data Privacy Law ............................................................................ 328
Algorithmic Transparency .............................................................. 329
Facial/Biometric Recognition ......................................................... 329
Evaluation ....................................................................................... 330
Poland ................................................................................................. 331
National AI Strategy ....................................................................... 331
Ultima Ratio .................................................................................... 335
Poland’s Position on AI and Fundamental Rights .......................... 335
OECD/G20 AI Principles ............................................................... 336
Human Rights ................................................................................. 337
Algorithmic Transparency .............................................................. 337
Data Protection................................................................................ 339
Public Participation ......................................................................... 340
Evaluation ....................................................................................... 340
Russia .................................................................................................. 341
National AI Strategy ....................................................................... 341
Digital Economy in Russia ............................................................. 342
AI Strategy for Russian start-up ..................................................... 342
AI Policy in Russia ......................................................................... 343
Digital Rights Law and AI Regulation ........................................... 344
Facial Recognition Controversy ..................................................... 344
Data Protection................................................................................ 345
Algorithmic Transparency .............................................................. 346
OECD/G20 AI Principles ............................................................... 346
Human Rights ................................................................................. 346
Evaluation ....................................................................................... 347
Rwanda ............................................................................................... 348
National AI Strategy ....................................................................... 348
AI System for Identity Management............................................... 349
Smart Cities ..................................................................................... 350
Drone Regulation ............................................................................ 351
Public Participation ......................................................................... 352
OECD AI Principles ....................................................................... 352
Fundamental Rights ........................................................................ 352
xxi
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection................................................................................ 353
Algorithmic Transparency .............................................................. 354
Evaluation ....................................................................................... 355
Saudi Arabia ....................................................................................... 356
National AI Strategy ....................................................................... 356
Global AI Summit ........................................................................... 359
Neom ............................................................................................... 360
Public Participation ......................................................................... 360
G-20 Meetings ................................................................................ 361
AI Oversight.................................................................................... 363
Data Protection................................................................................ 363
Data Governance ............................................................................. 364
OECD/G20 AI Principles ............................................................... 365
Human Rights ................................................................................. 365
Evaluation ....................................................................................... 366
Singapore ............................................................................................ 367
National AI Strategy ....................................................................... 367
AI System for Online Surveillance ................................................. 368
Smart Cities ..................................................................................... 370
Digital Identification and Surveillance ........................................... 370
Public Participation ......................................................................... 371
OECD/G20 AI Principles ............................................................... 372
Data Protection and Algorithmic Transparency ............................. 372
Human Rights ................................................................................. 374
Evaluation ....................................................................................... 374
Slovenia .............................................................................................. 376
National AI Strategy ....................................................................... 376
European Union .............................................................................. 377
European Council Presidency ......................................................... 378
International Research Center on Artificial Intelligence (IRCAI) .. 378
Privacy and Data Protection............................................................ 379
Algorithmic Transparency .............................................................. 379
OECD/G20 AI Principles ............................................................... 380
Human Rights ................................................................................. 380
Evaluation ....................................................................................... 381
South Africa ........................................................................................ 382
National AI Strategy ....................................................................... 382
Strides in Innovation ....................................................................... 383
Data Protection................................................................................ 384
COVID-19 Surveillance.................................................................. 385
Research and Development on AI .................................................. 385
xxii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation ......................................................................... 387
Public Opinion ................................................................................ 388
OECD/G20 Principles ..................................................................... 388
Human Rights ................................................................................. 389
Evaluation ....................................................................................... 389
Spain ................................................................................................... 391
National AI Strategy ....................................................................... 391
Artificial Intelligence Advisory Council ........................................ 394
Artificial Intelligence Supervision Agency .................................... 394
Charter on Digital Rights ................................................................ 395
Algorithmic Transparency .............................................................. 396
Lethal Autonomous Weapons ......................................................... 397
OECD/G20 AI Principles ............................................................... 397
Human Rights ................................................................................. 397
Evaluation ....................................................................................... 398
Sweden ................................................................................................ 399
National AI Strategy ....................................................................... 399
The Trellborg Controversy ............................................................. 403
Independent Oversight .................................................................... 404
Foreign Policy and AI ..................................................................... 406
Public Participation ......................................................................... 406
Facial recognition............................................................................ 407
Lethal Autonomous Weapons ......................................................... 409
OECD AI Principles ....................................................................... 410
Algorithmic Transparency .............................................................. 410
Human Rights ................................................................................. 411
Evaluation ....................................................................................... 412
Switzerland ......................................................................................... 413
National AI Strategy ....................................................................... 413
Swiss Foreign Policy and AI........................................................... 416
Independent AI oversight ................................................................ 418
Public Participation ......................................................................... 419
AI Events in Geneva ....................................................................... 420
AI and Criminal Justice .................................................................. 420
Algorithmic Transparency .............................................................. 421
OECD AI Principles ....................................................................... 422
Human Rights ................................................................................. 422
Evaluation ....................................................................................... 422
Taiwan ................................................................................................ 423
AI Action Plan ................................................................................ 423
AI Core Values ............................................................................... 424
xxiii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI R&D Guidelines ........................................................................ 425
Medical Data and AI Ethics ............................................................ 426
Covid-19 and Big Data Analytics ................................................... 427
Autonomous vehicles ...................................................................... 428
Facial Recognition .......................................................................... 429
Algorithmic Transparency .............................................................. 429
Use of AI for digital democracy ..................................................... 430
Data Protection................................................................................ 431
OECD AI Principles ....................................................................... 431
Fundamental Rights ........................................................................ 432
Evaluation ....................................................................................... 432
Thailand .............................................................................................. 433
National AI Strategy ....................................................................... 433
AI System for Surveillance ............................................................. 435
Anti-fake News Centre ................................................................... 435
Digital ID ........................................................................................ 436
Public Participation ......................................................................... 436
Fundamental Rights and OECD/G20 AI Principles ....................... 436
Data Protection................................................................................ 438
Lethal Autonomous Weapons ......................................................... 439
Evaluation ....................................................................................... 439
Turkey ................................................................................................. 440
National AI Strategy ....................................................................... 440
Public Participation ......................................................................... 441
Open Data Project ........................................................................... 443
Data Protection................................................................................ 443
OECD AI Principles ....................................................................... 445
Universal Guidelines for AI ............................................................ 445
Human Rights ................................................................................. 446
Algorithmic Transparency .............................................................. 446
Lethal Autonomous Weapons ......................................................... 447
United Arab Emirate ........................................................................... 449
National AI Strategy ....................................................................... 449
OECD/G20 Principles ..................................................................... 450
Public Participation ......................................................................... 451
Data Protection................................................................................ 452
Algorithmic Transparency .............................................................. 454
Human Rights ................................................................................. 454
Facial Recognition .......................................................................... 455
Predictive Policing .......................................................................... 455
Autonomous Weapons .................................................................... 456
xxiv
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Evaluation ....................................................................................... 456
United Kingdom ................................................................................. 458
National AI Strategy ....................................................................... 458
Public Participation ......................................................................... 460
Facial Recognition .......................................................................... 461
AI Grading Controversy ................................................................. 461
NGO Perspectives on AI in the UK ................................................ 462
Global Partnership on AI and OECD AI Principles ....................... 463
Data Protection................................................................................ 463
Algorithmic Transparency .............................................................. 464
Human Rights ................................................................................. 465
Evaluation ....................................................................................... 465
United States ....................................................................................... 467
National AI Strategy ....................................................................... 467
OMB AI Guidance for Agencies .................................................... 470
EU-U.S. Trade and Technology Council (TTC) ............................. 471
Facial Recognition .......................................................................... 472
National Security Commission on AI ............................................. 474
NIST Risk Management Framework .............................................. 475
JAIC ................................................................................................ 476
Algorithmic Transparency .............................................................. 476
OECD AI Principles ....................................................................... 478
Public Participation and Access to Documents .............................. 478
Human Rights ................................................................................. 479
Evaluation ....................................................................................... 481
Uruguay .............................................................................................. 482
National AI Strategy ....................................................................... 482
Global Partnership on AI ................................................................ 484
Global Privacy Assembly ............................................................... 484
Public Participation ......................................................................... 485
Data protection ................................................................................ 485
Data Protection Authority ............................................................... 486
Facial recognition............................................................................ 486
Human Rights ................................................................................. 487
Algorithmic Transparency .............................................................. 488
OECD/G20 AI Principles ............................................................... 488
Evaluation ....................................................................................... 488
COUNTRY EVALUATIONS ................................................................ 490
Evaluation Grid ................................................................................... 490
Tier I................................................................................................ 492
Tier II .............................................................................................. 492
xxv
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Tier III ............................................................................................. 492
Tier IV ............................................................................................. 493
Tier V .............................................................................................. 493
Country Distribution by Tier .............................................................. 494
Chart: AIDV Index by Country and Tier ............................................ 495
Chart: AIDV Index by Country, 2020 vs. 2021 .............................. 496
Chart: AIDV Country Scores by Individual Metrics ...................... 497
Metrics ................................................................................................ 498
Response Codes .................................................................................. 498
METHODOLOGY ................................................................................. 499
Scope ................................................................................................... 499
Time Period......................................................................................... 499
Factors ................................................................................................. 499
The Metrics ......................................................................................... 500
Q1. Has the country endorsed the OECD/G20 AI Principles? ....... 500
Q2. Is the country implementing the OECD AI Principles? ........... 501
Q3. Has the country endorsed the Universal Declaration of Human
Rights? ............................................................................................ 501
Q4. Is the country implementing the Universal Declaration for
Human Rights?................................................................................ 501
Q5. Has the country established a process for meaningful public
participation in the development of a national AI Policy? ............. 502
Q6. Are materials about the country’s AI policies and practices
readily available to the public? ....................................................... 502
Q7. Does the country have an independent (agency/mechanism) for
AI oversight? ................................................................................... 503
Q8. Do the following goals appear in the national AI policy:
“Fairness,” “Accountability,” “Transparency,” “Rule of Law,”
“Fundamental Rights”? ................................................................... 503
Q9. Has the country by law established a right to Algorithmic
Transparency? ................................................................................. 503
Q10. Has the country supported the Universal Guidelines for AI? 504
Q11. Has the country endorsed the UNESCO Recommendation on AI
Ethics? ............................................................................................. 504
Q12: Has the country’s Data Protection Agency endorsed the 2018
GPA Resolution on AI and Ethics and the 2020 GPA Resolution on
AI and Accountability? ................................................................... 505
Hong Kong ...................................................................................... 505
Exemplars ........................................................................................... 506
Scoring ................................................................................................ 506
xxvi
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Search Strategy ................................................................................... 506
Descriptive Summary ......................................................................... 506
Language ............................................................................................. 507
Citation Format ................................................................................... 507
Gender Balance and Diversity ............................................................ 507
Bias ..................................................................................................... 507
Private Sector Practices ...................................................................... 508
REFERENCES ....................................................................................... 509
AI Policy Surveys - International and Regional ............................. 509
AI Policy Surveys - Domestic......................................................... 510
Related Country Surveys ................................................................ 512
GLOSSARY ........................................................................................... 512
REFERENCE DOCUMENTS ................................................................ 516
OECD AI Principles ........................................................................... 516
OECD AI Policy Adherents ............................................................ 522
OECD Member Countries............................................................... 522
OECD Non-Member Countries ...................................................... 523
G-20 Countries ................................................................................ 523
Universal Guidelines for AI ................................................................ 524
UGAI Explanatory Memorandum .................................................. 526
UGAI References ............................................................................ 529
Universal Declaration of Human Rights ............................................. 531
GPA Declaration on Ethics and Data Protection in AI ....................... 537
GPA Resolution on AI and Accountability ........................................ 544
UNESCO Recommendation on AI Ethics .......................................... 548
Preamble ......................................................................................... 548
I. Scope of Application ................................................................... 552
II. Aims and Objectives .................................................................. 555
III. Values and Principles ................................................................ 556
IV. Areas for Policy Action ............................................................ 566
Policy Area 1: Ethical Impact Assessment ..................................... 567
Policy Area 2: Ethical Governance and Stewardship ..................... 568
Policy Area 3: Data Policy .............................................................. 572
Policy Area 4: Development and International Cooperation .......... 574
Policy Area 5: Environment and Ecosystems ................................. 575
Policy Area 6: Gender ..................................................................... 576
Policy Area 7: Culture .................................................................... 577
Policy Area 8: Education and Research .......................................... 579
Policy Area 9: Communication and Information ............................ 581
Policy Area 10: Economy and Labour ............................................ 582
xxvii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Policy Area 11: Health and Social Wellbeing ................................ 583
V. Monitoring and Evaluation ........................................................ 585
VI. Utilization and Exploitation of the Present Recommendation . 587
VII. Promotion of the Present Recommendation ............................ 587
VIII. Final Provisions...................................................................... 588
The Center for AI and Digital Policy ...................................................... 589
About this Report .................................................................................... 589
xxviii
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
xxix
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
EXECUTIVE SUMMARY
Purpose and Scope
Artificial Intelligence and Democratic Values is the first global
survey to assess progress toward trustworthy AI, based on detailed narrative
reports, combined with a methodology that produces ratings and rankings
for national AI policies and practices.
The AI Index has these objectives: (1) to document the AI policies
and practices, based on publicly available sources, (2) to establish a
methodology for the evaluation of AI policies and practices, based on global
norms, (3) to provide a basis for comparative evaluation, (4) to provide the
basis for future evaluations, and (5) to ultimately encourage all countries to
make real the promise of AI that is trustworthy, human-centric, and provides
broad social benefit to all.
Artificial Intelligence and Democratic Values focuses on human
rights, rule of law, and democratic governance metrics. Endorsement and
implementation of the OECD/G20 AI Principles is among the primary
metrics. Opportunities for the public to participate in the formation of
national AI policy, as well as the creation of independent agencies to
address AI challenges, is also among the metrics. Patents, publications,
investment, and employment impacts are important metrics for the AI
economy, but they are not considered here.
Artificial Intelligence and Democratic Values will be published on
an annual basis and will evolve as country practices change and new issues
emerge.
1
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The 2022 Edition
The 2022 edition of the report updates and expands the initial report.
Among the key changes:
• The number of countries assessed increased from 30 to 50
• Endorsement of the UNESCO Recommendation on AI Ethics is
now one of the key metrics to assess progress toward humancentric and trustworthy AI.
• For the metric concerning Implementation of the OECD AI
Principles, we have awarded top scores to Canada, France,
Japan, and Korea, the countries that were central to the
development and implementation of the first global framework
for AI policy.
• Additional efforts were made to normalize scores across key
metrics. For example, the determination of implementation of
the Universal Declaration for Human Rights now tracks the
designation of Freedom House for countries as “Free,” “Partly
Free,” and “Not Free.”
• The scores for country reports previously published were
reviewed and revised based on developments during the past
year concerning AI policies and practices.
• The number of researchers participating in the project has grown
significantly. The 2022 CAIDP Research Group now includes
over 100 participants from almost 40 countries.
• We acknowledge the comments of several reviewers who
recommended a more detailed approach to the review of human
rights. Additional recommendations concern expanded coverage
of AI and immigration and Al and criminal justice. We will
address these topics in the next edition.
2
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Findings
•
•
•
•
•
•
•
The OECD/G20 AI Principles have Framed the Global Debate
over AI policy. There are hundreds of frameworks for ethical AI,
but only the OECD/G20 Principles have significantly shaped the
policies and practices of national governments. Over 50
governments have formally endorsed the OECD/G20 AI Principles.
Governments have Both National Ambitions and Collaborative
Goals. National AI policies typically reflect ambitions to be a leader
in AI, to establish centers of AI excellence, and to promote
economic growth. Many of these ambitions will set countries in
competition for investment, personnel, and deployment. At the same
time, countries recognize the need for global cooperation in such
areas as public health, climate change, and sustainable development.
AI Safeguards Build on Data Protection Law. AI policy
safeguards follows from other laws and policy frameworks, most
notably data protection. The GDPR (Article 22), the Modernized
Council of Europe Privacy Convention (Article 9), and the recently
adopted California Privacy Rights Act in the US include explicit
provisions for AI. The Global Privacy Assembly, the international
conference of data protection officials, has recently adopted a
sweeping resolution on the need for AI accountability.
Facial Surveillance as an AI “Red Line.” Few AI applications are
more controversial than the use of AI for surveillance in public
spaces. The use of facial recognition on a general population has
raised widespread controversy with many NGOs stating it should be
prohibited. Other controversial AI applications include the scoring
of citizens, criminal sentencing, administrative service decisions,
and hiring assessments.
Concern About Autonomous Weapons Remains. The risk of
lethal autonomous weapons was among the first AI issues to focus
the attention of government policymakers. Although many other AI
policy issues have emerged in the last few years, concerns about
autonomous weapons remains.
NGOs are Powerful Advocates for the Public. In Europe, civil
society groups have published substantial reports on AI policy,
documented abuses, and called for reform. Their advocacy has also
strengthened democratic institutions which must now consider legal
measures to address public concerns.
AI Policy is in the Early Days, but the Pace is Accelerating. AI
research can be traced back to the 1950s but the effort of national
3
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
governments to develop formal frameworks for AI policy is a recent
phenomenon. Governments around the world are moving rapidly to
understand the implications of the deployment of AI as more
systems are deployed. We anticipate that the rate of AI
policymaking will accelerate in the next few years.
Recommendations
1. Countries must establish national policies for AI that implement
democratic values
2. Countries must ensure public participation in AI policymaking and
also create robust mechanisms for independent oversight of AI
systems
3. Countries must guarantee fairness, accountability, and
transparency in all AI systems
4. Countries must commit to these principles in the development,
procurement, and implementation of AI systems for public services
5. Countries must halt the use of facial recognition for mass
surveillance
New Recommendations (2022)
6. Countries must curtail the deployment of lethal autonomous
weapons
7. Countries must begin implementation of the UNESCO AI
Recommendation
8. Countries must establish a comprehensive, legally binding
convention for AI
4
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
THE GLOBAL AI POLICY LANDSCAPE
As a field of research, AI policy is in the very early stages. Only in
the last few years have national governments formally considered and
adopted policy frameworks that explicitly discuss “Artificial Intelligence.”2
While government funding for work on Artificial Intelligence goes back to
the mid-1950s, it would be many years before governments examined the
consequences of this research. That gap is now closing. Governments
around the world confront important decisions about AI priorities, AI
ambitions, and AI risks. Much of this report concerns the current policies
and practices of national governments.
In addition to national governments, many intergovernmental
organizations are pursuing AI policies and initiatives. This section provides
an overview of these organizations, listed in a simple A to Z. We also note
the important work of technical associations and civil society organizations
This section briefly summarizes these activities, as of early 2022.
The Council of Europe
The Council of Europe (COE) is the continent’s leading human
rights organization.3 The COE is comprised of 47 member states, 27 of
which are members of the European Union. All COE member states have
endorsed the European Convention of Human Rights, a treaty designed to
protect human rights, democracy and the rule of law. Article 8 of the
Convention, concerning the right to privacy, has influenced the
development of privacy law around the world.
The COE Convention 108 (1981) is the first binding international
instrument which protects the individual against abuses which may
accompany the collection and processing of personal data and which
regulates the transborder flow of personal data.4
In 2018, the Council of Europe amended Convention 108 and
opened for signature and ratification the COE Modernized Convention
108+.5 Article 9(1)(c) specifically addresses AI decision-making. As the
COE explains, the “modernised Convention extends the catalogue of
information to be transmitted to data subjects when they exercise their right
Marc Rotenberg, AI Policy Sourcebook (2019, 2020).
Council of Europe, Who we are, https://www.coe.int/en/web/about-us/who-we-are
4
Council of Europe, Treaty office, Details of Treaty No. 108,
https://www.coe.int/en/web/conventions/full-list/-/conventions/treaty/108
5
Council of Europe, Data Protection, Modernisation of Convention 108,
https://www.coe.int/en/web/data-protection/convention108/modernised
2
3
5
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of access. Furthermore, data subjects are entitled to obtain knowledge of the
reasoning underlying the data processing, the results of which are applied
to her/him. This new right is particularly important in terms of profiling of
individuals.”6 Forty-two states have signed the protocol amending the
Privacy Convention.7
Several new AI initiatives are underway at the Council of Europe,
including at the Council of Ministers, the COE Parliamentary Assembly.
Marija Pejčinović Burić, Secretary General of the Council of Europe, has
said “It is clear that AI presents both benefits and risks. We need to ensure
that AI promotes and protects our standards. I look forward to the outcome
of the work of the Ad hoc Committee on Artificial Intelligence (CAHAI), .
. . The Council of Europe has, on many occasions, demonstrated its ability
to pioneer new standards, which have become global benchmarks.”8
Citing the risks to privacy and data protection in 2021, the Council
of Europe called for strict rules to limit the use of facial recognition.9 The
guidelines were developed by the Consultative Committee of the Council
of Europe Convention for the Protection of Individuals with regard to
Automatic Processing of Personal Data,10 reflecting the close connection
between traditional rules for data protection and the emerging realm of AI
policy.
CAHAI
The COE Council of Ministers established the Ad Hoc Committee
on Artificial Intelligence (CAHAI) in September 2019.11 The aim of the
CAHAI is to “examine the feasibility and potential elements on the basis of
broad multi-stakeholder consultations, of a legal framework for the
Council of Europe, Data Protection, Modernisation of Convention 108: Overview of the
novelties, https://rm.coe.int/modernised-conv-overview-of-the-novelties/16808accf8
7
Council of Europe, Treaty Office, Chart of signatures and ratifications of Treaty 223
(Status as of Nov. 22, 2020), https://www.coe.int/en/web/conventions/full-list//conventions/treaty/223/signatures
8
Council of Europe, Artificial intelligence and human rights,
https://www.coe.int/en/web/artificial-intelligence/secretary-general-marija-pejcinovicburic
9
Council of Europe, Facial recognition: strict regulation is needed to prevent human
rights violations (Jan. 28, 2021), https://www.coe.int/en/web/portal/-/facial-recognitionstrict-regulation-is-needed-to-prevent-human-rights-violations10
Council of Europe, Details of Treaty No.108 of 1981,
https://www.coe.int/en/web/conventions/full-list?module=treaty-detail&treatynum=108
11
Council of Europe, The Council of Europe established an Ad Hoc Committee on
Artificial Intelligence - CAHAI (Sept. 11, 2019), https://www.coe.int/en/web/artificialintelligence/-/the-council-of-europe-established-an-ad-hoc-committee-on-artificialintelligence-cahai
6
6
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
development, design and application of artificial intelligence, based on the
Council of Europe’s standards on human rights, democracy and the rule of
law.”12 The Council of Ministers approved the first progress report of the
CAHAI in September 2020.13
The CAHAI held its final meeting in December 2021.14 At the end
of the meeting, the CAHAI adopted the “Possible elements of a legal
framework on artificial intelligence, based on the Council of Europe’s
standards on human rights, democracy and the rule of law.” The CAHAI
framework contains an outline of the legal and other elements which in the
view of the Committee could be included in legally binding or non-legally
binding instruments that will make up an appropriate legal framework on
AI of the Council of Europe. The document outlines the “Possible elements
of a legal framework on artificial intelligence, based on the Council of
Europe’s standards on human rights, democracy and the rule of law.” The
CAHAI framework will now be submitted to the Committee of Ministers
for further consideration.
Committee of Ministers
In September 2020 the Committee of Ministers approved
the CAHAI progress report, which concluded that the “Council of Europe
has a crucial role to play today to ensure that AI applications are in line with
human rights protections.”15 The Committee of Ministers asked the CAHAI
to draft a feasibility study on a legal instrument that could “regulate the
design, development and application of AI that have a significant impact on
human rights, democracy and the rule of law.” The Committee of Ministers
also proposed that the CAHAI should examine “human rights impact
assessments” and “certification of algorithms and AI systems.” The
Committee of Ministers will review the recommendation in early February
2022. These initiatives follow the 2020 Recommendation of the Committee
of Ministers to member States on the human rights impacts of algorithmic
Council of Europe, CAHAI - Ad hoc Committee on Artificial Intelligence,
https://www.coe.int/en/web/artificial-intelligence/cahai
13
Council of Europe, Ad hoc Committee on Artificial Intelligence (CAHAI): Progress
Report (Sept. 23, 2020),
https://search.coe.int/cm/Pages/result_details.aspx?ObjectID=09000016809ed062
14
Council of Europe, The CAHAI held its 6th and final plenary meeting (Dec. 2, 2021),
https://www.coe.int/en/web/artificial-intelligence/-/outcome-of-cahai-s-6th-plenarymeeting
15
Council of Europe, Ad hoc Committee on Artificial Intelligence (CAHAI): Progress
Report (Sept. 23, 2020),
https://search.coe.int/cm/Pages/result_details.aspx?ObjectID=09000016809ed062
12
7
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
systems16 and its 2019 Declaration on the manipulative capabilities of
algorithmic processes.17
In March 2021, the Committee of Ministers issued a comprehensive
declaration on the need to ensure that AI systems for social services respect
human rights.18 The Committee emphasized said that such systems should
be developed and implemented in accordance with the principles of legal
certainty, legality, data quality, non-discrimination, and transparency. The
Ministers also recommended effective arrangements to protect vulnerable
persons from serious or irreparable harm.
In November 2021, the Committee of Ministers also issued a
Recommendation on the protection of individuals with regard to automatic
processing of personal data in the context of profiling. The Committee
stressed that “respect for fundamental rights and freedoms, notably the
rights to human dignity and to privacy but also to freedom of expression,
and for the principle of non-discrimination and the imperatives of social
justice, cultural diversity and democracy, should be guaranteed, in both the
public and private sectors, during the profiling operations.”19
European Committee on Crime Problems
In September 2021, based on the results of the 2020 Feasibility
Study on a future Council of Europe instrument on artificial intelligence and
criminal law,20 the CDPC set up a Drafting Committee consisting of experts
appointed by the members of the CDPC tasked with the elaboration of an
Committee of Ministers, Recommendation CM/Rec(2020)1 on the human rights
impacts of algorithmic systems (Apr. 8, 2020)
https://search.coe.int/cm/pages/result_details.aspx?objectid=09000016809e1154
17
Committee of Ministers, Declaration on the manipulative capabilities of algorithmic
processes (Feb. 13, 2019)
18
Council of Europe, Declaration by the Committee of Ministers: the use of computerassisted or AI-enabled decision making by public authorities in the area of social
services must respect human rights (Mar. 17, 2021),
https://www.coe.int/en/web/artificial-intelligence/newsroom//asset_publisher/csARLoSVrbAH/content/declaration-by-the-committee-of-ministersthe-use-of-computer-assisted-or-ai-enabled-decision-making-by-public-authorities-in-thearea-of-social-servi
19
Committee of Ministers, Recommendation CM/Rec(2021)8 on the protection of
individuals with regard to automatic processing of personal data in the context of
profiling (Nov. 3, 2021)
https://search.coe.int/cm/pages/result_details.aspx?ObjectId=0900001680a46147
20
European Committee on Crime Problems, Feasibility Study on a future Council of
Europe instrument on artificial intelligence and criminal law (Sept.,4 2020)
https://rm.coe.int/cdpc-2020-3-feasibility-study-of-a-future-instrument-on-ai-andcrimina/16809f9b60
16
8
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
instrument on AI and criminal law related to vehicles and automated
driving.21 One of the main purposes of this instrument would be to “ensure
the development of AI systems in accordance with the fundamental rights
protected by Council of Europe instruments.”22 In November 2021, the
Drafting Committee held its first meeting but failed to agree on the
bindingness of the instrument.23
Parliamentary Assembly
In October 2020, the Parliament Assembly of the Council of Europe
has adopted a new resolution on the Need for Democratic Governance of
Artificial Intelligence.24 The Assembly called for “strong and swift action”
by the Council of Europe. The parliamentarians warned that “soft-law
instruments and self-regulation have proven so far not sufficient in
addressing these challenges and in protecting human rights, democracy and
rule of law.”
In a set of recommendations examining the opportunities and risks
of AI for democracy, human rights and the rule of law adopted in October
2020 as well, the Parliamentary Assembly called on the Committee of
Ministers to take into account the particularly serious potential impact of
the use of artificial intelligence “in policing and criminal justice systems”25
or “on the enjoyment of the rights to equality and non-discrimination”, 26
when assessing the necessity and feasibility of an international legal
framework for artificial intelligence.
21
European Committee on Crime Problems, Drafting Committee to elaborate an
instrument on artificial intelligence and criminal law – Terms of Reference / Working
methods (Sept. 16, 2021) https://rm.coe.int/cdpc-2021-2-terms-of-reference-cdpcaicl/1680a18ffe
22
European Committee on Crime Problems, Drafting Committee to elaborate an
instrument on artificial intelligence and criminal law – Terms of Reference / Working
methods, op. cit., p. 11.
23
European Committee on Crime Problems, 1st meeting of the Drafting Committee to
elaborate an instrument on Artificial Intelligence and Criminal Law (Nov. 15-16 2021)
24
Council of Europe, Parliamentary Assembly, Need for democratic governance of
artificial intelligence (Oct. 22, 2020), https://pace.coe.int/en/files/28803/html
25
Parliamentary Assembly, Recommendation 2182(2020) Justice by algorithm – The role
of artificial intelligence in policing and criminal justice systems (Oct. 22, 2020)
https://pace.coe.int/en/files/28806/html; See also, Resolution 2342 (2020)
https://pace.coe.int/en/files/28805
26
Parliamentary Assembly, Recommendation 2183 (2020) Preventing discrimination
caused by the use of artificial intelligence (Oct. 22, 2020)
https://pace.coe.int/en/files/28809/html; See also, Resolution 2343 (2020)
https://pace.coe.int/en/files/25318/html
9
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
European Court of Human Rights
The European Court of Human Rights has generated an abundant
amount of case law interpreting Article 8 of the European Convention on
Human Rights on the right to private life. The opinions of the Court on
privacy and data protection are widely regarded by other courts. The Court
has dealt with numerous aspects relating to the protection of personal data,
which has been deemed of fundamental importance to a person’s enjoyment
of a person’s right to respect for private and family life as guaranteed by
Article 8. The Court has addressed privacy challenges in relation to
telephone conversations, telephone numbers, computers, video
surveillance, voice recording, bulk interceptions of telecommunications and
the internet.27 However, to date, the Court has not addressed matters relating
to AI tools, including automated decision-making based on algorithms.
Commissioner for Human Rights
In January 2021, at a virtual event organized by the German Federal
Foreign Office and Federal Ministry of Justice and Consumer Protection as
part of Germany’s Chairmanship of the Committee of Ministers of the
Council of Europe, the Commissioner for Human Rights started her speech
on “Human Rights in the Era of AI – Europe as international Standard Setter
for Artificial Intelligence” by asserting that “Ensuring that technological
development works for and not against human rights, democracy and the
rule of law is one of the biggest tasks that states face”.28
Her speech refers to and builds on the 10-point Recommendation on
AI and human rights she addressed to Council of Europe member states in
May 2019.29 It focused more specifically on 1) Human rights impact
assessment, 2) Public consultations 3) Obligations of member states to
facilitate the implementation of human rights standards in the private sector
4) Information and transparency 5) Independent oversight 6) Nondiscrimination and equality 7) Data protection and privacy 8) Freedom of
For an overview of the case law, see European Court of Human Rights, Mass
surveillance (Jan. 2022)
https://www.echr.coe.int/documents/fs_mass_surveillance_eng.pdf; Personal data
protection (Jan. 2022) https://www.echr.coe.int/Documents/FS_Data_ENG.pdf.
28
Commissioner for Human Rights, Human Rights in the Era of AI – Europe as
international Standard Setter for Artificial Intelligence (Jan. 20, 2021)
https://rm.coe.int/german-cm-presidency-high-level-conference-human-rights-in-the-eraof-/1680a12379
29
Commissioner for Human Rights, Recommendation, Unboxing Artificial Intelligence:
10 steps to protect Human Rights (May 2019) https://rm.coe.int/unboxing-artificialintelligence-10-steps-to-protect-human-rights-reco/1680946e64
27
10
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
expression, freedom of assembly and association, and the right to work 9)
Remedies 10) Promotion of “AI literacy.”
Commission for the Efficiency of Justice
In December 2020, The European Commission for the Efficiency of
Justice (CEPEJ) adopted a feasibility study on the establishment of a
certification mechanism for artificial intelligence tools and services. The
study is based on the CEPEJ Charter on the use of artificial intelligence in
judicial systems. According to the CEPEJ, the Council of Europe, if it
decides to create such a mechanism, would be a pioneer in this field.30
In December 2021, the CEPEJ adopted the 2022-2025 Action plan:
“Digitalisation for a better justice.” The CEPEJ Action Plan sets out as the
priority assisting “States and courts in a successful transition towards
digitalisation of justice in line with European standards and in particular
Article 6 of the European Convention of Human Rights” on the right to a
fair trial, “while also ensuring that justice is human, efficient and of high
quality.” “Human justice” is presented as one of the main goals the CEPEJ
should take into account: “The digitalisation of justice shall make justice
more efficient but must never seek to replace the judge. The judge must
remain at the centre of the procedure.”31
The European Union
Many institutions in the European Union now play a significant role
in the development of AI policies and practices.
The European Commission
The European Commission plays an active role in developing the
EU’s overall strategy and in designing and implementing EU policies. The
Commission is the initiator of EU legislation. AI was identified as a priority
when the new Commission, under the Presidency of Ursula von der Leyen,
was established in late 2019.32 At that time, von der Leyen
Council of Europe, CEPEJ: Artificial intelligence and cyberjustice at the heart of the
discussions (Dec. 11, 2020), https://www.coe.int/en/web/portal/-/cepej-artificialintelligence-and-cyberjustice-at-the-heart-of-discussions
31
European Commission for the Efficiency of Justice, 2022-2025 CEPEJ Action plan:
“Digitalisation for a better justice”, CEPEJ(2021)12Final (Dec. 8-9, 2021)
https://rm.coe.int/cepej-2021-12-en-cepej-action-plan-2022-2025-digitalisationjustice/1680a4cf2c
32
CAID Update 1.3, European Commission Proposes Options for Ethical, (Aug. 3,
2020), https://dukakis.org/center-for-ai-and-digital-policy/center-for-ai-policy-updateeuropean-commission-proposes-four-options-for-ethical-ai/
30
11
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
recommended new rules on Artificial Intelligence that respect human safety
and rights.33
Von der Leyen’s proposal followed remarks by Chancellor Angela
Merkel at the G20 summit in 2019, who called for the European
Commission to propose comprehensive regulation for artificial intelligence.
“It will be the job of the next Commission to deliver something so that we
have regulation similar to the General Data Protection Regulation that
makes it clear that artificial intelligence serves humanity,” Merkel stated.
In February 2020, the Commission published the white paper On
Artificial Intelligence -A European Approach to Excellence and Trust for
public comment. The Commission subsequently proposed several options
for AI regulation. Speaking to the EU Ambassadors Conference in
November 2020, President von der Leyen said, “European rules on personal
data protection have inspired others to modernise their own privacy rules.
We must now put special focus on the international transfer of data,
particularly after a recent ruling of the European Court of Justice.”34 And in
remarks to the Council on Foreign Relations, she said “we must work
together on a human-centric vision on AI - a global standard aligned with
our values.”35
Following the U.S. election in November 2020, the European
Commission developed a new framework for transatlantic relations. On
December 2, 2020, the European Commission proposed a New EU-US
Agenda for Global Change. The New Agenda covers a wide range of topics,
but it is notable that the Commission states, “we need to start acting together
on AI - based on our shared belief in a human- centric approach and dealing
with issues such as facial recognition. In this spirit, the EU will propose to
start work on a Transatlantic AI Agreement to set a blueprint for regional
and global standards aligned with our values.”36 The Commission further
states, “We must also openly discuss diverging views on data governance
and see how these can be overcome constructively. The EU and the US
33
European Commission, A Union that Strives for more: the first 100 days (Mar. 6,
2020), https://ec.europa.eu/commission/presscorner/detail/en/ip_20_403
34
European Commission, Speech by President von der Leyen at the EU Ambassadors'
Conference 2020 (Nov. 10, 2020),
https://ec.europa.eu/commission/presscorner/detail/en/SPEECH_20_2064
35
Council on Foreign Relations, A Conversation with Ursula von der Leyen (Nov. 20,
2020), https://www.cfr.org/event
36
European Commission and High Representative of the Union for Foreign Affairs and
Security Policy, Joint Communication to the European Parliament, the European
Council and the Council: A New EU-US Agenda for Global Change (Dec. 2, 2020)
(emphasis in the original), https://ec.europa.eu/info/sites/info/files/joint-communicationeu-us-agenda_en.pdf
12
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
should intensify their cooperation at bilateral and multilateral level to
promote regulatory convergence and facilitate free data flow with trust on
the basis of high standards and safeguards.”
The Trade and Technology Council
At the US-EU Summit in Brussels in June 2021, President von der
Leyen launched together with US President Biden the EU-US Trade and
Technology Council (TTC). One of its main purposes is to coordinate
approaches to key technology issues and deepen transatlantic trade and
economic relations based on shared democratic values. The Trade and
Technology Council will include a working group on technology standards
cooperation including AI and another one on the misuse of technology
threatening security and human rights. For the EU, the TTC is co-chaired
by European Commission Executive Vice Presidents Valdis Dombrovskis
and Margrethe Vestager and for the US by Trade Representative Katherine
Tai, Secretary of Commerce Gina Raimondo and Secretary of State
Anthony Blinken.
In a joint statement following the TTC inaugural meeting in
Pittsburgh in September 2021, “the European Union and the United States
acknowledge that AI technologies yield powerful advances but also can
threaten our shared values and fundamental freedoms if they are not
developed and deployed responsibly or if they are misused. The European
Union and the United States affirm their willingness and intention to
develop and implement AI systems that are innovative and trustworthy and
that respect universal human rights and shared democratic values.”37 They
also agreed on the importance of public consultation as the TTC undertakes
its work. As a result, in October 2021, the Commission launched an online
consultation platform on the TTC38 allowing stakeholders to share their
views and make recommendations as well as be informed about its work.
The EU AI Act
In April 2021 The European Commission published the ”AI
package. This package consisted of: a Communication on Fostering a
European Approach to Artificial Intelligence; the Coordinated Plan with
Member States: 2021 update; a proposal for an AI Regulation laying down
harmonised rules for the EU (the ”AI Act”).39 In January 2022, The
EU-US Trade and Technology Council Inaugural Joint Statement (Sept. 29, 2021)
https://ec.europa.eu/commission/presscorner/detail/e%20n/statement_21_4951
38
Futurium Platform, Trade and Technology Council Community
https://futurium.ec.europa.eu/en/EU-US-TTC
39
European Commission, A European approach to artificial intelligence, https://digitalstrategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence.
37
13
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
European Commission proposed to define a set of principles for a humancentered digital transformation.40
The draft AI Act follows a risk-based approach and proposes to
categorize AI systems based on the four different risk levels they create: 1)
an unacceptable risk; 2) a high risk; 3) limited risk; or 4) minimal risk. No
limitations or requirements are set for use of AI systems creating minimal
or low risk.
The draft AI Act prohibits certain AI practices that create
unacceptable risk as they contradict EU values and fundamental rights. The
draft Act proposes to prohibit four AI practices: 1) deployment of
subliminal techniques beyond a person’s consciousness, 2) exploitation of
the vulnerabilities of specific vulnerable groups, 3) social scoring, and 4)
use of ‘real-time’ remote biometric identification systems in publicly
accessible spaces for law enforcement.
The draft AI Act sets out specific requirements for high-risk AI
systems, that create an adverse impact on safety or fundamental rights. This
includes AI systems that are product or safety components or systems used
in the areas listed in Annex III of the draft AI Act, including such areas as
biometric identification and categorization, education, employment, law
enforcement,
migration,
asylum
and
border
control.
For other AI systems that do not pose high risks, the draft AI Act imposes
limited transparency rules. The draft Act classifies as limited-risk AI
systems intended to interact with natural persons, emotion recognition
systems and biometric categorization systems, and AI systems used to
generate or manipulate image, audio or video content.
The Commission proposal is subject to review and amendment by
the Parliament and the Council, and then a subsequent negotiation, known
as the “trialogue.”41
Fundamental Rights in the Digital Age
In December 2021, the European Commission released its annual
report on the application of the Charter of Fundamental Rights in the EU. It
is the first thematic report and it focuses on the challenges in protecting
fundamental rights in the digital age.42 One of the key policy areas of the
40
European Commission, Declaration on European Digital Rights and Principles (Jan.
26, 2022), https://digital-strategy.ec.europa.eu/en/library/declaration-european-digitalrights-and-principles
41
European Council, Council of the European Union, The ordinary legislative procedure,
https://www.consilium.europa.eu/en/council-eu/decision-making/ordinary-legislativeprocedure/
42
European Commission, Protecting Fundamental Rights in the Digital Age – 2021
Annual Report on the Application of the EU Charter of Fundamental Rights, COM(2021)
14
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
report concerns “Safeguarding fundamental rights where artificial
intelligence is used” and another one “Supervising digital surveillance”
with a paragraph dedicated to remote biometric identification.
In January 2022, the European Commission proposed to define a set
of principles for a human-centered digital transformation in an
interinstitutional Declaration.43 This was one of the four cardinal points
identified by the Commission in its Digital Compass in which it set its vision
for a successful digital transformation of Europe by 2030.44 The European
Parliament and the Council are invited to discuss then endorse the
Declaration by the Summer of 2022.
The European Parliament
The European Parliament is co-legislator, together with the Council
of the European Union. The Parliament has convened hearings and adopted
resolution to outline the element of EU legislation.45 One resolution urged
the Commission to establish legal obligations for artificial intelligence and
robotics, including software, algorithms and data. A second would make
those operating high-risk AI systems strictly liable for any resulting
damage. And a third resolution on intellectual property rights makes clear
that AI should not have legal personality; only people may claim IP rights.
The European Parliament adopted all of these proposals in sweeping
majorities, across parties and regions. But even those proposals are unlikely
to meet the concerns of civil society. As Access Now and EDRi said of the
resolution on AI ethics, “They are cautious and restrained on fundamental
rights, taking only tentative steps to outline the biggest threats that artificial
intelligence pose to people and society, while also failing to propose a
legislative framework that would address these threats or provide any
substantive protections for people’s rights.”
819 final (Dec. 12, 2021),
https://ec.europa.eu/info/sites/default/files/1_1_179442_ann_rep_en_0.pdf
43
European Commission, Declaration on European Digital Rights and Principles (Jan.
26, 2022), https://digital-strategy.ec.europa.eu/en/library/declaration-european-digitalrights-and-principles
44
European Commission, 2030 Digital Compass: the European way for the Digital
Decade COM(2021) 118 final, (March 9, 2021) https://eur-lex.europa.eu/legalcontent/en/TXT/?uri=CELEX:52021DC0118
45
CAIDP Update 1.12, European Parliament Adopts Resolutions on AI (Oct. 24, 2020),
https://dukakis.org/center-for-ai-and-digital-policy/caidp-update-european-parliamentadopts-resolutions-on-ai/
15
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The influential LIBE Committee has also highlighted concerns
about AI and fundamental rights and AI in criminal justice.46 In February
2020, the Committee held a hearing on Artificial Intelligence and Criminal
Law, and examined the benefits and risks of AI, predictive policing, facial
recognition, as well as the ethical and fundamental rights implications.
LIBE worked in association with the United Nations Interregional Crime
and Justice Research Institute (UNICRI), the European Union Agency for
Fundamental Rights (FRA), and the Council of Europe (COE). In
November 2020, LIBE issued an opinion concerning AI and the application
of international law.47
The JURI Committee, responsible for Legal Affairs, also requested
a significant report on Artificial Intelligence and Civil Liability.48 The
report “demonstrates how technology regulation should be technologyspecific, and presents a Risk Management Approach, where the party who
is best capable of controlling and managing a technology-related risk is held
strictly liable, as a single entry point for litigation.” The report outlines the
application to four case studies. Following the European Parliament’s
October 2020 resolution on the topic, the European Commission published
an inception impact assessment on a likely legislative initiative to adapt the
EU liability rules to the digital age and circular economy in June 202149 and
launched a public consultation on the topic from October 2021 until January
2022.50
CAIDP Update 1.8 LIBE Committee of EU Parliament Examines AI Practices, Data
Protection, (Sept. 9, 2020), https://dukakis.org/center-for-ai-and-digital-policy/caidpupdate-libe-committee-of-eu-parliament-examines-ai-practices-data-protection/
47
European Parliament, Committee on Civil Liberties, Justice and Home Affairs, on
artificial intelligence: questions of interpretation and application of international law in
so far as the EU is affected in the areas of civil and military uses and of state authority
outside the scope of criminal justice (2020/2013 (INI)), (Nov. 23, 2020),
https://www.europarl.europa.eu/doceo/document/LIBE-AD-652639_EN.pdf
48
Policy Department for Citizens' Rights and Constitutional Affairs, Directorate-General
for Internal Policies, Artificial Intelligence and Civil Liability, PE 621.296 JURI (July 14,
2020),
https://www.europarl.europa.eu/RegData/etudes/STUD/2020/621926/IPOL_STU(2020)6
21926_EN.pdf
49
European Commission, Inception Impact Assessment - Adapting liability rules to the
digital age and circular economy (Jun. 30, 2021) https://ec.europa.eu/info/law/betterregulation/have-your-say/initiatives/12979-Civil-liability-adapting-liability-rules-to-thedigital-age-and-artificial-intelligence_en
50
European Commission, Public consultation on Civil liability – adapting liability rules
to the digital age and artificial intelligence (Oct. 18, 2021)
https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/12979-Civil46
16
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Following a report by the JURI Committee, the European
Parliament adopted in January 2021 a resolution on “artificial intelligence:
questions of interpretation and application of international law in so far as
the EU is affected in the areas of civil and military uses and of state authority
outside the scope of criminal justice”51. In its resolution, the Parliament
reiterated its call for an EU strategy to prohibit the use of lethal autonomous
weapon systems and for a ban on “killer robots”. It also called for the EU
to play a leading role in creating and promoting a global framework
governing the military use of AI. Regarding the use of AI in the public
sector, especially healthcare and justice, the Parliament stressed that “the
use of AI systems in the decision-making process of public authorities can
result in biased decisions that negatively affect citizens, and therefore
should be subject to strict control criteria regarding their security,
transparency, accountability, non-discrimination, social and environmental
responsibility, among others; urges Member States to assess the risks
related to AI-driven decisions connected with the exercise of State
authority, and to provide for safeguards such as meaningful human
supervision, transparency requirements and the possibility to contest such
decisions”. The Parliament also invited the Commission to “assess the
consequences of a moratorium on the use of facial recognition systems, and,
depending on the results of this assessment, to consider a moratorium on
the use of these systems in public spaces by public authorities and in
premises meant for education and healthcare, as well as on the use of facial
recognition systems by law enforcement authorities in semi-public spaces
such as airports, until the technical standards can be considered fully
fundamental rights-compliant, the results derived are non-biased and nondiscriminatory, and there are strict safeguards against misuse that ensure the
necessity and proportionality of using such technologies.”
In May 2020, the Directorate General for Parliamentary Research
Services of the European Parliament published The Impact of the General
Data Protection Regulation (GDPR) on Artificial Intelligence.52 The study
liability-adapting-liability-rules-to-the-digital-age-and-artificial-intelligence/publicconsultation_en
51
European Parliament, Resolution on artificial intelligence: questions of interpretation
and application of international law in so far as the EU is affected in the areas of civil
and military uses and of state authority outside the scope of criminal justice
(2020/2013(INI)), (Jan. 20, 2021) https://www.europarl.europa.eu/doceo/document/TA9-2021-0009_EN.html
52
European Parliament Think Tank, The impact of the General Data Protection
Regulation (GDPR) on artificial intelligence (June 25, 2020),
https://www.europarl.europa.eu/thinktank/en/document.html?reference=EPRS_STU(202
0)641530
17
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
examines the tensions and proximities between AI and data protection
principles, such as in particular purpose limitation and data minimization.
And in June 2020 the European Parliament established a Special Committee
on Artificial Intelligence to study the impact of AI and to propose a roadmap
for the EU. According to the decision of Parliament, the Committee should
pursue a “holistic approach providing a common, long-term position that
highlights the EU’s key values and objectives.”53
The work of the European Parliament on Artificial Intelligence also
intersects with the Digital Services Act, an initiative to overhaul the ECommerce Directive which has been the foundation of the digital single
market for the last twenty years.54 At the end of October, 2020,
European Margrethe Vestager said the proposed Digital Services Act
package will aim to make ad targeting more transparent and to ensure
companies are held accountable for their decisions.55 “The biggest
platforms would have to provide more information on the way their
algorithms work, when regulators ask for it,” Vestager said.
Committees – AIDA, IMCO, LIBE
There are three committees within the European Parliament that
have primary jurisdiction for the development of AI policy. The AIDA
Committee - the Special Committee on Artificial Intelligence in a Digital
Age – was established by the European Parliament on June 18, 2020 with
the goal of “setting out a long-term EU roadmap on Artificial Intelligence
(AI).”56 Over an 18-month period, AIDA organized hearings and
workshops with key stakeholders, including experts, policy-makers, and the
business community. In November 2021, members of the AIDA committee
met with policymakers, NGOs, and business groups in Washington, DC.
European Parliament, Setting up a special committee on artificial intelligence in a
digital age, and defining its responsibilities, numerical strength and term of office (June
18, 2020), https://www.europarl.europa.eu/doceo/document/TA-9-2020-0162_EN.html
54
European Parliament, Digital Services Act: Opportunities and Challenges for the
Digital Single Market and Consumer Protection,
https://www.europarl.europa.eu/RegData/etudes/BRIE/2020/652712/IPOL_BRI(2020)65
2712_EN.pdf
55
Matthew Broersma, New EU Rules ‘Would Open Tech Giants’ Algorithms To Scrutiny,
Silicon.co (Nov. 2, 2020) https://www.silicon.co.uk/workspace/algorithms-tech-giants348707
56
European Parliament, AIDA Committee, About: Welcome Words,
https://www.europarl.europa.eu/committees/en/aida/about
53
18
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In January 2022, the rapporteur of the AIDA Committee published
a draft report on artificial intelligence in a digital age.57 Approximately
1,400 amendments were received.58 AIDA committee anticipated the
finalization of the report and a vote on the associated resolution in March
2022.
Two committees in the European Parliament will then take the reins
for the proposed EU AI Act. The IMCO Committee is responsible for the
legislative oversight and scrutiny of EU rules on the single market,
including the digital single market, customs and consumer protection.59 The
LIBE Committee is “is responsible for the majority of legislation and
democratic oversight of policies that enable the European Union to offer its
citizens an area of freedom, security and justice (Article 3 TEU). While
doing so, we ensure, throughout the EU, the full respect of and compliance
with the EU Charter of Fundamental Rights, in conjunction with the
European Convention on Human Rights.”60
A joint hearing between IMCO and LIBE was held on January 25,
61
2022. The two rapporteurs expressed their views on the AI Act. Brando
Benifei, co-rapporteur for the Internal Market and Consumer Protection
Committee, stated “Our aim is to protect citizens and consumers, and
stimulate positive innovation at the same time, while focussing especially
on SMEs and start-ups. A legislative framework ensuring that AI systems
entering the EU single market are safe, human-centric and respect our
fundamental rights and freedoms will stimulate trust among citizens, which
is key to a successful and inclusive uptake of AI on our continent. That is
what we will strive for.” Dragoş Tudorache, co-rapporteur for the Civil
Liberties, Justice and Home Affairs Committee, said, “The AI Act is a
central piece of the European regulatory environment for the digital future
and the first of its kind worldwide. We have a chance to lead by example
Special Committee on Artificial Intelligence in a Digital Age, Draft Report on artificial
intelligence in a digital age, (2020/2266(INI)) (Nov. 2, 2021),
https://www.europarl.europa.eu/meetdocs/2014_2019/plmrep/COMMITTEES/AIDA/PR/
2022/01-13/1224166EN.pdf
58
AIDA, AIDA Commttee Meeting, Consideration of Amendments (Feb. 1, 2022),
https://www.europarl.europa.eu/committees/en/aida-committee-meeting-considerationof-/product-details/20220117CAN64673
59
European Parliament, About IMCO, Welcome Words,
https://www.europarl.europa.eu/committees/en/imco/about
60
European Parliament, About LIBE, Welcome Words,
https://www.europarl.europa.eu/committees/en/libe/about
61
Artificial Intelligence Act: EP lead committees to launch joint work on 25 January
(Jan. 24, 2022), https://portal.ieu-monitoring.com/editorial/artificial-intelligence-act-eplead-committees-to-launch-joint-work-on-25-january
57
19
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and to shape the rules of the digital world according to our values. As the
heart of European democracy, the European Parliament has a key role to
play: we need to find the right balance between enhancing the protection of
our fundamental rights and boosting Europe’s competitiveness and capacity
to innovate.”
According to a draft schedule, the IMCO and LIBE committees
anticipate that reports will be made final by June 20, 2022 and a vote on the
reports will occur September 29, 2022. The Parliament is expected to vote
on November 9, 2022.
The Two Councils
The European Council defines the EU's overall political direction
and priorities.62 Its members are the heads of state or government of the 27
EU member states, the European Council President, and the President of the
European Commission. The European Council is not one of the EU's
legislating institutions, so does not negotiate or adopt EU laws. This is the
prerogative of the Council of the European Union (“Council”), composed
of representatives of member states’ ministers.
In June 2020, the Council of the European Union set out
Conclusions for Shaping Europe’s Digital Future.63 Regarding AI, the
Council stressed, some “some artificial intelligence applications can entail
a number of risks, such as biased and opaque decisions affecting citizens’
well-being, human dignity or fundamental rights, such as the rights to nondiscrimination, gender equality, privacy, data protection and physical
integrity, safety and security, thus reproducing and reinforcing stereotypes
and inequalities. Other risks include the misuse for criminal or malicious
purposes such as disinformation.”
And then in October 2020, the European Council issued conclusions
on the charter of fundamental rights in the context of artificial intelligence
and digital change.64 “These conclusions are designed to anchor the EU's
fundamental rights and values in the age of digitalisation, foster the EU's
digital sovereignty and actively contribute to the global debate on the use of
artificial intelligence with a view to shaping the international framework,”
the Presidency of the Council stated.
European Council https://www.consilium.europa.eu/en/european-council/
Council of the European Union, Shaping Europe’s Digital Future (June 9, 2020),
https://data.consilium.europa.eu/doc/document/ST-8711-2020-INIT/en/pdf
64
Council of the European Union, Artificial intelligence: Presidency issues conclusions
on ensuring respect for fundamental rights (Oct. 21, 2020),
https://www.consilium.europa.eu/en/press/press-releases/2020/10/21/artificialintelligence-presidency-issues-conclusions-on-ensuring-respect-for-fundamental-rights/#
62
63
20
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Presidency recommended a “fundamental rights-based”
approach to AI and emphasized dignity, freedoms, equality, solidarity,
citizen’s rights, and justice.65 The Council urged the Union and Member
States to “consider effective measures for identifying, predicting and
responding to the potential impacts of digital technologies, including AI, on
fundamental rights.” The Council said the “Commission’s announced
proposal for a future regulatory framework for AI, should strengthen trust,
strike a fair balance between the various interests and leave room for
research and development and further innovation and technical and sociotechnical developments.” The Council also acknowledged the work of the
FRA on AI.
The Council of the European Union, through the Transport,
Telecommunications and Energy Council, has already set out proposed
changes to the EU AI Act. The Slovenian Presidency (July to December
2021) published in late November 2021 compromise text Articles 1 – 7 of
the AI Act.66 The text strengthens certain provisions but would also exempt
general purpose AI systems.
A compromise text dated January 13th was proposed by the French
Presidency (January to June 2022), addressing Articles 8-15 and Annex
IV.67 The French Presidency aims at adopting a Council position before July
2022. The French proposal would extend the ban on social scoring to private
actors, make clear that obligations for high-risk systems apply to both public
and private authorities, add insurance to the list of high-risk systems, expand
the definition of prohibited system that distort human behaviour, and
expand the limitation on remote identification systems. At the same time,
the French proposal would exclude AI systems “exclusively developed or
used for military purposes” as long as it is used only for military purposes.
A related provision would exclude AI systems that are exclusively
developed or used for national security purposes
The Czech Republic will take up the presidency of the Council for
the second half of 2022. However, since the Parliament is not expected to
finalize its position on the EU AI Act until late 2022, the final negotiations
COE, Presidency conclusions: The Charter of Fundamental Rights in the context of
Artificial Intelligence and Digital Change, 11481/20 (Oct. 21, 2020),
https://www.consilium.europa.eu/media/46496/st11481-en20.pdf
66
Council of the European Union, Presidency Compromise Text (Nov. 29, 2021),
https://data.consilium.europa.eu/doc/document/ST-14278-2021-INIT/en/pdf
67
Council of the European Union, Presidency compromise text - Articles 8-15 and Annex
IV (Jan. 13, 2022), https://www.caidp.org/app/download/8367910663/CAIDP-CongressTTC-Statement-01172022.pdf
65
21
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
between the Parliament, the Council, and the Commission are expected to
take place under the Swedish presidency (January – June 2023).
The Court of Justice of the European Union
Although the Court of Justice has yet to rule directly on AI policies,
the Court will play a significant role as AI policies evolve and AI law is
adopted.68 Judgments of the Court concerning data transfers will also
impact the development of AI systems. In the 2020 Schrems II judgment,
the Court struck down the Privacy Shield framework that permitted the
transfer of personal data from the European Union to the United States.69
The Schrems II judgment will likely limit the collection and use of personal
data for AI systems.
A case currently before the Court of Justice concerns the application
of Article 22 of the GDPR to credit scoring in Germany. The case was
referred by a German court and poses the question whether Article 22(1)
prohibits the “the automated establishment of a probability value
concerning the ability of a data subject to service a loan in the future.”70
AlgorithmWatch has established OpenSchufa, with the goal of making
credit report scoring transparent.71 According to AlgorithmWatch,
“Germany's leading credit bureau, SCHUFA, has immense power over
people's lives. A low SCHUFA score means landlords will refuse to rent
you an apartment, banks will reject your credit card application and network
providers will say 'no' to a new Internet contract.”72
The European Data Protection Board
The European Data Protection Board (EDPB) is an independent
European body, which contributes to the consistent application of data
protection rules throughout the European Union and promotes cooperation
between the EU’s data protection authorities.73
CAIDP Update 1.1, EU Privacy Decision Will Have Global Consequences, (July 19,
2020), https://dukakis.org/news-and-events/center-for-ai-and-digital-policy-update-euprivacy-decision-will-have-global-consequences/
69
CJEU, The Court of Justice invalidates Decision 2016/1250 on the adequacy of the
protection provided by the EU-US Data Protection Shield (July 16, 2020),
https://curia.europa.eu/jcms/upload/docs/application/pdf/2020-07/cp200091en.pdf
70
Court of Justice of the European Union, SCHUFA Holding, Case C-634-21, Request
for a Preliminary Ruling (Oct. 15, 2021).
71
OpenSchufa, cracking the Schufa Code, https://www.startnext.com/en/openschufa
72
AlgorithmWatch, OpenSCHUFA – shedding light on Germany’s opaque credit scoring
(May 22, 2018), https://algorithmwatch.org/en/openschufa-shedding-light-on-germanysopaque-credit-scoring-2/
73
EDPB, Who we are, https://edpb.europa.eu/about-edpb/about-edpb_en
68
22
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In a January 2020 letter to Sophie in’t Veld, EDPB Chair Andrea
Jelinek addressed “the appropriateness of the GDPR as a legal framework
to protect citizens from unfair algorithms” and also whether the EDPB
would issue guidance on the topic.74 Jelinek responded that the GDPR is a
“robust legal framework” to protect citizens’ right to data protection, and
highlighted several articles in the GDPR that would apply to AI systems,
including Article 22, regarding the legal effects of automated processing,
and Article 35, about the obligation to undertake Data Protection Impact
Assessments prior to processing.
Jelinek also warned of specific challenges arising from AI. The
“data maximization presumption of AI “creates an incentive for large and
possibly unlawful data collection and further processing of data.” She also
warned that the opacity of algorithms (the “black box”) can lead to lack of
transparency towards the data subject and also “a loss of human autonomy
for those working with algorithms.” But Jelinek concluded that it would be
“premature at this time” to issue guidance on what constitutes a “fair
algorithm.”
In a June 2020 letter to several members of the European Parliament
about facial recognition and the company ClearView AI, EDPB Chair
Jelinek stated “Facial recognition technology may undermine the right to
respect for private life and the protection of personal data . . .It may also
affect individuals’ reasonable expectation of anonymity in public spaces.
Such technology also raises wider issues from an ethical and societal point
of view.” But Jelinek failed to state whether the use of facial recognition in
public spaces was permissible under the GDRP.75
The European Data Protection Supervisor
The European Data Protection Supervisor is the European Union’s
independent data protection authority.76 The EDPS responsibilities include
the mission to “monitor and ensure the protection of personal data and
privacy when EU institutions and bodies process the personal information
of individuals.” In comments on the Commission’s White Paper on
Artificial Intelligence, the EDPS stated, “benefits, costs and risks should be
considered by anyone adopting a technology, especially by public
EDPB, Letter to MEP Sophie in’t Veld (OUT2020-0004),
https://edpb.europa.eu/sites/edpb/files/files/file1/edpb_letter_out2020_0004_intveldalgori
thms_en.pdf
75
EDPB Letter Members of the European Parliament (OUT2020-0052) (June 10, 2020),
https://edpb.europa.eu/sites/edpb/files/files/file1/edpb_letter_out_20200052_facialrecognition.pdf
76
EDPS, About, https://edps.europa.eu/about-edps_en
74
23
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
administrations who process great amounts of personal data.”77 The EDPS
also expressed support for a moratorium on facial recognition in public
space, “so that an informed and democratic debate can take place and until
the moment when the EU and Member States have all the appropriate
safeguards.”
In June 2021, the EDPB Chair Andrea Jelinek and the EDPS
Wojciech Wiewiórowski issued a joint opinion on the European
Commission’s Proposal for a Regulation laying down harmonized rules on
artificial intelligence (AI). 78 They stressed the need to make clear that
existing EU data protection legislation, including the GDPR, applies to the
processing of personal data falling under the scope of the draft AI
Regulation. They also proposed that compliance with legal obligations
arising from EU legislation - including on personal data protection - should
be a precondition for entering the European market as CE marked product.
They also recommended several “red lines” for AI deployment,
including general ban on any use of AI for automated recognition of human
features in publicly accessible spaces, such as recognition of faces, gait,
fingerprints, DNA, voice, keystrokes and other biometric or behavioral
signals. They proposed a ban on AI systems using biometrics to categorize
individuals into clusters based on ethnicity, gender, political or sexual
orientation, or other grounds on which discrimination is prohibited under
Article 21 of the Charter of Fundamental Rights. Furthermore, the EDPB
and the EDPS said that the use of AI to infer emotions of a natural person
should be prohibited, except for very specified cases. Andrea Jelinek, EDPB
Chair, & Wojciech Wiewiórowski, EDPS, said:
Deploying remote biometric identification in publicly
accessible spaces means the end of anonymity in those
places. Applications such as live facial recognition interfere
with fundamental rights and freedoms to such an extent that
they may call into question the essence of these rights and
freedoms. This calls for an immediate application of the
precautionary approach. A general ban on the use of facial
EDPS, Opinion 4/2020, EDPS Opinion on the European Commission’s White Paper on
Artificial Intelligence – A European approach to excellence and trust (June 29, 2020),
https://edps.europa.eu/sites/edp/files/publication/20-0619_opinion_ai_white_paper_en.pdf
78
EDPB, EDPB & EDPS call for ban on use of AI for automated recognition of human
features in publicly accessible spaces, and some other uses of AI that can lead to unfair
discrimination (June 21, 2021), https://edpb.europa.eu/news/news/2021/edpb-edps-callban-use-ai-automated-recognition-human-features-publicly-accessible_en
77
24
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
recognition in publicly accessible areas is the necessary
starting point if we want to preserve our freedoms and create
a human-centric legal framework for AI. The proposed
regulation should also prohibit any type of use of AI for
social scoring, as it is against the EU fundamental values and
can lead to discrimination.
Fundamental Rights Agency
The EU Agency for Fundamental Rights is also examining the
impact of AI. In 2018, the FRA launched a project on Artificial Intelligence,
Big Data and Fundamental Rights to assesses the use of AI for public
administration and business in the EU.79 A 2018 report explores
discrimination in AI80 and a 2019 FRA report examines facial recognition.81
In mid-December 2020, the German presidency of the EU, in
collaboration with the EU Fundamental Rights Agency and German
Ministry of Justice and Consumer Protection, organized a conference on AI
and the European Way.82 The conference highlighted recent papers on AI
policy from the FRA. The organizers reposted the 2018 FRA report on
discrimination in AI and the 2019 FRA report on facial recognition. One
paper also summarized FRA AI policy initiatives between 2016 and 2020.83
The German Government also provided its comments on the Commission
White Paper on AI84 and the detailed 2019 Opinion of the Data Ethics
Commission concerning algorithm-based decision-making, AI, and data.85
FRA, Artificial Intelligence, Big Data and Fundamental Rights (May 30, 2018),
https://fra.europa.eu/en/project/2018/artificial-intelligence-big-data-and-fundamentalrights
80
FRA, Big Data: Discrimination in data-supported decision-making (May 29, 2018),
https://fra.europa.eu/en/publication/2018/bigdata-discrimination-data-supported-decisionmaking
81
FRA, Facial recognition technology: fundamental rights considerations in the context
of law enforcement (Nov. 27, 2019), https://fra.europa.eu/en/publication/2019/facialrecognition-technology-fundamental-rights-considerations-context-law
82
Doing AI the European way: Protecting Fundamental Rights in an Era of Artificial
Intelligence (Dec. 14, 2020), https://eu2020-bmjv-european-way-on-ai.de/en/
83
Policy initiatives in the area of artificial intelligence (last updated Apr. 29, 2020),
https://eu2020-bmjv-european-way-onai.de/storage/documents/AI_policy_initiatives_(2016-2020).pdf
84
Die Bundesregierung, Comments from the Federal Government of the Federal
Republic of Germanyon the White Paper on Artificial Intelligence - A European Concept
for Excellence and Trust, COM (2020) 65 final, https://eu2020-bmjv-european-way-onai.de/storage/documents/Federal_Government's_Comments_on_the_AI_White_Paper.pdf
85
daten ethik commission, Opinion of the Data Ethics Commission (2019),
https://eu2020-bmjv-european-way-onai.de/storage/documents/Data_Ethics_Commission_Full_Report_in_English.pdf
79
25
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In December 2020, the FRA also issued a report on “Getting the future
right-Artificial intelligence and fundamental rights in the EU.”86
High-Level Expert Group on AI
Following the launch of the Artificial Intelligence Strategy in 2018,
the European Commission appointed a group of 52 experts to advice for its
implementation.87 The group members were selected following an open
selection process and comprised representatives from academia, civil
society and industry. The High-Level Expert Group on Artificial
Intelligence (AI HLEG) has produced four reports: Ethics Guidelines for
Trustworthy AI, Policy and Investment Recommendations for Trustworthy
AI, The final Assessment List for Trustworthy AI, and Sectoral
Considerations on the Policy and Investment Recommendations.
According to the ethical guidelines AI should be 1 lawful — respect
laws and regulations (including the EU Charter on Fundamental Rights, UN
Human Rights Treaties and the Council of Europe Convention on Human
Rights); 2. ethical - respect ethical principles and values and 3. robust —
from a technical perspective and with consideration of its social
environment.88 Since publication in 2019, the ethics guidelines have helped
frame EU policy processes with among others key requirements derived
form the guidelines in the European Commission’s 2021 ”AI Act”
proposal.89
International outreach for human-centric artificial intelligence initiative
In September 2021, The European Commission’s Service for
Foreign Policy Instruments (FPI) and the Directorate General for
Communications Networks, Content and Technology (DG CONNECT), in
collaboration with the European External Action Services (EEAS),
launched the International outreach for human-centric artificial intelligence
initiative (InTouchAI.eu) - a large foreign policy instrument project to
engage with international partners on regulatory and ethical matters and
FRA, Getting the future right- Artificial intelligence and fundamental rights (Dec. 14,
2020) https://fra.europa.eu/en/themes/artificial-intelligence-and-big-data
87
European Commission, High-Level Expert Group on Artificial Intelligence,
https://ec.europa.eu/digital-single-market/en/high-level-expert-group-artificialintelligence
88
European Union, Ethics guidelines for trustworthy AI (Nov. 8, 2019),
https://op.europa.eu/en/publication-detail/-/publication/d3988569-0434-11ea-8c1f01aa75ed71a1
89
European Commission, Proposal for a Regulation of the European Parliament and the
Council, Laying Down Harmonized Rules for Artificial Intelligence (Artificial
Intelligence Act, (Apr. 21, 2021), https://op.europa.eu/en/publication-detail//publication/d3988569-0434-11ea-8c1f-01aa75ed71a1
86
26
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
promote the responsible development of trustworthy AI at global level with
the main vision to ensure that AI “works for people and protects
fundamental rights.”90
G7
The Group of Seven (G7) is an inter-governmental political forum
consisting of Canada, France, Germany, Italy, Japan, the United Kingdom,
and the United States. The members constitute the wealthiest liberal
democracies. The group is officially organized around shared values of
pluralism and representative government. The G7 is also the incubator for
significant work on AI policy.
In advance of the 2016 G7 summit in Japan, then Prime Minister
Shinzo Abe urged his government to develop policies for AI that could
provide the basis for a global standard.91 At the subsequent meeting of G7
ICT ministers, Japan’s Communications Minister proposed international
rules that would make “AI networks controllable by human beings and
respect for human dignity and privacy.”92 She introduced eight basic
principles Japans proposed for AI. These principles are very similar to those
later adopted by the OECD and then the G20.
Prior to the 2018 G7 summit, France and Canada announced a joint
undertaking on Artificial Intelligence that led to the creation of the Global
Partnership on AI.93 According to the Mission Statement of the two
countries, the goal “will be to support and guide the responsible adoption of
AI that is human-centric and grounded in human rights, inclusion, diversity,
innovation and economic growth.”94
In advance of the 2019 G7 summit, hosted by France, leaders of
scientific societies set out a declaration on Artificial Intelligence and
European Commission, International outreach for human-centric artificial intelligence
initiative, https://digital-strategy.ec.europa.eu/en/policies/international-outreach-ai
91
CAIDP Update, Prime Minister Abe’s AI and Data Governance Legacy (Aug. 30,
2020), https://www.japantimes.co.jp/news/2016/04/29/national/japan-pushes-basic-airules-g-7-tech-meeting/
92
Japan Times, Japan pushes for basic AI rules at G-7 tech meeting (Apr. 29, 2016),
https://www.japantimes.co.jp/news/2016/04/29/national/japan-pushes-basic-ai-rules-g-7tech-meeting/
93
France Diplomacy, French-Canadian Declaration on Artificial Intelligence (June 7,
2018), https://www.diplomatie.gouv.fr/en/country-files/canada/events/article/frenchcanadian-declaration-on-artificial-intelligence-07-06-18
94
Canada, Prime Minister of Canada, Mandate for the International Panel on Artificial
Intelligence (Dec. 6, 2018),
https://pm.gc.ca/en/news/backgrounders/2018/12/06/mandate-international-panelartificial-intelligence
90
27
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Society in which they stated, “Artificial intelligence (AI) is one of the
technologies that is transforming our society and many aspects of our daily
lives. AI has already provided many positive benefits and may be a source
of considerable economic prosperity. It also gives rise to questions about
employment, confidentiality of data, privacy, infringement of ethical values
and trust in results.”95
At the 2021 G7 summit hosted by the UK, the G7 Leaders
committed to work together for a “values-driven digital ecosystem for the
common good that enhances prosperity in a way that is sustainable,
inclusive, transparent and human-centric.”96 They called for a “human
centric approach to artificial intelligence,” building on the work of the
Global Partnership for Artificial Intelligence (GPAI) advanced by the
Canadian and French G7 Presidencies in 2018 and 2019.
The G7 Leaders committed to work together for a “values-driven
digital ecosystem for the common good that enhances prosperity in a way
that is sustainable, inclusive, transparent and human-centric.” They called
for a “human centric approach to artificial intelligence,” building on the
work of the Global Partnership for Artificial Intelligence (GPAI) advanced
by the Canadian and French G7 Presidencies in 2018 and 2019, and looking
forward to the GPAI Summit in Paris in November 2021.
At the 2021 G7 privacy officials also issued a statement on Data
Free Flows with Trust.97 Regarding artificial intelligence, the officials said,
“human dignity, must be central to AI design; AI must be transparent,
comprehensible, and explainable; and the data protection principles of
purpose limitation and data minimization must apply to AI.” They further
said that “’red lines’ are needed for AI systems that are not compatible with
our values and fundamental rights.”
G20
The G20 is an international forum, made up of 19 countries and the
European Union, representing the world’s major developed and emerging
economies.98 Together, the G20 members represent 85 % of global GDP,
Summit of the G7 Science Academies, Artificial intelligence and society (Mar. 26,
2019), https://royalsociety.org/-/media/about-us/international/g-science-statements/2019g7-declaration-artificial-intelligence-and-society.pdf
96
The White House, Carbis Bay G7 Summit Communique (June 13, 2021),
https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/13/carbis-bayg7-summit-communique/
97
G7 United Kingdom 2021, Data Free Flows with Trust (Sept. 8, 2021),
https://www.caidp.org/app/download/8342900463/g7-attachment-202109.pdf
98
OECD, What is the G20? https://www.oecd.org/g20/about/
95
28
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
75% of international trade and two-thirds of the world’s population.
According to the OECD, because of its size and strategic importance, the
G20 has a crucial role in setting the path for the future of global economic
growth.
In the last few years, and in collaboration with the OECD, the G20
has taken a leading role in the promulgation of the global framework for AI
policy. At the Osaka summit in 2019, former Prime Minister Abe and
OECD Secretary General Gurria gathered support for the OECD AI
Principles from the G20 countries. The preparatory work for the 2020
summit in Riyadh provided the first opportunity to assess progress toward
implementation of the OECD AI Principles.99
In November 2020, the G20 Leaders Declaration addressed both
Artificial Intelligence and the digital economy. On AI, the G20 nations said,
“We will continue to promote multi-stakeholder discussions to advance
innovation and a human-centered approach to Artificial Intelligence (AI),
taking note of the Examples of National Policies to Advance the G20 AI
Principles. We welcome both the G20 Smart Mobility Practices, as a
contribution to the well-being and resilience of smart cities and
communities, and the G20 Roadmap toward a Common Framework for
Measuring the Digital Economy.”100
On the Digital Economy, the G20 said in 2020, “We acknowledge
that universal, secure, and affordable connectivity, is a fundamental enabler
for the digital economy as well as a catalyst for inclusive growth, innovation
and sustainable development. We acknowledge the importance of data free
flow with trust and cross-border data flows.” The G20 Declaration further
said, “We support fostering an open, fair, and non-discriminatory
environment, and protecting and empowering consumers, while addressing
the challenges related to privacy, data protection, intellectual property
rights, and security.”
The G20 advanced AI policy in the 2021 Leaders’ Declaration,
issued at the conclusion of the Summit in Rome.101 Recognizing the
“benefits stemming from the responsible use and development of
trustworthy human-centered Artificial Intelligence (AI),” the G20 Leaders
said in Rome they would encourage competition and innovation, “as well
99
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
100
G20 Riyadh Summit, Leaders Declaration (Nov. 21-22, 2020),
https://g20.org/en/media/Documents/G20%20Riyadh%20Summit%20Leaders%20Declar
ation_EN.pdf
101
G20 Information Centre, G20 Rome Leaders' Declaration (Oct. 31, 2021),
http://www.g20.utoronto.ca/2021/211031-declaration.html
29
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
as diversity and inclusion,” and the importance of international cooperation
to promote research, development, and application of AI
In advance of the 2021 Summit, the G20 Digital Economy Ministers
reaffirmed “their willingness to implement trustworthy Artificial
Intelligence (AI) and to commit to a human-centered approach, as . . .
guided by the G20 AI Principles, drawn from the OECD Recommendations
on AI.”102 The Ministers also noted that the “measurement of AI, notably
its diffusion and impact across the economy and the international
comparability of indicators on AI, needs to be improved.”
“Privacy and data protection” figured prominently in the 2021 G20
Leaders Statement with multiple references in policies concerning health
and COVID, transportation and travel, the digital economy and higher
education, data free flows with trust, and digital identity tools. The G20
Leaders also prioritized Gender Equality and Women’s Empowerment, a
focus area for AI policy. And the G20 Leaders said they would work in
2022 “towards enhancing confidence in the digital environment by
improving internet safety and countering online abuse, hate speech, online
violence and terrorism while protecting human rights and fundamental
freedoms.”
Global Privacy Assembly
The Global Privacy Assembly is the global network of privacy
officials and experts. The Global Privacy Assembly meets annually to
discuss emerging privacy issues and to adopt resolutions. In recent years,
the focus of the GPA has moved toward AI.103
The GPA adopted a foundational Declaration in 2018 on Ethics and
Data Protection in Artificial Intelligence.104 The 2018 GPA 2018 Resolution
on Ethics in AI emphasized fairness, vigilance, transparency and
intelligibility, and measures to reduce unlawful bias and discrimination.
In 2020, The Assembly adopted a significant Resolution on
Accountability and AI that urged organizations deploying AI systems to
G20 Information Centre, Declaration of G20 Digital Ministers: Leveraging
Digitalisation for a Resilient, Strong, Sustainable and Inclusive Recovery (Aug. 5, 2021),
http://www.g20.utoronto.ca/2021/210805-digital.html
103
CAIPD Update 1.15, Privacy Commissioners Adopt Resolutions on AI, Facial
Recognition (Oct. 19, 2020), https://dukakis.org/center-for-ai-and-digital-policy/caidpupdate-privacy-commissioners-adopt-resolutions-on-ai-facial-recognition/
104
International Conference of Data Protection and Privacy Commissioner, Declaration
on Ethics and Data Protection in Artificial Intelligence (Oct. 23, 2018),
http://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf. See also complete text in Reference section.
102
30
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
implement accountability measures, including a human rights impact
assessment.105 The Privacy Assembly also urged governments to make
changes to data protection law “to make clear the legal obligations regarding
accountability in the development and use of AI.” The 2020 GPA AI
Accountability Resolution builds on a recent a recent GPA survey that
identified accountability measures that are “very important or important for
either AI developers or AI users.” The GPA Resolution reiterated several
key principles for data protection, such as fairness and transparency, but
stopped short of endorsing a formal ban which had been urged by many
human rights advocates at the 2019 conference in Tirana. More than 100
organizations and 1,200 experts recommended that “countries suspend the
further deployment of facial recognition technology for mass surveillance”
and “establish the legal rules, technical standards, and ethical guidelines
necessary to safeguard fundamental rights and comply with legal
obligations before further deployment of this technology occurs.” The
Assembly said it would consider the “circumstances when facial
recognition technology poses the greatest risk to data protection and privacy
rights,” and develop a set of principles that could be adopted at the next
conference.
The OECD
The OECD is an international organization that “works to build
better policies for better lives.”106 The goal of the OECD is to “shape
policies that foster prosperity, equality, opportunity and well-being for all.”
The OECD has led the global effort to develop and establish the
most widely recognized framework for AI policy. This is a result of a
concerted effort by the OECD and the member states to develop a
coordinated international strategy. The OECD AI Principles also build on
earlier OECD initiatives such as the OECD Privacy Guidelines, a widely
recognized framework for transborder data flows and the first global
framework for data protection.107 OECD policy frameworks are not treaties,
do not have legal force, and are not directly applicable to OECD member
Global Privacy Assembly, Adopted Resolution on Accountability in the Development
and Use of Artificial Intelligence (Oct. 2020), https://globalprivacyassembly.org/wpcontent/uploads/2020/10/FINAL-GPA-Resolution-on-Accountability-in-theDevelopment-and-Use-of-AI-EN.pdf. See also complete text in Reference section
106
OECD, Who we are, https://www.oecd.org/about/
107
OECD, OECD Guidelines on the Protection of Privacy and Transborder Flows of
Personal Data (1981),
https://www.oecd.org/sti/ieconomy/oecdguidelinesontheprotectionofprivacyandtransbord
erflowsofpersonaldata.htm
105
31
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
states. However, there are many instances of countries adopting national
laws based on OECD policies, and a clear convergence of legal norms,
particularly in the field of data protection.
Following the publication of the OECD AI Principles in 2019, the
OECD continues extensive work on the adoption and implementation of AI
policies.108
Global Partnership on AI
The Global Partnership on Artificial Intelligence (GPAI) emerged
from the OECD Recommendation on Artificial Intelligence.109 GPAI
activities are intended to foster the responsible development of AI grounded
in “human rights, inclusion, diversity, innovation, and economic
growth.”110 The GPAI aims to “bridge the gap between theory and practice
on AI by supporting cutting-edge research and applied activities on AIrelated priorities.” The GPA developed within the G7 under the Canadian
and French presidencies. As of January 2022, GPAI’s members now include
Australia, Belgium, Brazil, Canada, Czech Republic, Denmark, France,
Germany, India, Ireland, Israel, Italy, Japan, Mexico, the Netherlands, New
Zealand, Poland, the Republic of Korea, Singapore, Slovenia, Spain,
Sweden, the United Kingdom, the United States, and the European Union.
The GPAI held the Montreal Summit in early 2020.111 The five key
themes at the first GPAI meeting were the Responsible Use of AI, Data
Governance, The Future of Work, AI and the Pandemic Response,
Innovation, and Commercialization. The organizers of the Montreal
Summit included an AI Art Session to learn how AI will “advance art
artistry.”
OECD AI Observatory
The OECD AI Observatory, launched in February 2020, provides
extensive data and multi-disciplinary analysis on artificial intelligence
across a wide range of policy areas.112 According to the OECD, the AI
Policy Observatory is based on multidisciplinary, evidence-based analysis,
and Global multi-stakeholder partnerships.
CAIP Update 1.13, OECD Report Examines Implementation of AI Principles (Oct. 5,
2020), https://dukakis.org/center-for-ai-and-digital-policy/caidp-update-oecd-reportexamines-implementation-of-ai-principles/
109
GPAI, The Global Partnership on Artificial Intelligence, https://gpai.ai
110
GPAI, About GPAI, https://gpai.ai/about/
111
GPAI, Montreal Summit 2020, https://www.c2montreal.com/en/lp/global-partnershipon-artificial-intelligence/
112
OECD, AI Policy Observatory, https://www.oecd.ai/
108
32
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
National Implementation
The OECD also published the first report that attempts to assess the
implementation of the OECD AI Principles among the G-20 nations in
2020.113 Examples of AI National Policies surveys “rationales and
illustrative actions” for the 10 principles that make up the OECD/G-20
Guidelines on AI policy. The report was prepared by the G20 Digital
Economy Task Force. Key observations from the Task Force report:
• G20 countries are moving quickly to build trustworthy AI
ecosystems, though most initiatives are very recent
• Many national AI strategies address multiple G20 AI Principles
simultaneously, which the OECD contends reinforce the strong
complementarity of the Principles
• So far, few national policies emphasize Principles of
robustness, security and safety, and accountability,
• Many national policies emphasize R&D, fostering a digital
ecosystem, human capacity, and international cooperation
The Task Force also found that “there is potential for steering public
research towards socially oriented applications and issues, and for
leveraging R&D activities to make progress on issues such as
accountability, explainability, fairness and transparency.” The Task Force
emphasized that there “is currently a critical window for G20 members to
continue their leadership on AI policy issues and to promote
implementation of the G20 AI Principles. Development, diffusion and use
of AI technologies are still at a relatively early level of maturity across many
countries and firms, and policy-making on AI is in an active experimental
phase.”114
A second report on implementation was published in 2021.115 The
report builds both on the expert input provided at meetings of the OECD.AI
Network of Experts working group on national AI policies that took place
online from February 2020 to April 2021 and on the EC-OECD database of
national AI strategies and policies. The expert group leveraged the OECD
AI Policy Observatory www.oecd.ai (OECD.AI), containing a database of
national AI policies from OECD countries and partner economies and the
EU. These resources help policy makers keep track of national initiatives to
CAIP Update 1.13, OECD Report Examines Implementation of AI Principles (Oct. 5,
2020), https://dukakis.org/center-for-ai-and-digital-policy/caidp-update-oecd-reportexamines-implementation-of-ai-principles/
114
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
115
OECD, State of Implementation of the OECD AI Principles: Insights from National AI
Polices (June 2021), https://oecd.ai/en/policies
113
33
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
implement the recommendations to governments contained in the OECD AI
Principles. National policy makers are the primary audience for this report.
The expert group met monthly between June 2020 and March 2021 to
discuss case studies from selected countries during 90-minute online
meetings. Over this period, 24 case studies were discussed during ten virtual
meetings. These discussions provided “deep dives” into national
experiences in implementing AI policies and were rich in lessons learned
and good practices identified for each phase of the AI policy cycle.
OECD Secretary General Angel Gurria remarks at the 2020 G-20
Digital Economy Ministers Meeting in Riyadh also provide insight into the
work of the OECD on AI.116 Secretary Gurria, addressing the global
challenges of the COVID-19 crisis, urged countries to “use digital
technologies to build our economies back in a better way: more resilient,
inclusive and sustainable.” He also spoke about the need to bridge the digital
divide, to shift to smart mobility practices, and to continue work on
measurement of the digital economy.
“As this year’s G20 AI Dialogue showed,” Secretary Gurria said in
2020, “AI’s full potential is still to come. To achieve this potential, we must
advance a human-centred and trustworthy AI, that respects the rule of law,
human rights, democratic values and diversity, and that includes appropriate
safeguards to ensure a fair and just society. This AI is consistent with the
G20 AI Principles you designed and endorsed last year, drawing from the
OECD’s AI Principles.”
The OECD ONE PAI
The OECD has also established a Working Group on Policies for AI
(ONE PAI).117 The Working Group is developing practical guidance for
policymakers on a wide array of topics: investing in AI R&D; data,
infrastructure, software & knowledge; regulation, testbeds and
documentation; skills and labor markets; and international co-operation.
The ONE PAI leverages lessons learned by other OECD bodies, as
well as analysis of national AI policies. The working group is focusing on
the practical implementation of the OECD AI Principles throughout the AI
policy cycle for:
• Policy design – focusing on national AI governance policies
and approaches;
CAIP Update 1.2, OECD’s Gurria Underscores AI Fairness at G-20 (July 26, 2020),
https://dukakis.org/center-for-ai-and-digital-policy/center-for-ai-policy-update-oecdsgurria-underscores-ai-fairness-at-g-20-meeting/
117
OECD, OECD Network of Experts on AI (ONE AI), https://oecd.ai/network-ofexperts
116
34
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Policy implementation – focusing on lessons learned to date
through national implementation examples;
• Policy intelligence – identifying different evaluation methods
and monitoring exercises; and
• Approaches for international and multi-stakeholder cooperation on AI policy.
The OECD ONE PAI held five virtual meetings between June and
September 2020 which provided “deep dives” into national experience in
implementing AI policies in practice.
•
United Nations
The United Nations launched work on AI in 2015 with the General
Assembly event Rising to the Challenges of International Security and the
Emergence of Artificial Intelligence.118 In 2015, the UN Interregional
Crime and Justice Research Institute (UNICRI) launched a program on AI
and Robotics.
The Secretary General
In its 2020 Roadmap for Digital Cooperation, the UN Secretary
General stated that “Digital technologies provide new means to advocate,
defend and exercise human rights, but they can also be used to suppress,
limit and violate human rights," noting with emphasis lethal autonomous
weapons and facial recognition.119 He also announced the creation of an
advisory body on global artificial intelligence cooperation to provide
guidance to the Secretary General and the international community on
artificial intelligence that is trustworthy, human-rights based, safe and
sustainable and promotes peace. The advisory body will comprise Member
States, relevant United Nations entities, interested companies, academic
institutions, and civil society groups.
The Roadmap echoes the UN Secretary General 2018 Strategy on
New Technologies whose goal was to "define how the United Nations
system will support the use of these technologies to accelerate the
UNICRI, Rising to the Challenges of International Security and the Emergence of
Artificial Intelligence (Oct. 7, 2015),
http://www.unicri.it/news/article/cbrn_artificial_intelligence
119
UN Secretary General, Report - Roadmap for Digital Cooperation (June 2020,
https://www.un.org/en/content/digital-cooperationroadmap/assets/pdf/Roadmap_for_Digital_Cooperation_EN.pdf); see also UN Secretary
General, The Highest Aspiration - A Call to Action for Human
Rights (2020) https://www.un.org/sg/sites/www.un.org.sg/files/atoms/files/The_Highest_
Asperation_A_Call_To_Action_For_Human_Right_English.pdf
118
35
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
achievement of the 2030 Sustainable Development Agenda and to facilitate
their alignment with the values enshrined in the UN Charter, the Universal
Declaration of Human Rights and the norms and standards of International
Laws" with the first principle: "Protect and Promote Global Values" and the
second principle: "Foster inclusion and transparency."120
In a 2021 report Our Common Agenda, the UN
Secretary General also proposed the creation of a Digital Global
Compact which could "promote regulation of artificial intelligence to
ensure that this is aligned with shared global values." The
Compact would be agreed on during a Summit of the Future, prepared in
part by "a multi-stakeholder digital technology track."121
On January 26, 2022, Maria-Francesca Spatolisano was designated
as the Acting UN Envoy on Technology. She is in charge of coordinating
the implementation of the Secretary-General’s Roadmap on Digital
Cooperation and advancing work towards the Global Digital Compact
proposed in the Common Agenda, in close consultation with Member
States, the technology industry, private companies, civil society, and other
stakeholders.122
In
December
2021, Secretary-General
Antonio
Guterres encouraged the Review Conference of the U.N.’s Convention on
Certain Conventional Weapons "to agree on an ambitious plan for the future
to establish restrictions on the use of certain types of autonomous
weapons."123 This follows his call for an international legal ban on LAWS
which he qualified in a 2019 message to Meeting of the Group of
Governmental Experts on Emerging Technologies in the Area of Lethal
Autonomous Weapons Systems as "politically unacceptable, morally
repugnant."124
UN Secretary General, Strategy on New Technologies (Sept.
2018, https://www.un.org/en/newtechnologies/images/pdf/SGs-Strategy-on-NewTechnologies.pdf)
121
UN Secretary General, Report: Our Common Agenda
(2021), https://www.un.org/en/content/common-agendareport/assets/pdf/Common_Agenda_Report_English.pdf).
122
https://www.un.org/techenvoy/content/about
123
United Nation, Secretary-General's message to the Sixth Review Conference of High
Contracting Parties to the Convention on Certain Conventional Weapons (Dec. 13,
2021) https://www.un.org/sg/en/node/261134
124
United Nations, Secretary-General's message to Meeting of the Group of
Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous
Weapons Systems (March 25, 2019) https://www.un.org/sg/en/content/sg/statement/201903-25/secretary-generals-message-meeting-of-the-group-of-governmental-expertsemerging-technologies-the-area-of-lethal-autonomous-weapons-systems
120
36
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
UNESCO Recommendation on AI Ethics
In 2020 UNESCO embarked on a two-year project to develop a
global standard for Artificial Intelligence. UNESCO Director General
Audrey Azoulay stated, "Artificial intelligence can be a great opportunity
to accelerate the achievement of sustainable development goals. But any
technological revolution leads to new imbalances that we must
anticipate.”125
In 2020 UNESCO published a draft Recommendation on the Ethics
of Artificial Intelligence. UNESCO stated that the Recommendation “aims
for the formulation of ethical values, principles and policy
recommendations for the research, design, development, deployment and
usage of AI, to make AI systems work for the good of humanity,
individuals, societies, and the environment." The UNESCO draft
Recommendation sets out about a dozen principles, five Action Goals, and
eleven Policy Actions. Notable among the UNESCO recommendations is
the emphasis on Human Dignity, Inclusion, and Diversity. UNESCO also
expresses support for Human Oversight, Privacy, Fairness, Transparency
and Explainability, Safety and Security, among other goals.
Understandably, UNESCO is interested in the scientific, educational, and
cultural dimensions of AI, the agency’s program focus.
The UNESCO Recommendation was adopted on November 24,
2021, at the 41st General Conference at its 41st session. This is the first
global agreement on the Ethics of Artificial Intelligence.126 UNESCO
Director General Audrey Azoulay stated, "The world needs rules for
artificial intelligence to benefit humanity. The recommendation on the
ethics of AI is a major answer. It sets the first global normative framework
while giving member states the responsibility to apply it at their level.
UNESCO will support its 193 member states in its implementation and ask
them to report regularly on their progress and practices.”
The UNESCO Recommendation was the outcome of a multi-year
process and was drafted with the assistance of more than 24 experts.127
According to UNESCO, the “historical text defines the common values and
principles which will guide the construction of the necessary legal
UNESCO, Artificial intelligence with human values for sustainable development,
https://en.unesco.org/artificial-intelligence
126
UNESCO, Recommendation on the Ethics of Artificial Intelligence (2021),
https://unesdoc.unesco.org/ark:/48223/pf0000380455
127
UNESCO, Preparation of a draft text of the Recommendation: Ad Hoc Expert Group,
https://en.unesco.org/artificial-intelligence/ethics#aheg
125
37
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
infrastructure to ensure the healthy development of AI.”128 UNESCO
explained, “The Recommendation aims to realize the advantages AI brings
to society and reduce the risks it entails. It ensures that digital
transformations promote human rights and contribute to the achievement of
the Sustainable Development Goals, addressing issues around transparency,
accountability and privacy, with action-oriented policy chapters on data
governance, education, culture, labour, healthcare and the economy.” The
key achievements of the UNESCO AI Recommendation include:
1. Protecting data. The UNESCO Recommendation calls for action
beyond what tech firms and governments are doing to guarantee
individuals more protection by ensuring transparency, agency and
control over their personal data.
2. Banning social scoring and mass surveillance. The UNESCO
Recommendation explicitly bans the use of AI systems for social
scoring and mass surveillance.
3. Monitoring and Evaluation. The UNESCO Recommendation
establishes new tools that will assist in implementation, including
Ethical Impact Assessments and a Readiness Assessment
Methodology.
4. Protecting the environment. The UNESCO Recommendation
emphasizes that AI actors should favor data, energy and resourceefficient AI methods that will help ensure that AI becomes a more
prominent tool in the fight against climate change and on tackling
environmental issues.
The Recommendation aims to provide a basis to make AI systems
work for the good of humanity, individuals, societies and the environment
and ecosystems, and to prevent harm. It also aims at stimulating the peaceful
use of AI systems. The Recommendation provides a universal framework
of values and principles of the ethics of AI. It sets out four values: respect,
protection and promotion of human rights and fundamental freedoms and
human dignity; environment and ecosystem flourishing; ensuring diversity
and inclusiveness; living in peaceful, just and interconnected societies.
UNESCO, UNESCO member states adopt the first ever global agreement on the
Ethics of Artificial Intelligence (Nov. 25, 2021), https://en.unesco.org/news/unescomember-states-adopt-first-ever-global-agreement-ethics-artificial-intelligence
128
38
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Further, the Recommendation outlines 10 principles –
proportionality and do no harm, safety and security, fairness and nondiscrimination, sustainability, right to privacy and data protection, human
oversight and determination, transparency and explainability, responsibility
and accountability, awareness and literacy – backed up by more concrete
policy actions on how they can be achieved. The Recommendation also
introduces red-lines to unacceptable AI practices. For example, it states that
“AI systems should not be used for social scoring or mass surveillance
purposes.”
The Recommendation focuses not only on values and principles, but
also on their practical realization, via concrete eleven policy actions. It
encourages Member States to introduce frameworks for ethical impact
assessments, oversight mechanisms etc. Member States should ensure that
harms caused through AI systems are investigated and redressed, by
enacting strong enforcement mechanisms and remedial actions, to make
certain that human rights and fundamental freedoms and the rule of law are
respected.
UN High Commissioner for Human Rights
In the Roadmap for Digital Cooperation, the Secretary General
stated, "To address the challenges and opportunities of protecting and
advancing human rights, human dignity and human agency in a digitally
interdependent age, the Office of the United Nations High Commissioner
for Human Rights will develop system-wide guidance on human rights due
diligence and impact assessments in the use of new technologies, including
through engagement with civil society, external experts and those most
vulnerable and affected."129
In September 2021, the UN High Commissioner for Human
Rights Michelle Bachelet called for a moratorium on the sale and use of AI
that pose a serious risk to human rights until adequate safeguards are put in
place.130 She also called for a ban on AI applications that do not comply
with international human rights law. “Artificial intelligence can be a force
for good, helping societies overcome some of the great challenges of our
UN Secretary General, Report - Roadmap for Digital Cooperation (June
2020) https://www.un.org/en/content/digital-cooperationroadmap/assets/pdf/Roadmap_for_Digital_Cooperation_EN.pdf)
130
UN Human Rights, Office of the High Commissioner, Artificial intelligence risks to
privacy demand urgent action – Bachelet (Sept. 15, 2021),
https://www.ohchr.org/EN/NewsEvents/Pages/DisplayNews.aspx?NewsID=27469&Lang
ID=E; see also UN Urges Moratorium on AI that Violates Human Rights, CAIDP Update
2.34 (Sept. 15, 2021), https://www.caidp.org/app/download/8343909663/CAIDP-Update2.34.pdf
129
39
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
times. But AI technologies can have negative, even catastrophic, effects if
they are used without sufficient regard to how they affect people’s human
rights,” Bachelet said.
The High Commissioner’s statement accompanied the release of
a new report on The Right to Privacy in the Digital Age. The UN Report
details how AI systems rely on large data sets, with information about
individuals collected, shared, merged and analyzed in multiple and
often opaque ways. The UN Report finds that data used to guide AI systems
can be faulty, discriminatory, out of date or irrelevant. Long-term storage
of data also poses particular risks, as data could in the future be exploited in
as yet unknown ways.131
International Telecommunications Union
In 2017 and 2018, the International Telecommunications Union
(ITU) organized the AI for Good Global Summits, “the leading United
Nations platform for dialogue on AI.”132 Houlin Zhao, Secretary General of
the ITU stated, “As the UN specialized agency for information and
communication technologies, ITU is well placed to guide AI innovation
towards the achievement of the UN Sustainable Development Goals. We
are providing a neutral platform for international dialogue aimed at building
a common understanding of the capabilities of emerging AI technologies.”
The 2018 ITU report Artificial Intelligence for global good focused on the
relationship between AI and progress towards the United Nations’
Sustainable Development Goals (SDGs).133
UN Special Rapporteur
An extensive 2018 report by a UN Special Rapporteur explored the
implications of artificial intelligence technologies for human rights in the
information environment, focusing in particular on rights to freedom of
opinion and expression, privacy and non-discrimination.134 The Report of
the Special Rapporteur on the promotion and protection of the right to
Human Rights Council, The right to privacy in the digital age, Report of the United
Nations High Commissioner for Human Rights (Sept. 13, 2021),
https://www.ohchr.org/EN/HRBodies/HRC/RegularSessions/Session48/Documents/A_H
RC_48_31_AdvanceEditedVersion.docx
132
ITU, AI for Good Global Summit 2018, https://www.itu.int/en/ITUT/AI/2018/Pages/default.aspx
133
ITU News Magazine, Artificial Intelligence for global good (Jan. 2018),
https://www.itu.int/en/itunews/Documents/2018/2018-01/2018_ITUNews01-en.pdf
134
UN Special Rapporteur, Report of the Special Rapporteur on the promotion and
protection of the right to freedom of opinion and expression, A/73/348 (Aug. 29, 2018),
https://freedex.org/wp-content/blogs.dir/2015/files/2018/10/AI-and-FOE-GA.pdf
131
40
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
freedom of opinion and expression report defines key terms “essential to a
human rights discussion about artificial intelligence”; identifies the human
rights legal framework relevant to artificial intelligence; and presents
preliminary to ensure that human rights are considered as AI systems
evolve. The report emphasizes free expression concerns and notes several
frameworks, including the International Covenant on Civil and Political
Rights and the UN Guiding Principles on Business and Human Rights.
Among the Recommendations, the Special Rapporteur proposed
“Companies should make all artificial intelligence code fully auditable and
should pursue innovative means for enabling external and independent
auditing of artificial intelligence systems, separately from regulatory
requirements. The results of artificial intelligence audits should themselves
be made public.” The report emphasizes the need for transparency in the
administration of public services. “When an artificial intelligence
application is being used by a public sector agency, refusal on the part of
the vendor to be transparent about the operation of the system would be
incompatible with the public body’s own accountability obligations,” the
report advises.
UN and Lethal Autonomous Weapons
One of the first AI applications to focus the attention of global
policymakers was the use of AI for warfare.135 In 2016, the United Nations
established the Group of Governmental Experts (GGE) on Lethal
Autonomous Weapons Systems (LAWS) following a review of the High
Contracting Parties to the Convention on Certain Conventional Weapons
(CCW).136 In November 2019,137 the CCW High Contracting Parties
endorsed 11 Guiding Principles for LAWS.138 But concerns about future of
The Computer Professionals for Social Responsibility (CPSR), a network of computer
scientists based in Palo Alto, California, undertook early work on this topic in the 1980s.
CPSR History, http://cpsr.org/about/history/. See also David Bellin and Gary Chapman,
Computers in Battle Will They Work? (1987).
136
United Nations, 2018 Group of Governmental Experts on Lethal Autonomous
Weapons Systems (LAWS),
https://www.unog.ch/80256EE600585943/(httpPages)/7C335E71DFCB29D1C12582430
03E8724
137
Meeting of the High Contracting Parties to the Convention on Prohibitions or
Restrictions on the Use of Certain Conventional Weapons Which May Be Deemed to Be
Excessively Injurious or to Have Indiscriminate Effects, Final Report (Dec. 13, 2019),
https://undocs.org/Home/Mobile?FinalSymbol=CCW%2FMSP%2F2019%2F9&Languag
e=E&DeviceType=Desktop
138
Group of Governmental Experts on Emerging Technologies in the Area of Lethal
Autonomous Weapons System, Report of the 2019 session of the Group of Governmental
Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons Systems
135
41
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
regulation of lethal autonomous weapons remain. At present, some
countries believe that current international law “mostly suffices” while
others believe new laws are needed.139 Human Rights Watch provided an
important overview of country positions on the future of banning fully
autonomous weapons in August 2020.140 Concerns over killer reports also
arose at the 75th UN Assembly in October 2020.141 Pope
Francis warned that lethal autonomous weapons systems would
“irreversibly alter the nature of warfare, detaching it further from human
agency.” He called on states to “break with the present climate of distrust”
that is leading to “an erosion of multilateralism, which is all the more
serious in light of the development of new forms of military technology.”142
The Permanent Representative of the Holy See to the UN called for a ban
on autonomous weapons in 2014.143
The Vatican
Pope Francis has emerged as a leading figure the world of AI policy.
In addition to his statements on autonomous weapons, in November 2020
the Pope warned that AI could exacerbate economic inequalities around the
world if a common good is not pursued. “Artificial intelligence is at the
(Sept. 25, 2019), https://documents-ddsny.un.org/doc/UNDOC/GEN/G19/285/69/PDF/G1928569.pdf?OpenElement
139
Dustin Lewis, An Enduring Impasse on Autonomous Weapons, Just Security (Sept. 28,
2020), https://www.justsecurity.org/72610/an-enduring-impasse-on-autonomousweapons/
140
Human Rights Watch, Stopping Killer Robots: Country Positions on Banning Fully
Autonomous Weapons and Retaining Human Control (Aug. 10, 2020),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#
141
Stop Killer Robots, 75th UN Assembly (Oct. 30, 2020),
https://www.stopkillerrobots.org/2020/10/un-diplomacy/
142
Address of His Holiness Pope Francis to the Seventy-fifth Meeting of the General
Assembly of the United Nations, The Future We Want, the United Nations We Need:
Reaffirming our Joint Commitment through Multilateralism (Sept. 25, 2020),
https://reachingcriticalwill.org/images/documents/Disarmamentfora/unga/2020/25Sept_HolySee.pdf
143
Statement by H.E. Archibishop Silvano M. Tomasi, Permanent Representative of the
Holy See to the United Nations and Other International Organizations in Geneva at the
meeting of Experts on Lethal Autonomous weapons systems of the High Contracting
Parties to the Convention, On Prohibitions or Restrictions on the Use of Certain
Conventional Weapons which may be Deemed to be Excessively Injurious or to have
Indiscriminate Effect (May 13, 2014),
https://www.unog.ch/80256EDD006B8954/(httpAssets)/D51A968CB2A8D115C1257C
D8002552F5/$file/Holy+See+MX+LAWS.pdf
42
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
heart of the epochal change we are experiencing. Robotics can make a better
world possible if it is joined to the common good. Indeed, if technological
progress increases inequalities, it is not true progress. Future advances
should be oriented towards respecting the dignity of the person and of
Creation.”144
Earlier in 2020, the Pope endorsed the Rome Call for AI Ethics.145
The goal of the Rome Call is to “support an ethical approach to Artificial
Intelligence and promote a sense of responsibility among organizations,
governments and institutions.” The Pope said, "The Call’s intention is to
create a movement that will widen and involve other players: public
institutions, NGOs, industries and groups to set a course for developing and
using technologies derived from AI.” The Pope also said that the Rome Call
for Ethics is the “first attempt to formulate a set of ethical criteria with
common reference points and values, offering a contribution to the
development of a common language to interpret what is human.”146
The key principles of the Rome Call are 1) Transparency: AI
systems must be explainable; 2) Inclusion: the needs of all human beings
must be taken into consideration so that everyone can benefit and all
individuals can be offered the best possible conditions to express
themselves and develop; 3) Responsibility: those who design and deploy
the use of AI must proceed with responsibility and transparency; 4)
Impartiality: do not create or act according to bias, thus safeguarding
fairness and human dignity; 5) Reliability: AI systems must be able to work
reliably; 6) Security and privacy: AI systems must work securely and
respect the privacy of users. These principles are described as “fundamental
elements of good innovation.”
Technical Societies
Technical societies have also played a leading role in the articulation
of AI principles. The IEEE led several initiatives, often in cooperation with
government policymakers, to develop and promote Ethically Aligned
Design (EAD).147 The initial report A Vision for Prioritizing Human Wellbeing with Autonomous and Intelligent Systems was published in 2015. The
Vatican News, Pope’s November prayer intention: that progress in robotics and AI
“be human” (Nov. 2020), https://www.vaticannews.va/en/pope/news/2020-11/popefrancis-november-prayer-intention-robotics-ai-human.html
145
Rome Call AI Ethics, https://romecall.org
146
Pontifical Academy for Life, Rome Call for Ethics (Feb. 28, 2020),
http://www.academyforlife.va/content/pav/en/events/intelligenza-artificiale.html
147
IEEE Ethics in Action in Autonomous and Intelligent Systems,
https://ethicsinaction.ieee.org
144
43
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
IEEE published the second edition in 2017.148 In 2019 the IEEE issued a
Positions Statement on Artificial Intelligence, concluding that “AI systems
hold great promise to benefit society, but also present serious social, legal
and ethical challenges, with corresponding new requirements to address
issues of systemic risk, diminishing trust, privacy challenges and issues of
data transparency, ownership and agency.”149
ACM, an international society of computer scientists and
professionals, has also contributed to the global AI policy landscape.150 In
2017 ACM released a Statement on Algorithmic Transparency and
Accountability, identifying key principles to minimize bias and risks in
algorithmic
decision-making
systems,
including
transparency,
accountability, explainability, auditability, and validation.151 In 2020, in
response to growing concerns about the use of facial recognition
technologies in public spaces, ACM released another statement addressing
the unique issues of biometric data systems and the potential bias and
inaccuracies that have significant consequences for violation of human
rights.152
Civil Society
Latin America
In Latin America, NGOs have been active in AI-related aspects,
particularly in connection with the use of facial recognition technology. In
Argentina, the Association for Civil Rights (Asociación por los Derechos
Civiles), a very-well known Argentinian human rights organization has
criticized the increasing and unaccountable use of facial recognition
technology. These efforts have led to the creation of a national campaign
using the slogan ‘“Con mi Cara No” (“No with my face”). The organization
IEEE Standards Association, IEEE Releases Ethically Aligned Design, Version 2 to
show "Ethics in Action" for the Development of Autonomous and Intelligent Systems
(A/IS) (Dec. 12, 2017), https://standards.ieee.org/news/2017/ead_v2.html
149
IEEE, Artificial Intelligence (June 24, 2019), https://globalpolicy.ieee.org/wpcontent/uploads/2019/06/IEEE18029.pdf
150
Association for Computing Machinery, www.acm.org/public-policy
151
ACM, US Public Policy Council, Statement on Algorithmic Transparency and
Accountability, (Jan. 12, 2017),
https://www.acm.org/binaries/content/assets/publicpolicy/2017_usacm_statement_algorit
hms.pdf.
152
ACM, US Technology Policy Committee, Statement on Principles and Prerequisites
for the Development, Evaluation and Use of Unbiased Facial Recognition Technologies
(June 30, 3030), https://www.acm.org/binaries/content/assets/public-policy/ustpc-facialrecognition-tech-statement.pdf
148
44
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
aims to raise awareness about the dangers of facial recognition technology,
particularly when their data is included within opaque and unaccountable
systems.153 Furthermore, during 2020, the Association made contributions
to Future City: AI Strategy (Ciudad Futuro: Plan Estratégico Inteligencia
Artificial) of the Autonomous City of Buenos Aires.154 It also participates
in the Trustworthy Artificial Intelligence (TAI) program organized by
Mozfest, and the working group “Making use of the Civic Voice in AI
Impact Assessment” with more than thirty members of different civil
society organizations around the world.
The Igarape Institute, an independent Brazilian think tank, also
publishes AI-related research: in 2019, the Institute published a study on
Future Crime providing an overview of the opportunities and pitfalls of new
technologies to fight crime and stated recommendations to ensure
transparency and accountability.155 The emphasis was on predictive
analytics and the Institute recommended that enforcement agencies are
informed about the challenges and caveats associate applying these new
crime prediction platforms. The principles of transparency and
accountability were also highlighted, as well as the need to ensure the
safety, dignity and rights of people in the crime forecasting process,
including when advanced software packages are deployed. Predictive tools
need not replace the intuition and experience of law enforcement officers,
but rather complement them in an agile and auditable manner.
Furthermore, in relation to the São Paulo Metro operator,
ViaQuatro, that installed and used an AI crowd analytics system that claims
to predict the emotion, age, and gender of metro passengers without
processing personal data, Access Now filed an expert opinion criticizing
this initiative.156
Fundoción Karisma, another civil society organization dedicated to
supporting the responsible use of tech highlights the pitfalls of these
systems. In their report titled Discreet Cameras, they point out that
surveillance technology and biometric identification systems in Colombia
only take into consideration the technical and impact considerations while
assessing systems. There is no analysis using necessity, proportionality or
Asociación por los Derechos Civiles, https://conmicarano.adc.org.ar/
Asociación por los Derechos Civiles, Yearbook 20021 https://adc.org.ar/wpcontent/uploads/2022/01/ADC-Yearbook-2021.pdf
155
Igarape Institute, Future Crime - Assessing twenty first century crime prediction (Feb.
3, 2019),
https://igarape.org.br/en/future-crime-assessing-twenty-first-century-crime-prediction/
156
Brazilian Institute of Consumer Protection, Autos no.: 1090663-42.2018.8.26.0100,
https://www.accessnow.org/cms/assets/uploads/2020/06/Expert_Opinion_Brazil_Facial_
Categorization.pdf
153
154
45
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the possible effect of the technology on human rights. Although the
government tries to ensure transparency by sharing the location of video
surveillance systems that use facial recognition technology, the right to
privacy and other fundamental rights of individuals are still ignored.157
In addition, when Uruguay began developing a facial identification
database some civil society organizations warned that “this system was
approved using the National Budget Act as an ‘omnibus law,’ thus
preventing proper discussion about the issue due to the tight deadlines for
approval of this type of law.158
More broadly, several civil society organizations under the banner
“Al Sur” in Latin America that seeks to strengthen human rights in the
digital environment responded to the public consultation on “Ethics and
Data Protection in Artificial Intelligence: continuing the debate” promoted
by the International Conference of Data Protection and Privacy
Commissioners (ICDPPC).159
Africa
In relation to Africa, research shows more limited engagement with
AI-related questions. In relation to Nigeria, Paradigm Initiative, which
operates regional offices in Cameroon, Kenya, Nigeria, Senegal, Zambia,
and Zimbabwe, has observed that Nigeria conducts surveillance activities
without judicial oversight and a comprehensive framework for data
protection and recommended the enactment of a comprehensive framework
for data protection and privacy and judicial oversight over surveillance.160
With regard to AI, Paradigm Initiative has published policy briefs and
factsheets, providing a series of recommendations, namely: assessment of
Nigeria’s strategic priorities, strengths and weaknesses, alignment with
supranational AI standards, concerns regarding the use of AI in certain
sectors, such as law enforcement, criminal justice, immigration and national
Fundación Karisma, Discreet Cameras, (Feb. 2, 2018),
https://web.karisma.org.co/camaras-indiscretas/
158
DATYSOC, Organizaciones de la sociedad civil y académicas expresan su
preocupación por reconocimiento facial en el Proyecto de Ley de Presupuesto de
Uruguay (Nov. 17, 2020), https://datysoc.org/2020/11/17/organizaciones-de-la-sociedadcivil-y-academicas-expresan-su-preocupacion-por-reconocimiento-facial-en-el-proyectode-ley-de-presupuesto-de-uruguay/
159
Ethics and Data Protection in Artificial Intelligence: continuing the debate. A
contribution from Latin America & the Caribbean, https://web.karisma.org.co/ethics-anddata-protection-in-artificial-intelligence-continuing-the-debate-a-contribution-from-latinamerica-the-caribbean/
160
https://paradigmhq.org/wp-content/uploads/2021/05/Digital-Rights-and-Privacy-inNigeria_0.pdf
157
46
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
security; a human-centric approach to data governance; reinforcing the
responsibility of the Nigerian State to protect citizens human rights, and the
responsibility of businesses to respect these rights; prioritizing local AI and
ensuring a transparent procurement process for AI systems from abroad:
and calling for AI upskilling and reskilling.161
Paradigm Initiative has also published a policy brief on the AI policy
of Kenya highlighting the challenges faced in the adoption of AI systems,
which include the lack of relevant data for the development of the systems,
lack of regulatory framework governing the AI ecosystem in the country,
lack of relevant AI skills, connectivity divide in the country, and the lack of
investment in research on development of AI systems and protection of
human rights.162 Paradigm Initiative also stressed the risks posed by the use
of AI systems on human rights, focusing not only on bias caused by the
systems, but also the weaponization of AI systems by the Government
which may undermine freedom of expression and association, surveillance
through the use of facial recognition technologies, and violation of rights
through contents moderation.
In 2019, Witness and the Centre for Human Rights at the University
of Pretoria, hosted an expert meeting on deepfakes and other forms of AIenabled synthetic media.163 The Centre for Human Rights also launched the
#Tech4Rights initiative to, among several purposes, build stronger regional
partnerships for advocacy on the effective use of digital technologies for
human rights protection.164
The African Internet Rights Alliance (AIRA) is made up of nine
civil society organizations based in countries across Central, East, Southern
and West Africa.165 The work of AIRA is rooted in four values:
accountability, transparency, integrity, and good governance. Using these
values as a guide, AIRA undertakes collective interventions and executes
strategic campaigns that engage the government, private sector, media and
Paradigm Initiative, Towards A Rights-Respecting Artificial Intelligence Policy for
Nigeria,
(November 2021), https://paradigmhq.org/wp-content/uploads/2021/11/Towards-ARights-Respecting-Artificial-Intelligence-Policy-for-Nigeria.pdf
162
Paradigm Initiative, Artificial Intelligence in Kenya, (January 2022),
https://paradigmhq.org/wp-content/uploads/2022/02/Artificial-Inteligence-in-Kenya1.pdf
163
Centre for Human Rights and Witness Host Africa’s first ‘deepfakes’ workshop in
Pretoria (Nov. 28, 2019), https://www.chr.up.ac.za/news-archive/2019/1929-witnessand-centre-for-human-rights-host-africa-s-first-deepfakes-workshop-in-pretoria
164
Centre for Human Rights, #Tech4Rights: Rethinking a human rights-based approach
to new technologies in Africa (Oct. 26, 2021), https://www.chr.up.ac.za/tech4rights
165
Africa Internet Rights Alliance, About Us, https://aira.africa/about-us/
161
47
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
civil society to institute and safeguard digital rights. In February 2022, the
Alliance hosted a seminar on “Artificial Intelligence in Africa:
Opportunities, Challenges, and Ethical Imperatives.”
Furthermore, the Digital Transformation Center, a GermanRwandan innovation hub, among other tasks, organises events about current
ICT topics and trends, organizes training and capacity-development, as well
as networking opportunities.166
Moreover, the Rwandan government has engaged Future Society,
an independent think tank, to support the development of Rwanda’s national
artificial intelligence strategy, along with AI ethical guidelines, and a
practical implementation strategy fit for the local context.167 In 2021, the
Future Society also organised workshops for employees working specific
banks with branches in Africa regarding the concept of responsible AI,
existing corporate guidelines, the ethical challenges raised by the use of
algorithmic prediction for credit lending, and potential impact of facial
recognition technologies (FRT) in the banking sector.168 The Future Society
has also published a briefing about the opportunities and challenges of AI
in Healthcare in Africa, based on research conducted in TFS’ Responsible
AI for Development (RAI4D) program.169
Asia
In China, the Beijing Academy of Artificial Intelligence (BAAI) is
a non-profit research institute aimed at promoting collaboration among
academia and industries, as well as fostering top talents and a focus on longterm research on the fundamentals of AI technology. In 2019, the BAAI
released the Beijing AI Principles for the research and development, use,
and governance of AI.170
For example see Luisa Olaya Hernandez, How Rwanda’s AI policy helps to shape the
evolving AI ecosystem, (Oct. 11, 2021), https://digicenter.rw/how-rwandas-ai-policyhelps-to-shape-the-evolving-ai-ecosystem/
167
The Future Society, The Development of Rwanda’s National Artificial Intelligence
Policy, (Aug. 31, 2020) https://thefuturesociety.org/2020/08/31/development-of-rwandasnational-artificial-intelligence-policy/
168
The Future Society, Leveraging Responsible AI in the Banking Sector in Africa, (Oct.
21, 2021), https://thefuturesociety.org/2021/10/21/leveraging-responsible-ai-in-thebanking-sector-in-africa/
169
The Future Society, Opportunities & Challenges of AI in Healthcare in Africa, (Jul.
21, 2021),
https://thefuturesociety.org/2021/07/22/opportunities-challenges-of-ai-in-healthcare-inafrica/
170
Beijing Principles, https://www.baai.ac.cn/news/beijing-ai-principles-en.html
166
48
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In India, the Artificial Intelligence Foundation Trust aims to spread
and promote the quality education in the area of Artificial Intelligence and
concerned engineering streams.171 The trust will also explore the
applications of artificial intelligence in the life, i.e. agriculture, healthcare
sector, business, social media, navigation and travel, banking and finance,
security and surveillances, e-commerce and many other unexplored
application areas.
In Indonesia, the Institute for Policy Research and Advocacy
(ELSAM) is a civil society organisation that works to enhance the
democratic political order by empowering civil society. With regard to
Indonesia’s national strategy on AI, ELSAM’s researcher Alia Yofira
Karunian said the national strategy should be centered around human needs
and uphold principles of fairness, accountability and transparency as pillars
in AI implementation.172 The Big Data and AI Association (ABDI) is also
concerned with AI developments; in relation to the national strategy its
Chairman Rudi Rusdiah commented that the government should prioritize
trade and industrial affairs in AI development to reap the economic
benefits.173
Furthermore, the Association for Civil Rights in Israel, which is the
oldest and most influential civil and human rights organization advocating
across the broad spectrum of human rights and civil liberties, has been
active in this field. It was one of the groups that brought before the Israel’s
Supreme Court a case concerning the Israeli Security Agency tracing the
phone location of those who may be infected with Covid-19, eventually
banned by the Court.174
In Russia, the Human Rights Watch and Amnesty International have
criticized the expansion of the use of facial recognition and highlighted
threats to privacy taking into account Russia’s track record of rights
violations.175 Amnesty International has also been critical of Russia’s plans
to broaden the use of widespread facial-recognition systems, saying their
Artificial Intelligence Foundation Trust https://www.aifoundation.in/index.php
The Jakarta Post, Indonesia sets sights on artificial intelligence in new national
strategy (Aug. 14, 2020), https://www.thejakartapost.com/news/2020/08/13/indonesiasets-sights-on-artificial-intelligence-in-new-national-strategy.html
173
ibid. See ABID, https://www.abdi.id/
174
BBC News, Coronavirus: Israeli court bans lawless contact tracing (Apr. 27, 2020),
https://www.bbc.com/news/technology-52439145
175
Human Rights Watch, Russia Expands Facial Recognition Despite Privacy Concerns
- Lack of Accountability, Oversight, Data Protection (Oct. 2, 2020),
https://www.hrw.org/news/2020/10/02/russia-expands-facial-recognition-despite-privacyconcerns
171
172
49
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
expected deployment during public gatherings will “inevitably have a
chilling effect” on protesters.176
Europe
Civil Society organizations, particularly in Europe, are also shaping
national AI policies and practices. Group such as Access Now have
published detailed assessment of AI regulatory proposals177 and a report on
“trustworthy AI.”178 AlgorithmWatch has drawn attention to controversies
in the use of AI-based decision-making systems.179 BEUC, the European
consumer organization, has surveyed public attitudes toward AI,180 and in
October 2020 proposed specific AI rights for consumers.181 Privacy
International has examined the impact of AI in several context, including
advertising, welfare, and migration.182
The European Commission’s White Paper on AI provided an
opportunity for these groups to express their views on regulatory options.
Several European NGOs said that the Commission has moved too slowly to
establish a legislative framework and has placed too much emphasis on
ethics rather than fundamental rights. Access Now and EDRi said that the
Commission’s “risk-based approach” fails to safeguard fundamental
rights.183 As they explained, “the burden of proof to demonstrate that an AI
system does not violate human rights should be on the entity that develops
Radio Free Europe, Watchdog Warns About 'Chilling Effect' Of Russia’s Use Of
Facial-Recognition Technology (Jan. 31, 2020), https://www.rferl.org/a/watchdog-warnsabout-chilling-effect-of-russia-s-use-of-facial-recognition-technology/30410014.html
177
AccessNow, Mapping Regulatory Proposals for Artificial Intelligence in Europe
(Nov. 2018),
https://www.accessnow.org/cms/assets/uploads/2018/11/mapping_regulatory_proposals_
for_AI_in_EU.pdf
178
AccessNow, Europe’s Approach to Artificial Intelligence: How AI Strategy is
Evolving (Dec. 7, 2020), https://www.accessnow.org/eu-trustworthy-ai-strategy-report/
179
AlgorithmWatch, Automating Society Report 2020 (Oct. 2020),
https://automatingsociety.algorithmwatch.org
180
BEUC, Survey: Consumers see potential of artificial intelligence but raise serious
concerns (Sept. 7, 2020), https://www.beuc.eu/publications/survey-consumers-seepotential-artificial-intelligence-raise-serious-concerns/html
181
BEUC, AI Rights for Consumers (2019), https://www.beuc.eu/publications/beuc-x2019-063_ai_rights_for_consumers.pdf
182
Privacy International, Artificial Intelligence (“AI has the potential to revolutionise
societies, however there is a real risk that the use of new tools by states or corporations
will have a negative impact on human rights.”)
https://privacyinternational.org/learn/artificial-intelligence
183
Access Now and EDRi, Attention EU regulators: we need more than AI “ethics” to
keep us safe (Oct. 21, 2020), https://edri.org/our-work/attention-eu-regulators-we-needmore-than-ai-ethics-to-keep-us-safe/
176
50
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
or deploys the system” and “such proof should be established through a
mandatory human rights impact assessment.”
BEUC wrote “a strong regulatory framework is necessary” to
“facilitate innovation and guarantee that consumers can fully reap the
benefits of the digital transformation of our societies but are protected
against the risks posed by AI.”184 The German consumer association vzbv
has also said that the EC recommendation is too narrow.185 Risky
applications that can cause immense harm to consumers’ self-determination
would then most likely be out of the scope, such as insurance, e-commerce,
and smart personal assistants like Amazon Echo/Alexa. The European
Commission’s plan also appears to include only machine-learning
applications. This would exclude a range of expert systems, such as. the
German credit scoring system “Schufa.” According to vzbv, this is not
technology neutral as it should be.
In the fall of 2020, more than a dozen NGOs in Europe joined
together to ban biometric mass surveillance.186 The Reclaim Your Face
coalition demands “transparency, red lines, and respect for humans,” and
has specifically objected to the deployment of facial recognition in
Belgrade. According to the organizations, “ReclaimYourFace is a European
movement that brings people’s voices into the discussion around biometric
data used to monitor the population. We question why these sensitive data
are being used and raise the alarm on the impact on our freedoms in public
spaces.”187
In 2021, the Reclaim YourFace campaign continued to gather
support. On January 7, 2021, the European Commission formally
recognized the campaign as a European Citizen Initiative.188 As of February
2022, approximately 68,000 signatures in support had been received.189
Signatures will continue to be gathered until August 2022.
BEUC, BEUC’s Response to the European Commission’s White Paper on Artificial
Intelligence (June 2020), https://www.beuc.eu/publications/beuc-x-2020049_response_to_the_ecs_white_paper_on_artificial_intelligence.pdf
185
Vzbv, White Paper on Artificial Intelligence: Proposals of the Federation of German
Consumer Organisations (May 11, 2020),
https://www.vzbv.de/sites/default/files/downloads/2020/06/18/20_06_11_vzbv_ec_white
paper_ai_comment_final.pdf
186
Reclaim Your Face, https://reclaimyourface.eu
187
Reclaim Your Face, The Movement, https://reclaimyourface.eu/the-movement/
188
European Commission, European Citizen Initiative, Civil society initiative for a ban
on biometric mass surveillance practices, ECI(2021)000001, https://europa.eu/citizensinitiative/initiatives/details/2021/000001_en
189
Reclaim Your Face, https://reclaimyourface.eu (Accessed Feb. 11, 2022).
184
51
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In response to the release of the Commission proposal, Access Now
urged stronger action, and called for redline for applications of AI that are
incompatible with fundamental rights.190 Later in the year, Fanny Hidvégi,
Europe Policy Manager at Access Now, stated “Access Now’s priority is
not to have an EU law on AI, but to have one that is an effective instrument
to protect people’s rights,” said Fanny Hidvégi, Europe Policy Manager at
Access Now, stated “We’ve laid out the steps needed to boost the proposed
regulation’s human rights standards, and are looking forward to working
with the Council and Parliament to guarantee they are achieved.”191
United States
In the United States, the AI Now Institute at New York University
has organized important conferences192 and issued expert reports193 on
several AI topics. The AI Now Institute also recently provided a statement
to the New York City Council on discrimination in automated employment
decision tools.194 The Electronic Privacy Information Center (EPIC) has
pursued several innovative complaints concerning AI with the US Federal
Trade Commission,195 provided comments on AI to federal agencies,196
expert statements to Congress,197 and pursued public release of materials
concerning the activities of the National Security Commission on AI.198
EPIC has also pursued open government cases concerning the use of
proprietary forensic techniques in the criminal justice system.
Fight for the Future, an independent NGO, organized a national
campaign in the US to ban facial recognition.199 Amazon also came under
AccessNow, EU takes minimal steps to regulate harmful AI systems, must go further
to protect fundamental right (Apr. 21, 2021), https://www.accessnow.org/eu-minimalsteps-to-regulate-harmful-ai-systems/
191
AccessNow, The EU needs an Artificial Intelligence Act that protects fundamental
rights (Nov. 30, 2021), https://www.accessnow.org/eu-artificial-intelligence-actfundamental-rights/
192
AI Now Institute, Bias, https://ainowinstitute.org/symposia.html
193
AI Now Institute, Reports, https://ainowinstitute.org/reports.html
194
Dr. Sarah Myers West, AI Now Institute, Ethical Implications of Using Artificial
Intelligence and Automated Decision Systems, New York City Council (Nov. 13, 2020),
https://ainowinstitute.org/ai-now-city-council-testimony-fair-shot-act.pdf
195
EPIC, In re HireVue, https://epic.org/privacy/ftc/hirevue/
196
Fight for the Future, Ban Facial Recognition, https://www.banfacialrecognition.com
197
EPIC Urges Congress to Regulate AI Techniques, Promotes 'Algorithmic
Transparency' (Dec. 12, 2017), https://epic.org/2017/12/epic-urges-congress-toregulat.html
198
EPIC v. National Security Commission on AI, No. 19-2906 (D.D.C. Dec. 3, 2019),
https://www.epic.org/foia/epic-v-ai-commission/
199
Fight for the Future, Ban Facial Recognition, https://www.banfacialrecognition.com
190
52
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
widespread criticism from many US NGOs in 2018 about the company’s
facial recognition system Rekognition.200 In June 2020, Amazon agreed to
“pause” the police use of its facial recognition software.201 IBM and
Microsoft also agree to halt the development of facial recognition.
According to MIT Technology Review, the decision “mark[s] a major
milestone for researchers and civil rights advocates in a long and ongoing
fight over face recognition in law enforcement.”202
The Algorithmic Justice League (AJL) has advised the US Congress
on AI policy203 and facial recognition technology.204 The AJL has also
proposed the creation of a federal agency, similar to the FDA, to regulate
facial recognition technology.205 And the AJL published a landmark report
on AI bias - Gender Shades: Uncovering Gender and skin-Type bias in
Commercial AI Products.206
200
Letter from Nationwide Coalition to Amazon CEO Jeff Bezos Regarding Rekognition
(June 18, 2018), ttps://www.aclu.org/letter-nationwide-coalition-amazon-ceo-jeff-bezosregarding-rekognition
201
Karen Weise and Natasha Singer, Amazon Pauses Police Use of Its Facial
Recognition Software, New York Times (June 10, 2020),
https://www.nytimes.com/2020/06/10/technology/amazon-facial-recognitionbacklash.html
202
Karen Hao, The two-year fight to stop Amazon from selling face recognition to the
police, MIT Technology Review (June 12, 2020),
https://www.technologyreview.com/2020/06/12/1003482/amazon-stopped-selling-policeface-recognition-fight/
203
Joy Buolamwini, Artificial Intelligence; Societal and Ethical Implications, United
States House Committee on Science, Space and Technology (June 26, 2019),
https://science.house.gov/imo/media/doc/Buolamwini%20Testimony.pdf
204
Joy Buolamwini, Facial Recognition Technology (Part 1): Its Impact on our Civil
Rights and Liberties, United States House Committee on Oversight and Government
Reform (May 22, 2019),
https://docs.house.gov/meetings/GO/GO00/20190522/109521/HHRG-116-GO00-WstateBuolamwiniJ-20190522.pdf
205
AJL, Federal Recognition Technologies: A Call for a Federal Office (May 29, 2020),
https://www.ajl.org/federal-office-call
206
AJL, Gender Shades: Uncovering Gender and skin-Type bias in Commercial AI
Products, http://gendershades.org
53
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
COUNTRY REPORTS
Argentina
National AI Strategy
The Argentinean government published the National Strategy for
Artificial Intelligence (Plan Nacional de Inteligencia Artificial) in 2019,
through the Office of the President (Presidencia de la Nación). Known as
ARGENIA, for short, the Strategy builds upon pre-existing policies such as
the Digital Agenda Argentina 2030 (Agenda Digital Argentina 2030) and
the National Strategy for Science, Technology, and Innovation, Argentina
Innovates 2030 (Estrategia Nacional de Ciencia, Tecnología e Innovación
Argentina Innovadora 2030).207
With the ultimate goal of positioning Argentina as a regional leader
on AI, the ten-year Strategy seeks to transform the country through AI,
leveraging the technology in pursuit of developmental objectives built on
the UN’s sustainable Development Goals (SDGs).
Taking a people-centered approach,208 the Strategy aims to
minimize the potential risks of AI development and implementation for the
Argentinean society, by protecting personal data and individual privacy
through guidelines for the design of AI systems consistent with ethical and
legal principles. The Strategy also proposes to analyze the impact in the
production scheme, resulting effects on labor forces and prevent automate
systems from reproducing or reinforcing discriminatory or exclusionary
stereotypes.209 The Strategy addresses the following areas:
• Talent and education
• Data
• Research & Development and Innovation
• Supercomputing infrastructure
• Actions to facilitate job transitions
• Facilitating public-private co-operation on data use
Office of the President, https://www.casarosada.gob.ar/informacion/actividadoficial/9-noticias/44081-el-gobierno-presento-la-nueva-agenda-digital-2030
208
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.oecd.org/g20/summits/riyadh/examples-of-ai-national-policies.pdf
209
OECD.ai, Artificial Intelligence National Plan (Plan Nacional de Inteligencia
Artificial de la República de Argentina), https://oecd.ai/en/dashboards/policyinitiatives/http:%2F%2Faipo.oecd.org%2F2021-data-policyInitiatives-24309
207
54
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
Public services and manufacturing (as target sectors for AI
development)
The cross-cutting themes in the Strategy are:
• Ethics and regulation
• Communication and awareness building
• International co-operation
As can be seen from the wide range of topics covered, the Strategy
requires a whole-of-government effort that brings together different
government ministries under the leadership of the Digital Agenda Executive
Roundtable (Mesa Ejecutiva Agenda Digital). This effort will be supported
by twenty different government agencies, as well as a Multi-sectoral
Committee of Artificial Intelligence and a Scientific Committee of experts.
While the AI Strategy for Argentina sets out ambitious goals that
build upon other national strategies, it should be noted that the Strategy was
developed under the former president, Mauricio Macri in 2019.210 Although
the former Argentinean government set out milestones to meet specific
goals,211 there’s no indication as to how the new administration will
approach the AI strategy, and most crucially, its implementation efforts
going forward.212
Regional/Provincial
The Autonomous City of Buenos Aires launched Future City: AI
Strategy (Ciudad Futuro: Plan Estratégico Inteligencia Artificial) in
August 2021. The Plan outlines the following three objectives:
• Use AI for the city’s development
• Use AI for the benefit of the citizens
• Use cross-cutting tools to ensure the city’s sustainability
Under this strategy, the Buenos Aires government has established Buenos
Aires AI Lab (BA Laboratorio IA), which provides opportunities for
training and professional development to the youth and serves as a hub for
Jasmine Kendall, Oxford Insights’ AI Strategy Series: Argentina and Uruguay (Jan. 13
2021), https://www.oxfordinsights.com/insights/2021/1/8/oxford-insights-ai-strategyseries-argentina-and-uruguay
211
Presidencia de la Nacion, Plan Nacional de Inteligencia Artificial (Aug. 2018),
https://uai.edu.ar/ciiti/2019/buenos-aires/downloads/B1/JA-Plan-Nacional-IA.pdf
212
TMG Telecom, Repaso de las Políticas y Desarrollos Latinoamericanos sobre
Inteligencia Artificial 21 (Feb. 2020); OECD, State of Implementation of the OECD AI
Principles 16 (June 2021); OECD G20 Digital Economy Task Force, Examples of AI
National Policies 9 (2020).
210
55
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
facilitating R&D and application of AI.213 As with the National Plan, one of
the key aspects of the Buenos Aires strategy is that it aims to foster
mechanisms and tools for a development and use of AI technology that
respect fundamental values and human rights.214
Public Participation
To define the specific risks and opportunities that the AI plan should
address, the government organized several meetings to gather the
perspectives of people from different disciplines and sectors. Through the
32 working tables that were assembled, experts from the government, the
private sector, the scientific community, the academia, civil society and
international organizations collaborated actively in this effort of priorities
definition. The outcomes provided the basis for the strategic objectives and
lines of action reflected in the pan. Several creative workshops and
“unconferences” were held as well. However, mechanisms for ordinary
citizens to express their views regarding AI were not identified.
Privacy
Article 43 of Argentina’s Constitution guarantees an individual’s
access to personal data in private and public registries, and exercise agency
over how that data is used. The current data protection act of Argentina,
Law 25,326 (Personal Data Protection Law), follows international standards
on basic personal data rules, and has even been deemed adequate by the
European Commission under the former Directive 95/46/EC.215 A new
proposal has been put forward by the former administration to reform Law
25,326 and auxiliary legislation. Since 2018 a legislative draft has been
under consideration by the Argentinean National Congress, with no formal
decision made in that regard as of this date. The purpose of this reform effort
aims not only that the country keeps its international status as a jurisdiction
that provides an adequate level of protection, particularly after the passing
of the European General Data Protection Regulation, but also to keeps its
data protection regime up to date to the technological and legal
developments that have taken place in recent years. As expressed by the
former President: [“t]he objective of the proposed regime is to provide our
Buenos Aires Ciudad, Plan de Inteligencia Artificial
https://www.buenosaires.gob.ar/jefaturadegabinete/innovacion/plan-de-inteligenciaartificial
214
Buenos Aires Ciudad,
https://www.buenosaires.gob.ar/sites/gcaba/files/plan_de_inteligencia_artificial_de_la_ci
udad.pdf
215
DLA Piper, Data Protection Laws of the World – Argentina (Jan. 24, 2022),
https://www.dlapiperdataprotection.com/index.html?t=law&c=AR
213
56
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
country with a more modern legislation that respects the rights and
guarantees established by our National Constitution and that, at the same
time, adapts to new technologies and regulatory changes that have occurred
in comparative law.”216
The Agencia de Acceso a la Información Pública was a cosponsor
of both the 2018 Global Privacy Assembly on AI and Ethics and the 2020
Resolution on AI and Accountability.
Algorithmic Transparency
The current Argentinean data protection law does not contain formal
legal prescriptions that recognizes the right of citizens to receive
information about automated-decision system or to object to a decision
based solely on automatic data processing methods, among other
protections. Nevertheless, the Argentinean data protection agency, Agencia
de Acceso a la Información Pública (AAIP) has provided guidance through
a resolution issued on 2019,217 in which it recognizes that, under the right
of access enshrined in the current data protection law, data subjects have the
right to request from data controllers an explanation about the logic used by
any system that reaches decisions solely based on automated processing of
data and which can affect citizens or have pernicious legal effects on
them.218 In fact, and in the absence of a formal AI oversight agency both
under the current legal framework as well as in the National Strategy, it
seems apparent that the AAIP is poised to fill this vacuum, at least with
regard to AI uses with data protection implications.219 Although the AAIP
enjoys functional autonomy by law, the agency remains under the National
Executive Branch from a structural perspective; an aspect that, along with
the absence of proper mechanisms in place, has led civil society groups to
question the impartiality and independence of the appointment process of
its Executive Director.220 The proposed reform act includes the right of
citizens to get information about “the existence of automated decision
systems, including those that create digital profiles, as well as “meaningful
Letter 147/2018 sent by the former Argentinean President to the National Congress
through which was submitted the draft of the new data protection act from the Executive
Branch to the Legislative Branch (Sept. 19, 2018).
217
Resolución 4/2019, RESOL-2019-4-APN-AAIP (Sept. 13, 2019).
218
Anex I of Resolution 4/2019 (IF-2019-01967621-APN-AAIP).
219
Gustavo P. Giay, Diego Fernández and Manuela Adrogué, “Argentina: The Use of
Artificial Intelligence”, DataGuidance (Mar. 2021),
https://www.dataguidance.com/opinion/argentina-use-artificial-intelligence
220
La Asociación por los Derechos Civiles, Observaciones de la ADC a la candidatura
propuesta para la Dirección de la Agencia de Acceso a la Información Pública (Mar. 17,
2021, https://adc.org.ar/2021/03/17/observaciones-de-la-adc-a-la-candidatura-propuestapara-la-direccion-de-la-agencia-de-acceso-a-la-informacion-publica/
216
57
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
information” about the logic applied by those systems.221 A formal right to
object to a decision based solely on automatic processing methods it is also
included in the proposal.222
Human Rights
According to Freedom House, Argentina is considered a “free”
country under the organization’s Global Freedom Scores. receiving overall
a score of 84/100.223 There are concerns about the independence of the
Argentinean judiciary. In the international arena, Argentina has shown a
strong commitment to the protection of human rights, including
international and regional initiatives that pertain to AI. Argentina has
participated (through the AAIP), as part of the Ibero-American Network of
Data Protection Authorities (RIPD), in the drafting of two key guidelines:
the General Recommendations for the Processing of Personal Data in
Artificial Intelligence, and in the Specific Guidelines for Compliance with
the Principles and Rights that Govern the Protection of Personal Data in
Artificial Intelligence Projects. As the country’s Data protection Agency,
the AAIP is co-sponsor of both the 2018 Global Privacy Assembly
Declaration on Ethics and Data Protection in Artificial Intelligence and the
2020 Resolution on Accountability in the Development and Use of
Artificial Intelligence.
OECD/G20 AI Principles
As part of the G20 and as a prospective member to the OECD,
Argentina has endorsed the OECD/G20 AI Principles. In fact, according to
an OECD report, several policies of Argentina’s national AI strategy align
with the G20 AI principles. These include the comprehensive, humancentered and human rights-focus nature, which aligns with the Principles
for Responsible Stewardship of Trustworthy AI (Section 1); while
Argentina’s investment initiatives, the focus on conditions for AI
development, educational plans and international engagements implement
Section 2 of the G20 AI Principles (National Policies and International Cooperation for Trustworthy AI).224
Article 28 (h) of the draft bill.
Article 32 of the draft bill.
223
Freedom House, Global Freedom Scores,
https://freedomhouse.org/countries/freedom-world/scores
224
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
page 66 (Table A).
221
222
58
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Facial Recognition
Several documented cases of facial recognition technology use have
been reported in various cities and localities as well as at the provincial level
in the country. Facial recognition systems being deployed, according to
authorities, for the identification and capture of fugitives (in the
Autonomous City of Buenos Aires);225 for the identification of missing
persons and people with criminal backgrounds (town of Tigre, Buenos
Aires province)226; for the use of the police to surveille massive gatherings
(Mendoza province);227 or for the prevention and persecution of crimes
(Salta province).228 The program in the City of Buenos Aires in particular
was denounced by the Special Rapporteur from the United Nations for the
Right of Privacy when visiting the city, as a technology whose
“proportionality” was questionable when compared to the “serious privacy
implications” for people not related to any crime and for not carefully
updating and checking for accuracy.229 Human Rights Watch also
denounced the system, noting the illegal exposure of minor’s personal
information.230 The City legislature approved a bill in 2020 to authorized
use for the purpose of capturing fugitives.231 But it has been alleged that this
fact does not alter the unconstitutional character of the Buenos Aires
Al Sur, “Facial Recognition in Latin America: Trends in the Implementation of a
Perverse Technology” (2021), page 11, https://www.alsur.lat/sites/default/files/202111/ALSUR_Reconocimiento_facial_en_Latam_ES.pdf
226
Ambito, Tigre lanzó un nuevo sistema de reconocimiento facial (May 10, 2019),
https://www.ambito.com/municipios/municipios/tigre-lanzo-un-nuevo-sistemareconocimiento-facial-n5030978
227
El Sol, Reconocimiento facial: hallaron a más de 100 personas con pedido de captura
(May 20, 2019), https://www.elsol.com.ar/reconocimiento-facial-hallaron-a-mas-de-100personas-con-pedido-de-captura
228
Las cámaras de reconocimiento facial permitieron detener a una persona con pedido
de captura (June 19, 2019),
https://salta.gob.ar/prensa/noticias/las-camaras-de-reconocimiento-facial-permitierondetener-a-una-persona-con-pedido-de-captura-64939
229
OHCHR, OHCHR | Statement to the media by the United Nations Special Rapporteur
on the right to privacy, on the conclusion of his official visit to Argentina, 6-17 May 2019
(May 17, 2019),
https://www.ohchr.org/en/NewsEvents/Pages/DisplayNews.aspx?NewsID=24639&LangI
D=E
230
Human Rights Watch, Argentina publica en línea datos personales de niños y niñas
acusados de cometer delitos (Oct 9, 2020),
http://www.hrw.org/es/news/2020/10/09/argentina-publica-en-linea-datos-personales-deninos-y-ninas-acusados-de-cometer
231
Asociación por los Derechos Civiles, La Legislatura porteña debe rechazar el uso de
la tecnología de reconocimiento facial para la vigilancia del espacio público (Oct. 21,
2020), https://adc.org.ar/2020/10/21/la-legislatura-portena-debe-rechazar-el-uso-de-latecnologia-de-reconocimiento-facial-para-la-vigilancia-del-espacio-publico/
225
59
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
program.232 The increasing and unaccountable use of this technology led to
the creation of a national campaign by Association for Civil Rights
(Asociación por los Derechos Civiles), a very well-known Argentinean
human rights organization. With the slogan: “Con mi Cara No” (“No with
my face”), the organization aims to raise awareness about the dangers facial
recognition technologies poses to citizens, particularly when their data is
included within opaque and unaccountable systems.233
Lethal Autonomous Weapon Systems
Argentina has been very critical about the development and use of
lethal autonomous weapons systems, particularly those without significant
human involvement. Argentina has set out a strong position in public
statements as well as within international organizations, including the U.N.
Human Rights Council as part of the meetings regarding the Convention on
Conventional Weapons. Within the framework of those meetings,
Argentina stressed the need “to preserve meaningful human control at all
phases of the development and use” of weapons systems.234 On behalf of
the Group of Latin American and Caribbean Countries, Argentina raised
several concerns over fully autonomous weapons, including the risks of
reprisal, retaliation and terrorism.235 And Argentina has called for a
“preemptive prohibition of the development of lethal autonomous
systems.236
Evaluation
Argentina’s comprehensive, ambitious and human-centered
national strategy reflects the country’s interest in matching socioeconomic
development with strong human rights commitments in the design and
development of AI. Despite the initial enthusiasm that surrounded the
launching of the Plan, there’s no clear indication about the direction that the
Government will provide in the short and middle term. Recent incidents of
human rights and democratic violations, such as the deployment of facial
iProfessional, ¿Ahora vienen por tu cara?: este experto advierte sobre los peligros del
reconocimiento facial (May 10, 2020), https://www.iprofesional.com/tecnologia/338236reconocimiento-facial-advierten-sobre-peligros-en-argentina
233
Asociación por los Derechos Civiles, https://conmicarano.adc.org.ar/
234
Government of Argentina, Statement to the Convention on Conventional Weapons
Group of Governmental Experts on lethal autonomous weapons systems, March 26,
2019.
235
Government of Argentina, Statement to the UN Human Rights Council (May 30,
2013),
236
Government of Argentina, Statement to the Convention on Conventional Weapons
Fifth Review Conference, December 12, 2016.
232
60
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
recognition systems, has raised widespread concern that AI could be used
for other pernicious purposes. Argentina has the resources and the
infrastructure to pursue regional leadership. Argentina has also shown a
strong commitment to global AI ethics principles as well as an active
involvement in international and regional AI initiatives. The question
remains about the feasibility of implementing these ambitious objectives
with the economic, political, and social challenges that the country faces.
61
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Australia
National AI Strategy
In November 2019, the Australia government published a Roadmap
for AI, to “help develop a national AI capability to boost the productivity of
Australian industry, create jobs and economic growth, and improve the
quality of life for current and future generations.”237 Australia’s AI
Technology Roadmap is intended to help guide future investment in AI and
provide a pathway to ensure Australia captures the full potential of AI.238
As well as identifying three high potential areas of AI specialization, the
Roadmap elaborates the foundations needed in terms of skills, data
governance, trust research, infrastructure and ethics, underscoring the
mutual complementarity of the OECD AI Principles.
The Roadmap identifies three domains of AI development and
application where AI could transform Australian industry, based on existing
strengths and comparative advantages, opportunities to solve Australian
problems, and opportunities to export solutions to the rest of the world.
These domains are Heath, Aging and Disability; Cities, Town and
Infrastructure (including connected and automated vehicle technology); and
National Resources and Environment (especially building on strengths
related to mining and agriculture).
CSIRO, the national science agency, has said that AI “represents a
significant opportunity to boost productivity and improve the national
economy.”239 The agency is deploying AI for gene sequencing in crops,
sustainable fishing, to predict the failure of infrastructure, and in hospitals
to forecast demand to ensure access to emergency care.
Standards Australia also launched Australia’s AI Standards
Roadmap in March 2020 to support the implementation of the OECD AI
principles. The roadmap provides a framework for Australians to shape the
development of standards for AI internationally. It explores standards that
can promote and develop the opportunities of responsible AI, delivering
business growth, improving services and protecting consumers.240
Data61, Artificial Intelligence Roadmap: Australia’s artificial intelligence roadmap,
developed by CSIRO’s Data61 for the Australian Government.
https://data61.csiro.au/en/Our-Research/Our-Work/AI-Roadmap
238
Australian Government, CSIRSO, and Data 61, Artificial Intelligence: Solving
problems, growing the economy and improving our quality of life (2019),
https://data61.csiro.au/~/media/D61/AI-Roadmap-assets/1900346_DATA61_REPORT_AI-Roadmap_WEB_191111.pdf
239
CSIRO, Artificial Intelligence, https://www.csiro.au/en/Research/AI
240
Standards Australia (2020), An Artificial Intelligence Standard Roadmap: Making
Australia’s Voice, Heard (Final Report), Standards Australia, Sydney,
237
62
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Australia has also published an AI Ethics Framework to “help guide
businesses and governments looking to design, develop, and implement AI
in Australia.”241 Key goals are to achieve better outcomes, reduce the risk
of negative impact, and practice the highest standards of ethical business
and good governance. The eight AI Ethics Principles are Human, social and
environmental wellbeing, Human-centered values, Fairness, Privacy
protection and security, Reliability and safety, Transparency and
Explainability, Contestability, and Accountability.242 The Australian
government notes that the principles are derived from the Ethically Aligned
Design report by IEEE.
The Roadmap and the Ethics Framework were put forward at the AI
Technology Summit in 2019.243 Over 100 leaders and experts in artificial
intelligence (AI) technology gathered at the public summit to “help shape
Australia’s AI future.”
Public Participation
Standards Australia is embarking on a consultation process with
Australian representatives of industry, government, civil society and
academia to examine how technical specifications and related material can
support artificial intelligence in Australia.244 In March 2020, Standards
Australia published Artificial Intelligence Standards Roadmap: Making
Australia’s Voice Heard.245
The development of Australia’s AI Ethics Framework followed a
public consultation. The Minister for Industry, Science and Technology
Karen Andrews released a discussion paper to encourage conversations on
https://www.standards.org.au/getmedia/ede81912-55a2-4d8e-849f9844993c3b9d/O_1515An-Artificial-Intelligence-Standards-Roadmap-soft_1.pdf.aspx
241
Australian Government, Department of Industry, Science, Energy and Resources, AI
Ethics Framework, https://www.industry.gov.au/strategies-for-the-future/artificialintelligence
242
Australian Government, Department of Industry, Science, Energy and Resources, AI
Ethics Principles, https://www.industry.gov.au/data-and-publications/building-australiasartificial-intelligence-capability/ai-ethics-framework/ai-ethics-principles
243
Australian Government, Department of Industry, Science, Energy and Resources,
Techtonic: Shaping Australia’s AI Future (Nov. 27, 2019),
https://www.industry.gov.au/news-media/techtonic-shaping-australias-ai-future
244
Standards Australia, Standards Australia sets priorities for Artificial Intelligence
(Mar. 2020), https://www.standards.org.au/news/standards-australia-sets-priorities-forartificial-intelligence
245
Standards Australia, FINAL REPORT: An Artificial Intelligence Standards Roadmap:
Making Australia’s Voice Heard (Mar. 2020),
https://www.standards.org.au/getmedia/ede81912-55a2-4d8e-849f9844993c3b9d/R_1515-An-Artificial-Intelligence-Standards-Roadmap-soft.pdf.aspx
63
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
how to design, develop, deploy and operate AI in Australia.246 In particular,
the Australian government sought feedback on the draft AI Ethics Principles
presented in the discussion paper. The Minister received more than 130
submissions from government, business, academia, non-government
organizations and individuals. According to the Minister, the submissions
generally supported a principles-based framework to guide the design,
development, deployment and operation of AI in Australia. There were
questions about how the draft principles can be applied in practice. The Law
Council of Australia provided extensive comments on the Ethics
Framework. The Council expressed concerns about the administrative law
implications of AI, “an AI involved in a government decision should be able
to explain its decision-making process.”247
Noted Australian AI ethics researcher Roger Clarke published a
critical assessment of the AI Ethics Principles. In 2019 Clarke undertook an
extensive survey of AI policy frameworks and identified 10 themes and 50
principles.248 Clarke concluded that the AI Ethics Principles for Australia
adequately addressed only 13 of the 50 Principles.249 “An additional 19 are
partly or weakly addressed, and 18 are not addressed at all.” Clarke states
that “the key to achieving trust is to ensure trustworthiness of the
technologies and of organisations' uses of the technologies. That requires a
comprehensive set of principles of real substance; articulation of them for
each stage of the supply chain; educational processes; means of encouraging
their application and discouraging behaviour in breach of the principles; a
credible regulatory framework; and the enforcement of at least baseline
standards.”
A 2020 survey of Australian attitudes toward AI found high levels
of support for the use of AI to address social, humanitarian and
environmental challenges.250 The survey also found high levels of support
The Hon Karen Andrews MP, Minister for Industry, Science and Technology, Seeking
feedback on ethics of artificial intelligence (Apr. 5, 2019),
https://www.minister.industry.gov.au/ministers/karenandrews/media-releases/seekingfeedback-ethics-artificial-intelligence
247
Law Council of Australia, Artificial Intelligence: Australia’s Ethics Framework (June
28, 2019), https://www.lawcouncil.asn.au/docs/b3ebc52d-afa6-e911-93fe005056be13b5/3639%20-%20AI%20ethics.pdf
248
Roger Clarke, Responsible AI Technologies, Artefacts, Systems and Applications: The
50 Principles, http://www.rogerclarke.com/EC/AIP.html#App1
249
Roger Clarke, The Australian Department of Industry's 'AI Ethics Principles'
of September / November 2019: Evaluation against a Consolidated Set of 50 Principles
(Nov. 12, 2019), http://www.rogerclarke.com/EC/AI-Aust19.html
250
Monash Data Futures Institute, AI FOR SOCIAL GOOD? Australian public attitudes
toward AI and society (Aug. 2020),
246
64
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
for legislation to ban the use of lethal autonomous weapons, ensure the
safety of autonomous vehicles, and protect data privacy.
There is currently a public consultation on “Mapping Australia's
Artificial Intelligence and Autonomous Systems Capability.”251 Part A of
the survey seeks information at an organizational level about Australia’s
national artificial intelligence and autonomous systems capabilities. Part B
of the survey focusses on unique, world-leading and significant Australian
case studies and projects.
Data Protection
The Online Privacy Bill, released in 2021 as an exposure draft,
would enable a binding online privacy code for social media and certain
other online platforms. Once passed, the industry has 12 months to
develop a code or Australia’s data protection authority, the Office of the
Australian Information Commissioner, can step in to develop it. The
“other measures” in the bill will substantially increase the civil penalties
for breaches of the Privacy Act and improve the extraterritorial reach of
the Privacy Act to protect the information collected from individuals in
Australia, regardless of where the collecting entity is located or
incorporated. We can expect to see the introduction of the Privacy
Legislation Amendment (Enhancing Online Privacy and Other Measures)
Bill and further progress on the Australian government’s review of the
Privacy Act.
In October 2021, the government released a discussion paper
containing proposals and options to modernize the Privacy Act. Some
proposals and options have been inspired by other jurisdictions such as
the EU, including introducing individual rights to object and to erasure.
The government will now consider the feedback and consult with
stakeholders on specific issues before concluding its review report, which
it intends to make public after consideration. The release of an exposure
draft of Privacy Act amendments will then follow the review report, likely
in the second half of 2022 or into 2023.252
https://www.monash.edu/__data/assets/pdf_file/0019/2313262/MDFI_AI_for_Social_Go
od_report_Final.pdf
251
Australian Government, Department of Industry, Science, Energy and Resources,
Mapping Australia's Artificial Intelligence and Autonomous Systems Capability (Oct. 2,
2020) (closes Nov. 29, 2020), https://consult.industry.gov.au/digital-economy/mappingaustralias-ai-capability/
252
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
65
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Australia is a cosponsor of the 2020 GPA Resolution on
Accountability in the Development and Use of Artificial
Intelligence.253
Global Partnership on AI
Australia joined the Global Partnership on AI as a founding member
in June 2020.254 Minister Andrews stated, “Australia is committed to
responsible and ethical use of AI. Membership of the GPAI will allow
Australia to showcase our key achievements in AI and provide international
partnership opportunities which will enhance our domestic capability.”
Andrews further stated, “Membership of the GPAI will build on the work
the Government started at last year’s National AI Summit, which brought
together 100 AI experts to discuss the challenges and opportunities which
AI will present for the Australian economy.”
Australia and Singapore, building on their pre-existing trade
agreement, also signed the Singapore-Australia Digital Economy
Agreement (SADEA) in the same year, where Parties agreed to advance
their cooperation on AI.255
Algorithmic Transparency
The concept of Algorithmic Transparency is briefly addressed in the
AI Ethics Framework. The Victorian Information Commissioner warns of
risks associated with “corporate cooption” of transparency and
accountability mechanisms.256 The paper argues that “significant resources
must be invested in developing the necessary skills in the public sector for
deciding whether a machine learning system is useful and desirable, and
how it might be made as accountable and transparent as possible.”
International Conference of Data Protection and Privacy Commissioners, Resolution
on Accountability in the Development and Use of Artificial Intelligence (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/FINAL-GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN-1.pdf
254
The Hon Karen Andrews MP, Minister for Industry, Science and Technology,
Australia joins global partnership on artificial intelligence (June 16, 2019),
https://www.minister.industry.gov.au/ministers/karenandrews/articles/australia-joinsglobal-partnership-artifical-intelligence
255
The Government of the Republic of Singapore and the Government of Australia,
Memorandum of Understanding between the Government of the Republic of Singapore
and the Government of Australia on Cooperation on Artificial Intelligence,
https://www.mti.gov.sg/-/media/MTI/Microsites/DEAs/Singapore-Australia-DigitalEconomy-Agreement/MOUs/MOU-on-Cooperation-on-Artificial-Intelligence.pdf.
256
Goldenfein, Jake, Algorithmic Transparency and Decision-Making Accountability:
Thoughts for Buying Machine Learning Algorithms (Aug. 31, 2019),
https://ssrn.com/abstract=3445873
253
66
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In early 2019, the Australian Human Rights Commission called for
an AI Policy Council to guide companies and regulators as artificial
intelligence technology. "When companies use AI decision-making
systems, they must build them in a way that allows a person to understand
the basis of decisions that affect them. This is fundamental to ensuring
accountability and will be really important for all companies that use AI,"
Human Rights Commissioner Ed Santow said.257
In a 2020 paper, Santow called on the Australian government to
modernize privacy and human rights laws to take into account the rise of
artificial intelligence.258 "We need to apply the foundational principles of
our democracy, such as accountability and the rule of law, more effectively
to the use and development of AI," he said.
OECD/G20 AI Principles
Australia has endorsed the OECD and the G20 AI Principles.
Regarding implementation of the AI Principles, the OECD notes the
Australia Roadmap for AI, the AI Ethics Framework, and the Australia’s
AI Standards Roadmap, “currently under development and intended to
identify priority areas for AI standards development and a pathway for
Australian leadership on international standardisation activities for AI.”259
The OECD also notes the work of Australia on trustworthy AI for health.
Human Rights
Australia is a signatory to many international human rights treaties
and conventions. Freedom House ranked Australia very highly (97/100) in
2020 and 2021 and reported that, “Australia has a strong record of
advancing and protecting political rights and civil liberties. Challenges to
these freedoms include the threat of foreign political influence, harsh
James Eyers, Call for 'AI policy council' to govern how algorithms use personal
information, Financial Review (Mar. 15, 2020), https://www.afr.com/technology/call-forai-policy-council-to-govern-how-algorithms-use-personal-information-20190315-h1cej1
258
Australian Human Rights Commission, Human Rights and Technology: Discussion
Paper (Dec. 2019), https://tech.humanrights.gov.au/sites/default/files/201912/TechRights2019_DiscussionPaper.pdf
259
OECD, G20 Digital Economy Task Force, Examples of National AI Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf; OECD
(2021), State of implementation of the OECD AI Principles: Insights from national AI
policies (Jun 18, 2021), https://doi.org/10.1787/1cd40c44-en; OECD (2021), An overview
of national AI strategies and policies (August 2021),
https://goingdigital.oecd.org/data/notes/No14_ToolkitNote_AIStrategies.pdf.
257
67
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
policies toward asylum seekers, and ongoing difficulties ensuring the equal
rights of indigenous Australians.”260
Evaluation
Australia has set out an AI Roadmap and an AI Ethics Framework.
Australia has encouraged public participation in the development of AI
policy, joined the Global Partnership on AI and has a strong record on
human rights. Australia has independent agencies, including a national
regulator for privacy and freedom of information261 and a human rights
commission that is engaged in AI oversight. Australia was also a cosponsor
of the GOA resolution on Accountability in the development and use of AI.
While there is no express support for the Universal Guidelines for AI,
Australia’s adopted policies are similar to those recommended in the UGAI.
Questions have also been raised about the adequacy of the Ethics
Framework.
Freedom House, Freedom in the World 2020 – Australia (2020),
https://freedomhouse.org/country/australia/freedom-world/2020;;Freedom House,
Freedom in the World 2020 – Australia (2021),
https://freedomhouse.org/country/australia/freedom-world/2021.
261
Australian Government, Office of the Australian Information Commission, Human
Rights and Technology Discussion Paper (Dec. 2019),
https://tech.humanrights.gov.au/sites/default/files/201912/TechRights2019_DiscussionPaper.pdf
260
68
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Austria
National AI Strategy
The Austrian Government presented the national AI strategy in
August 2021.262 The objectives are:
“a broad use of AI oriented to the common good.
Positioning Austria as an innovation location for AI in key
areas and fields of strength.
• The development and use of AI is intended to secure the
competitiveness of Austria as a location for technology and
business.”
The Austrian strategy is oriented towards the two cornerstones of
the European AI strategy (ecosystem for trust and ecosystem for excellence)
and applies these to the future use of AI in Austria. Austria thus not only
supports increased cooperation at the European level, as proposed in the
White Paper263 and the latest AI package,264 but also intends to shape
national AI ecosystems in line with European goals.265
The Austrian government emphasized a human-centered approach
to ensure that resources are used for the benefit of fundamental European
values and respecting and guaranteeing fundamental and human rights, such
as privacy and the principle of equality. To this end, the Federal
Government plans to initiate and promote the involvement of citizens.266
Furthermore, the Austrian government intends to define ethical
principles.267 For this purpose, Austria developed guidelines for the ethical
•
•
Federal Ministry of the Republic of Austria, Digital and Economic Affairs, Strategy of
the Austrian Federal Government for Artificial Intelligence "AIM AT 2030"
https://www.bmdw.gv.at/en/Topics/Digitalisation/Strategy/Artificial-Intelligence.html.
263
European Commission, White Paper on Artificial Intelligence: a European approach to
excellence and trust, COM(2020) 65 final, (Feb. 19, 2020),
https://ec.europa.eu/info/sites/default/files/commission-white-paper-artificialintelligence-feb2020_en.pdf.
264
European Commission, Proposal for an AI Regulation laying down harmonised rules
on artificial intelligence, COM/2021/206 final (Apr. 21, 2021), https://eurlex.europa.eu/legalcontent/EN/TXT/?qid=1623335154975&uri=CELEX%3A52021PC0206.
265
AIM AT 2030, at 20.
266
Id. at 22.
267
Whereas the Council for Robotics and Artificial Intelligence [Rat für Robotik und
Künstliche Intelligenz (ACRAI)] laid the foundation to this discussion in the 2018 white
paper “Die Zukunft Österreichs mit Robotik und Künstlicher Intelligenz positiv
gestalten”: https://www.acrai.at/wpcontent/uploads/2019/04/ACRAI_whitebook_online_2018.pdf
262
69
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
use of trustworthy AI.268 According to these, AI systems must fulfil three
basic principles to be considered trustworthy. They must:
● “be lawful by respecting all existing laws and regulations;
● respect ethical principles and values such as equality and
fairness; and
● be robust, both in a technical sense and from a societal
perspective.”
A further pinpoint for the Austrian government is to implement a
legal framework for AI. The goal is a human-centered use of AI that serves
the common good while at the same time promoting competitiveness and
innovation. To achieve this, a clear legal framework shall be created that
releases the innovation in science and economy, reduces uncertainties and
at the same time guarantees legal certainty. The Austrian Federal
Government supports the creation of a Europe-wide legal framework for AI
applications to avoid isolated national solutions.
In conclusion, an interministerial working group chaired by the
Federal Ministry for Climate Protection, Environment, Energy, Mobility,
Innovation and Technology and Federal Ministry of Science, Research and
Economy is to be set up to accompany the implementation of the strategy
and to promote regular updates. In addition, the ongoing involvement of the
relevant stakeholders and the public will be ensured. This is to be done in
part through 64 defined measures.269
Public Participation
Experts and other stakeholders were involved in the development of
the national AI strategy.270 The strategy also provides for broad
participation of civil society organizations, intermediaries, and citizens in
the implementation of the measures. Furthermore, the Federal Government
endeavors to formulate its target provisions in close coordination and
comprehensive agreement with the fundamental values and objectives of
the European Union and the Community measures. With this strategy,
Austria is thus also contributing to the promotion of Europe's industrial and
technical performance and supporting the spread of AI throughout the
European Union's economy.
High-Level Expert Group on Artificial Intelligence (2018): Ethik-Leitlinien für eine
vertrauenswürdige KI. ec.europa.eu/digital-single-market/en/news/ ethics-guidelinestrustworthy-ai
269
AIM AT 2030, at 62.
270
The Austrian Council on Robotics and AI (ACRAI) served as an advisor to the
Government and its departments until Oct. 23,2021, https://www.acrai.at/en/home/
268
70
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Facial Recognition
After a one-year test phase, facial recognition for law enforcement
has been in regular operation in Austria since August 2020.271 The use of
the software is regulated by the Security Police Act (SPG) and the Code of
Criminal Procedure (StPO). According to these, the police can only use
digital image comparison if there is a suspicion of the commission of an
intentional judicially punishable act in the case of unknown perpetrators.
The Ministry of the Interior published information about the use of the
software after parliamentary inquiries. According to the report, the Federal
Criminal Police Office may use the system to investigate intentional acts,
authorized by the judiciary, regardless of the level of punishment for an
offense. The EDE database includes about 600,000 persons.
Predictive Policing
Efforts to base police work on computer-assisted forecasts have
existed in Austria since 2004.272 Due to the increasing importance of big
data and AI, an increase and expansion of predictive policing methods is to
be expected in the next few years. Most of the predictive policing methods
developed or applied in Austria do not affect the scope of protection of the
right to respect for privacy (Art 8 ECHR, Art 7 GRC) or the fundamental
right to data protection (Art 1 § 1 Abs 1 DSG, Art 8 GRC), and are intended
in particular to support the patrol service and burglary prevention.273 The
situation was different with the project called INDECT, where an Austrian
university, the FH Technikum Wien, was also involved.274 In this project,
personal data from social media was to be combined with retained data and
video recordings in order to be able to identify "abnormal behaviour" at an
early stage. The project was funded by the European Commission during its
term from 2009 to 2014.275 Neither the official project website, nor the
Federal Ministry of the Interior, Query Response Parliamentary Question No. 2648/J:
"Findings from the Test Operation of the Face Recognition System (2662/AB)," (Sept. 4,
2020) https://www.parlament.gv.at/PAKT/VHG/XXVII/AB/AB_02662/index.shtml#]
(accessed 25 November 2021).
272
Adensamer/Klausner, Ich weiss, was du nächsten Sommer getan haben wirst:
Predictive Policing in Österreich, in: juridikum, Zeitschrift für Kritik | Recht |
Gesellschaft, 2019, 419, https://doi.org/10.33196/juridikum201903041901
273
Id.
274
Laub, INDECT: Anonymous macht gegen totale Überwachung mobil, in:
derStandard.at, (July 20, 2012) [https://derstandard.at/1342139631592/INDECT-TotaleUeberwachung- als-EU-Projekt.
275
Tajani, Answer to a written question - Indect project, data protection breach
- E-1332/2010 and E-1385/2010, (May 3, 2020)
https://www.europarl.europa.eu/doceo/document/E-7-2010-1332-ASW_EN.html
271
71
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
official project website of the FH Technikum Wien show further
information on this project.276
AI Oversight
The Federal Ministry Republic of Austria Climate Action,
Environment, Energy, Mobility, Innovation and Technology is tasked with
AI oversight. Under the broad topic of innovation, the Ministry oversees
topics such as digital, human resources, international/EU, future mobility,
research technology and innovation policy in Austria, production of the
future, publications, Austrian Registry for Space Objects, Space
Technology, Technologies for Sustainable Development, Technology
Transfer.277
The Austrian Council on Robotics and Artificial Intelligence
(ACRAI) which consisted of experts on robotics, and artificial intelligence
from industry, research and teaching advised the Federal Ministry Republic
of Austria Climate Action, Environment, Energy, Mobility, Innovation and
Technology on top priorities, topical issues, challenges, risks, on the use of
artificial intelligence as well as robotics and autonomous systems. The
Robots Council also commissioned a survey of 1,000 Austrians where twothirds asked that a strategy be developed for AI and for handling robots in
the country. On the usage of robots in Austria, Minister Leichtfried states
that “we decide where the journey goes, the human being must always be
the centre of attention.”278
Austria was not a signatory to either the 2018 GPA Declaration on
Ethics and Data Protection in Artificial Intelligence or the 2020 Resolution
on Accountability in the Development and Use of Artificial Intelligence.
Algorithmic Transparency
Austria is subject to the transparency obligation in the GDPR for the
processing of personal data. Austria is a member of the Council of Europe
but has not yet ratified the modernized Privacy Convention (“108+), which
includes a provision on algorithmic transparency
Adensamer/Klausner, Ich weiss, was du nächsten Sommer getan haben wirst:
Predictive Policing in Österreich, p. 8-10, in: juridikum, Zeitschrift für Kritik | Recht |
Gesellschaft, 2019, 419, https://doi.org/10.33196/juridikum201903041901.
277
Federal Ministry Republic of Austria Climate Action, Environment, Energy, Mobility,
Innovation and Technology “Innovation”
https://www.bmk.gv.at/en/topics/innovation.html
278
Die Zukunft Österreichs mit Robotik und Künstlicher Intelligenz positiv gestalten
White Paper des Österreichischen Rats für Robotik und Künstliche Intelligenz
https://www.acrai.at/wp-content/uploads/2020/04/ACRAI_White_Paper_DE_.pdf.
276
72
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Research show that the Public Employment Service Austria (AMS)
makes use of algorithmic profiling of job seekers, and there have been
concerns about notable discrimination and bias.279Algorithm Watch also
disclosed that the sorting algorithm used in Austria for employment gives
lower scores to the disabled and women, and women with children are given
even more negative weight.280
OECD AI Principles
Austria endorsed the G20 AI Principles. Regarding implementation
of the AI Principles, the OECD notes Austria’s active involvement in
relevant international organisations, the EU and other processes and the
specific addressing of “human-centred values and fairness, robustness,
security and safety, inclusive growth, sustainable development and wellbeing, investing in AI R&D and providing an enabling policy environment
for AI” within the AI strategy. 281
Human Rights
Human Rights are defined in the context of this report as the rights
that accrue to every human being by reason of their being human. This in
line with the definition of the United Nations who define “Human rights are
rights inherent to all human beings, regardless of race, sex, nationality,
ethnicity, language, religion, or any other status. Human rights include the
right to life and liberty, freedom from slavery and torture, freedom of
opinion and expression, the right to work and education, and many more.
Everyone is entitled to these rights, without discrimination.”282
These rights are usually contained in international human rights law,
regionally and locally and the starting point here is usually the United
Nations Universal Declaration of Human rights.283 According to Freedom
House, Austria scores highly for political rights and civil liberties (93/100),
and is designated as “Free.”284Austria was the 70th country that joined the
United Nations and actively participates in its activities and serves
279
Austria’s employment agency rolls out discriminatory algorithm, sees no problem
Nicolas Kayser-Bril https://algorithmwatch.org/en/austrias-employment-agency-amsrolls-out-discriminatory-algorithm/.
280
Id.
281
OECD AI Observatory, AI Mission Austria 2030, https://oecd.ai/en/dashboards/policyinitiatives/http:%2F%2Faipo.oecd.org%2F2021-data-policyInitiatives-24233.
282
United Nations, Human Rights https://www.un.org/en/global-issues/human-rights.
283
United Nations, Universal Declaration of Human Rights (1948),
https://www.un.org/en/about-us/universal-declaration-of-human-rights] (accessed 1
November 2021).
284
Freedom House, Freedom in the World 2021 – Austria,
https://freedomhouse.org/country/austria/freedom-world/2021
73
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
alongside New York, Nairobi and Geneva as one of the four headquarters
of the United Nations.285 It further regards the policies of the United Nations
as centrally relevant to its foreign policy.286
With regards to AI policy, as part of the objectives contained in its
AI Strategy, Austria states that it will deploy AI responsibly targeting the
common good relying on the fundamental human rights.287 Austria intends
to apply AI in sectors such as health care, education and in addressing
climate change. If harnessed properly, this can be useful in supporting
human rights such as the right to healthcare, education and a safe and
healthy environment.
The strategy further stipulates that a secure framework is created in
partnership with its European partners ensuring that issues and challenges
that would arise on the basis of fundamental human rights including the
prohibition of discrimination, data protection and the right to equality is
developed. It will achieve this by leaning into international human rights
and humanitarian law framework ensuring the AI, their digital space and
standards are ethical as illustrated by the image below.
Lethal Autonomous Weapons
Austria supports a legally binding instrument that would ban
autonomous weapons and systems that are not meaningfully controlled by
humans.288 At the virtual conference, “Safeguarding Human Control over
Autonomous Weapon Systems” held in September 2021,289 the Austrian
Ministry for European and International Affairs through the Federal
Minister for European and International Affairs of Austria, situated
themselves as the vanguard of many disarmaments, non-proliferation, and
arms control issues. They also talked about the challenges of AI, and
questioned algorithms which make death or life decisions based on ethics,
United Nations, The United Nations in Vienna, https://www.unvienna.org/
Permanent Mission of Austria to the United Nations, Austria at the UN,
https://www.bmeia.gv.at/oev-wien/austria-at-the-un/
287
AIM AT 2030.
288
DW Akademie “Austria wants ethical rules on battlefield killer robots”
https://www.dw.com/en/austria-wants-ethical-rules-on-battlefield-killer-robots/a55610965 accessed 01/11/2021
289
Austrian Ministry for European and International Affairs
https://www.reachingcriticalwill.org/resources/calendar/event/15378-safeguardinghuman-control-over-autonomous-weapon-systems accessed 20/11/2021
285
286
74
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
morality and law and called for a legal norm in the form of a treaty to ensure
human control.290
Evaluation
The Austrian AI strategy, released in late 2021, follows the larger
goals of the EU strategy, emphasizing both excellence and the protection of
fundamental rights. Austria has emphasized public participation in the
development of the national AI strategy and receives expert advice from the
Austrian Council on Robotics and Artificial Intelligence, which has
emphasized the importance of human-centric AI. Austria ranks highly for
traditional human rights and is active at the OECD. However, Austria is not
a member of the Global Partnership on AI and has not ratified the
modernized privacy convention of the Council of Europe which includes an
important provision on algorithmic transparency. And concerns have
emerged about the use of AI techniques for facial surveillance and
predictive policing.
Austrian Ministry for European and International Affairs
https://www.reachingcriticalwill.org/resources/calendar/event/15378-safeguardinghuman-control-over-autonomous-weapon-systems accessed 20/11/2021
290
75
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Bangladesh
National AI Strategy
Bangladesh published its National Strategy on Artificial Intelligence
in March 2020.291 The goal is to make Bangladesh a “technologically
advanced nation by the next decade.” The National AI Strategy for
Bangladesh is driven by the slogan- “AI for Innovative Bangladesh”. The
AI Strategy identified seven national priority sectors, which are:
1) public service delivery
2) manufacturing
3) agriculture
4) smart mobility and transportation
5) skill & education
6) finance & trade
7) health
To create a “sustainable AI Ecosystem,” the report proposes six strategic
pillars, namely:
1) research and development,
2) skilling and reskilling of AI workforce
3) data and digital infrastructure
4) ethics, data privacy, security & regulations
5) funding and accelerating AI startups
6) industrialization for AI technologies
Each pillar consists of a strategic brief, a roadmap, action plan,
related stakeholders and lead ministries. Finally, a summary roadmap in the
report includes steps for the development of AI over the next five years.
Public Participation
The National AI Strategy of Bangladesh identified engagement with
media and civil societies for creating a “robust ethics, data privacy, security
and regulations guideline” for emerging technologies.292 In March 2020 as
part of its National Internet of Things (IoT) Strategy, the Bangladesh
291
Information and Communication Technology Division Government of the People’s
Republic of Bangladesh, National AI Strategy (March 2020),
https://ictd.portal.gov.bd/sites/default/files/files/ictd.portal.gov.bd/policies/e57f1366_a62
c_4d1a_8369_a9d3bc156cd5/National%20Strategy%20for%20Artificial%20Intellgence
%20-%20Bangladesh%20.pdf
292
Bangladesh National AI Strategy 40-41 (2021),
https://ictd.portal.gov.bd/sites/default/files/files/ictd.portal.gov.bd/policies/e57f1366_a62
c_4d1a_8369_a9d3bc156cd5/National%20Strategy%20for%20Artificial%20Intellgence
%20-%20Bangladesh%20.pdf
76
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
government proposed to establish “An Advisory Committee (AC) including
representatives from Government, industry, academia and community” to
provide ongoing guidance in the emerging areas of IoT.293
Data Privacy Law
The Bangladeshi data protection regime comprises several laws,
however, there is no general law on data protection yet.294 Although the
Constitution of Bangladesh does not explicitly grant the fundamental right
to privacy, Article 43 of the constitution with certain restrictions recognises
this right under certain restrictions and states that, “every citizen shall have
the right, subject to any reasonable restrictions imposed by law in the
interests of the security of the State, public order, public morality or public
health – (a) to be secured in his home against entry, search and seizure; and
b) to the privacy of his correspondence and other means of
communication.”295
The Telecommunications Act (2000) is a law “for the purpose of
development and efficient regulation of telecommunication system and
telecommunication services in Bangladesh.”296 Under Section 67 (b) of the
Act no person can “intercept any radio communication or
telecommunication nor shall utilize or divulge the intercepted
communication, unless the originator of the communication or the person
to whom the originator intends to send it has consented to or approved the
interception or divulgence.” Under Section 97 (Ka) of the Act the
government may ask the telecommunication operator to maintain records
relating to the communications of a specific user under the broad definition
of National Security and Public Interest.
The Information Communication Technology Act (2006) imposes
responsibility on any individual or body corporate handling personal or
sensitive data and requires them to maintain and implement reasonable
293
Bangladesh National AI Strategy 11 (2021),
https://bcc.portal.gov.bd/sites/default/files/files/bcc.portal.gov.bd/page/bdb0a706_e674_4
a40_a8a8_7cfccf7e9d9b//2020-10-19-15-04-9807d52e24da56e66f7ec89f7eb540ec.pdf
294
UNCTAD, Cyberlaw Tracker: The case of Bangladesh (Apr. 2020),
https://unctad.org/page/cyberlaw-tracker-country-detail?country=bd
295
Silvee, Sadiya S. and Hasan, Sabrina and Hasan, Sabrina, The Right to Privacy in
Bangladesh in the Context of Technological Advancement (Dec. 8, 2018). International
and Comparative Law Journal 1(2), Available at SSRN:
https://ssrn.com/abstract=3298069 or http://dx.doi.org/10.2139/ssrn.3298069
296
Bangladesh Telecommunication Regulatory Commission, The Bangladesh
Telecommunications Act 2001,
http://www.btrc.gov.bd/sites/default/files/telecommunication_act_english_2001.pdf
77
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
security practices for this. 297 But Section 46 of the Act states that the state
can intercept, monitor or decrypt data if it is in the interest of:
1) “the sovereignty, integrity, or security of the state;
2) friendly relations with foreign states;
3) public order;
4) for preventing incitement to the commission of any cognisable
offence relating to the above;
5) for investigation of any offence”.
The Digital Security Act (2018) came into force in full on 8 October
2018 and pertains to “offences committed through digital device.”298
Section 26 of the Act “Punishment for unauthorized collection, use etc. of
identity information.” Under the Digital Security Act (2018) two entities
have been formed: The National Data Security Council (NDSC) and the
Digital Security Agency (DSA) “to carry out the purposes of the Act”. The
NDSC under Section 13 of the Act “shall provide necessary direction and
advice to the Agency” and the DSA shall have the power to “remove or
block some data-information.”
The government is working on the draft data protection law.299
Biometric Recognition
Since 2008, the Election Commission of Bangladesh has issued a
National Identity Card (NID) which is compulsory for every Bangladeshi
citizen above the age of 18 for voting and for availing 22 types of services,
including banking, taxpayer identity number (TIN), driving license and
passport. In 2016, the government started issuing a machine readable ‘smart
NID card’ with a chip that can store encrypted data such as biometric and
identification data for enhancing security and reducing forgery.300
Algorithmic Transparency
Under Strategy 04 of the AI national roadmap: Ethics, Data Privacy,
Security & Regulations, the Bangladeshi government will create a new set
297
Bangladesh Computer Council, ICT Act 2006,
https://bcc.portal.gov.bd/site/page/8a843dba-4055-49af-83f5-58b5669c770d/298
Bangladesh e-Government Computer Incident Response Team, Digital Security Act
2020 https://www.cirt.gov.bd/wp-content/uploads/2020/02/Digital-Security-Act-2020.pdf
299
Islam, Zyma, “Bangladesh’s New Peronal Data Protection: Door Ajar for Misuse” (15
Sep 2021) https://www.theweek.in/news/sci-tech/2021/09/15/personal-data-protectionlaw-door-ajar-for-misuse.html
300
Mizan Rahman, Bangladesh launches smart national ID cards, Gulf Times (Oct. 16,
2016), https://web.archive.org/web/20180517001739/http://www.gulftimes.com/story/515953/Bangladesh-launches-smart-national-ID-cards
78
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of AI ethics guidelines to address issues such as fairness, safety,
cybersecurity, and transparency. By 2023-2024, its ICT Division and
Ministry of Law, Justice and Parliamentary Affairs intend to formulate RTE
(Right To Explanation) Guideline For AI Algorithm.301
While there is no public declaration of adapting the Human Rights
framework for AI policy in Bangladesh as of 2021, its government
acknowledged the lack of transparency of machine learning. The national
AI strategy explicitly stated302 that:
1) The EU General Data Protection Regulation (GDPR) can be a
good solution to the challenge on rules about who will be
responsible for an unwanted inversion.
2) There should be a rule of ‘right to get an explanation’ in each
and every process.
3) The impacts AI will bring to human physiology, dignity and
autonomy is a core challenge.
4) A strong legal and ethical framework on how AI would be
implemented in applications is a must.
5) AI ethics should be righteous, fundamentally sound, assessable,
reversible and inclusive.
Human Rights
According to the Freedom House report, Bangladesh has received a
score of 39/100 for political and civil rights and is considered “partly
free.”303 The report states that, “The ruling Awami League (AL) has
consolidated political power through sustained harassment of the opposition
and those perceived to be allied with it, as well as of critical media and
voices in civil society. Corruption is a serious problem, and anticorruption
efforts have been weakened by politicized enforcement. Due process
guarantees are poorly upheld and security forces carry out a range of human
right abuses with near impunity.” 304
Lethal Autonomous Weapons
Bangladesh expressed its support for multilateral talks on lethal
autonomous weapons systems at the UN General Assembly in October
301
Bangladesh National AI Strategy 40-41 (2021),
https://ictd.portal.gov.bd/sites/default/files/files/ictd.portal.gov.bd/policies/e57f1366_a62
c_4d1a_8369_a9d3bc156cd5/National%20Strategy%20for%20Artificial%20Intellgence
%20-%20Bangladesh%20.pdf
302
Id. At 47-48.
303
Freedom House, Freedom in the World 2021 – Bangladesh,
https://freedomhouse.org/country/bangladesh/freedom-world/2021
304
Id.
79
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
2016.305 The country participated for the first time in CCW306 meetings on
lethal autonomous weapons systems in 2019, but did not make any
statements. As of 2020, it has never expressed its views on calls to ban them
through a new international treaty.307
Evaluation
Bangladesh has set out an ambitious national strategy for AI that
recognizes the importance of AI ethics. Although Bangladesh does not have
a comprehensive data protection law, there is support in the national AI
strategy for a GDPR-style law and also for an explicit right of explanation.
Biometric identification is in widespread use though at present there is little
deployment of facial recognition for mass surveillance.
305
Human Rights Watch, Stopping Killer Robots:Country Positions on Banning Fully
Autonomous Weapons and Retaining Human Control (Apr. 2021),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#_ftn34
306
Certain Conventional Weapons
307
Id.
80
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Belgium
National AI Strategy
In October 2020, the Belgium government, along with thirteen other
countries, published a position paper on innovative and trustworthy AI.308
This paper sets out two visions for the EU’s development of AI: (1)
Promoting innovation, while managing risks through a clear framework and
(2) Establishing trustworthy AI as a competitive advantage.
The countries call for a borderless single market for AI in the EU.
They state that “The main aim must be to create a common framework
where trustworthy and human-centric AI goes hand in hand with
innovation, economic growth and competitiveness in order to protect our
society, maintain our high-quality public service and benefit our citizens
and businesses. This can help the EU to protect and empower their citizens,
underpin innovation and progress in society and ensure that their values are
protected.”
The 2020 Position Paper follows the 2019 AI4Belgium policy
recommendation. The AI4Belgian strategy was commissioned by the
Minister of Digital Affairs and written by the AI4Belgium coalition in
cooperation with 40 technology experts. The AI4Belgium strategy aims to
position Belgium as a leader in the European AI landscape. 309 The strategy
lays out five areas of implementation:
• Set up a new learning deal
• Develop a responsible data strategy
• Support private sector AI adoption
• Innovate and radiate”
• Improve public service and boost the ecosystem
The Responsible Data Strategy specifically targets the ethical use of
AI and proposes to:
• Share guidelines and best practices on how to address ethical
topics in business and public institutions
• Demand from the private and public sectors to communicate and
be transparent about their AI ethics policies
Position Paper on Behalf of Denmark, Belgium, the Czech Republic, Finland, France
Estonia, Ireland, Latvia, Luxembourg, the Netherlands, Poland, Portugal, Spain and
Sweden, Innovative and Trustworthy AI: Two Sides to the Same Coin (Oct. 8,
2020), https://www.permanentrepresentations.nl/binaries/nlatio/documents/publications/2
020/10/8/non-paper---innovative-and-trustworthy-ai/Non-paper++Innovative+and+trustworthy+AI+-+Two+side+of+the+same+coin.pdf
309
AI4Belgium, AI4Belgium Strategy, https://www.ai4belgium.be/wpcontent/uploads/2019/04/report_en.pdf
308
81
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Create a Belgian ethical committee to provide industry, authorities
and society with guidance on ethical and regulatory topics
In the introduction on the AI4Belgium website, the president of
Belgium, Alexander De Croo and Philippe De Backer, the Minister of
Administrative Simplification, Digital Agenda, Postal Services and
Telecom write: “This is an initial step towards an ambitious and official
Belgian AI strategy. We will start implementing some of the coalition’s
recommendations. It will also be up to our next government to uphold this
ambition and put recommendations into practice, together.”310 The
AI4Belgium coalition also encouraged the federal government to
commission a National AI strategy.311
In 2019 the “Information Report on the necessary cooperation
between the Federal State and the federated entities regarding the impact,
opportunities, possibilities and risks of the digital “smart society” was
released by a working group created by the Belgian Senate that has been
meeting since 2018.312 Their findings are grouped in six chapters:
1) Governance, ethics and human rights, and legislation
2) Economy, labour market and taxation
3) Education and training
4) Attention economy: impact on people
5) Privacy and Cybersecurity
6) Research and development
Further, recommendations are made for each of these areas. The report
states: “The development and use of artificial intelligence shall be based on
the following guiding principles: prudence, vigilance (3), loyalty (4),
reliability, justification and transparency, accountability, limited autonomy,
humanity (5), human integrity (6), and balancing of individual and
collective interests.” and “Fundamental rights, in particular human dignity
and freedom, and privacy, must be the basis and starting point for all actions
and legislation in the field of artificial intelligence.”313
•
AI4Belgium, About: Introduction, https://www.ai4belgium.be/introduction/
European Commission, Belgium AI Strategy Report, August 2020,
https://ec.europa.eu/knowledge4policy/ai-watch/belgium-ai-strategy-report_en
312
US Library of Congress, Regulation of Artificial Intelligence: Europe and Central
Asia: Belgium (July 2020) [DT], https://www.loc.gov/law/help/artificialintelligence/europe-asia.php#belgium
313
Sénat de Belgique, Rapport d’information relatif à la nécessaire collaboration entre
l’État fédéral et les entités fédérées en ce qui concerne les retombées, les opportunités, les
potentialités et les risques de la « société intelligente » numérique (Mar. 2019)
https://www.senate.be/www/webdriver?MItabObj=pdf&MIcolObj=pdf&MInamObj=pdfi
d&MItypeObj=application/pdf&MIvalObj=100664119
310
311
82
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Chamber of Representatives also formed the Working Group
on a Robo-Digital Agenda in Parliament which was tasked with designing
an agenda for the establishment of an “inclusive and sustainable robo-digital
agenda.”314315 This Working Group held its first meeting in 2018.316
Regional/Community Strategies
Belgium is a federal government. This means that there are many
different levels of government. Belgium has three regions as well as three
communities, all of which have their own governments and many of which
have also developed strategies and initiatives on digitalization or AI. The
German, French and Flemish-speaking communities are language based.
They are responsible for language, culture, education, audiovisual things
and aid to people in need. The regions, Flemish, Brussels Capital and
Wallon-region, are territory based. They are responsible for economy,
employment, housing, public works, energy transportation, environmental
and spatial planning and have some things to say concerning international
affairs. The federal government is responsible for foreign affairs, defense,
justice, finance, social security, healthcare and internal affairs.317
The Flemish region released the Vlaanderen Radicaal Digitaal in
2019 which is an action plan to foster AI which also includes supporting
awareness and training skills needed for new technology. This also entailed
5 million euros for initiatives specifically related to AI ethics and
education.318 The Walloon government published a the “Digital Wallonia
2019-2024” strategy “based on values including a cross-disciplinary
approach, transparency, coherence, openness and flexibility.”319 The
US Library of Congress, Regulation of Artificial Intelligence: Europe and Central
Asia: Belgium (July 2020), https://www.loc.gov/law/help/artificial-intelligence/europeasia.php#belgium
315
Chambres des Représentants de Belgique, Proposition de Résolution relative à la
création d’un agenda robonumérique inclusif et durable (July 2020),
https://www.lachambre.be/doc/flwb/pdf/54/2643/54k2643001.pdf#search%3D%22intelli
gence%20artificielle%20%2054%20%3Cin%3E%20keywords%22
316
Gilles van den Burre, Première réunion du groupe de travail sur l’agenda
robonumérique au Parlementn (Jan. 2018)
https://gillesvandenburre.be/2018/01/18/premiere-reunion-groupe-de-travail-lagendarobonumerique-parlement/
317
Wikipedia, Communities, regions and language areas of Belgium,
https://en.wikipedia.org/wiki/Communities,_regions_and_language_areas_of_Belgium
318
Flanders: Department for Economy, Science and Innovation, Vlaams actieplan
Artificiële Intelligentie gelanceerd (Mar. 22, 2019),
https://www.ewi-vlaanderen.be/nieuws/vlaams-actieplan-artificiele-intelligentiegelanceerd
319
Digitalwallonia.be, Digital Wallonia 2019-2024 (June 2018),
https://www.digitalwallonia.be/en/posts/digital-wallonia-2019-2024
314
83
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
government also launched DigitalWallonia4.ai which, amongst other
things, calls for “awareness-raising and training initiatives” and “it includes
practical actions to support companies that want to incorporate artificial
intelligence into their business through to developing prototypes.”320 The
Brussels region also funds several awareness and educational programs
through its regional innovation funding body, Innoviris.321322 Finally, the
Federation Wallonie Bruxelles, which is the French community of Belgium,
has appointed a digital ethics coordinator and is also planning to set up an
ethics boaffard.323
Public Opinion
A 2019 opinion survey by AI4Belgium examined the public
perception of AI, the perceived impact, and the role the government should
play in AI implementation.324 According to the survey, 76% of the
respondents hold a positive attitude towards technological developments,
while only 6% hold a negative attitude. Most respondents were worried
about the loss of privacy, security and integrity of their personal information
(85%), less use of human common sense (85%), less human interaction
(83%) and the loss of trust and control over robots and artificial intelligence
(77%).
When asked which activity to prioritize, the highest priority was
"The management of ethical risks around AI. For example, discrimination,
privacy, etc." (74%). This was followed by "supporting employees and
employers in the transition to AI in the workplace" (65%), "improving
public service through AI" (58%), "supporting research and development
(R & D) and innovation in the field of AI" (52%), "facilitating and
supporting enterprise access to AI technologies" (48%), and "supporting
start-ups engaged in AI" (45%). The majority of citizens suspect that AI
will increase inequality between highly educated and low- or unskilled
people (66%) and between persons with a privileged background and
persons without a privileged background (60%).
Digitalwallonia.be, DigitalWallonia4.ai,
https://www.digitalwallonia.be/en/projects/digitalwallonia4-ai#contacts
321
Innoviris.brussels, Get funded, https://innoviris.brussels/get-funded
322
European Commission, Belgium AI Strategy Report (Aug. 2020),
https://ec.europa.eu/knowledge4policy/ai-watch/belgium-ai-strategy-report_en
323
OECD.ai, Approach of the Federation Wallonie Bruxelles (Oct. 2019),
https://www.oecd.ai/dashboards/policy-initiatives/2019-data-policyInitiatives-24911/
324
AI4Belgium, Perceptie Artificiële Intelligentie (Feb. 2019),
https://www.ai4belgium.be/wp-content/uploads/2019/04/enquete_en.pdf
320
84
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Independent AI oversight
The Belgium Privacy Commission was reformed in 2018 due to the
implementation of GDPR. It is now called the Belgian Data Protection
Authority and has direct sanctioning powers as well as extended
enforcement competencies. It also completely restructured the entire entity
into six bodies.325326
As a further result of GDPR, the Supervisory Body for Police
Information, “the oversight body which looks at how the police use
information (Controleorgaan op politionele infomatie, COC) was reformed
to function as an independent data protection body.” This body is intended
to oversee how the police use data.327328
Furthermore, in 2019 the Parliament established the National
Human Rights Institution (NHRI). This step was welcomed by the UN and
many human rights organizations, as there were gaps in human rights
oversight on a national level. The Institution’s main goal is to facilitate
cooperation between the existing human right oversight mechanisms and
fill the gaps in the existing landscape.329330331
Belgium was not a signatory to either the 2018 GPA Declaration on
Ethics and Data Protection in Artificial Intelligence or the 2020 Resolution
on Accountability in the Development and Use of Artificial Intelligence.
Public Participation
AI4Belgium.be not only provides information on the national AI
strategy but also offers information on AI implementation. The section
“News” lists news articles on the latest happenings related to AI policy and
PWC Legal, The new Belgian Data Protection Authority: who’s who and how will it
work (Jan. 23, 2019), https://www.pwclegal.be/en/news/the-new-belgian-data-protectionauthority---whos-who-and-how-wil.html
326
Hunton Andrews Kurth, Belgium Adopts Law Reforming the Belgian Privacy
Commission (Jan. 18, 2018), https://www.huntonprivacyblog.com/2018/01/18/belgiumadopts-law-reforming-belgian-privacy-commission/
327
Algorithm Watch, Automating Society Report 2020: Belgium, 2020,
https://automatingsociety.algorithmwatch.org/report2020/belgium/
328
Supervisory Body for Police Information, https://www.controleorgaan.be/en/
329
LibertiesEU, Belgium Approves Law Creating Long Overdue Human Rights
Institution (June 15, 2020), https://www.liberties.eu/en/news/civicus-monitor-belgiumupdate-june-2019/18043
330
European Networks of National Human Rights Institutions, ENNHRI welcomes new
law adopted on National Human Rights Institution in Belgium (May 9, 2019),
http://ennhri.org/news-and-blog/ennhri-welcomes-new-law-adopted-on-national-humanrights-institution-in-belgium/
325
Amnesty International, Belgium 2019, https://www.amnesty.org/en/countries/europeand-central-asia/belgium/report-belgium/
331
85
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
industry.332 Further, there is a form to contact the coalition as well as an
opportunity to join the coalition.333334 Anyone can join, including
organizations, technology experts, policy makers as well as civil society are
encouraged to join. According to Alexander De Croo and Philippe De
Backer: “This is a coalition open to anyone who wants to build a better
Belgium.”335
Further, several regional websites, such as digitalwallonia.be,
provide information on the region’s specific initiatives and projects.336
Facial Recognition
According to AlgorithmWatch, the Belgian government is using AI
for facial recognition at the Brussels Airport, at school registrations, football
matches, and for healthcare.337 A “smart” video surveillance system is also
in use to locate criminals, solve theft cases and collect statistical
information. According to AlgorithmWatch, there is no legal framework
governing this activity by police. The Belgian Oversight Body for Police
Information (COC) has criticized the use of facial recognition at the
Brussels airport, stating that there is “too little information about the
implementation and risks of the technology as there was no clear policy or
data protection impact assessment conducted to come to a conclusion or
offer advice.” They are asking for a temporary ban of the pilot project.338
Algorithmic Transparency
Belgium is subject to the General Data Protection Regulation which
established rights to “meaningful information about the logic involved” as
well as about “the significance and the envisaged consequences.”339 The
scope of protection in Belgium is wide, meaning that “any “significant
effect” can trigger the protection of Article 22.” Further, only one of several
safeguards is mentioned, namely: the right to obtain human intervention.
AI4Belgium, News, https://www.ai4belgium.be/news/
AI4Belgium, Join, https://www.ai4belgium.be/join-ai4belgium/
334
AI4Belgium, Contact, https://www.ai4belgium.be/contact/
335
AI4Belgium, About: Introduction, https://www.ai4belgium.be/introduction/
336
Digitalwallonia.be, https://www.digitalwallonia.be/fr/projets
337
AlgorithmWatch, Automating Society 2020, (Oct. 2020),
https://automatingsociety.algorithmwatch.org/report2020/belgium/
338
COC, Visitatie-Toezichtrapport Executive Summary Publieke Versie, 2020,
https://www.controleorgaan.be/files/DIO19005_Onderzoek_LPABRUNAT_Gezichtsher
kenning_Publiek_N.PDF
339
GDPR Art. 22, Art. 13.2.f.
332
333
86
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The right to contest, express his/her
information/explanation is not mentioned.340
view,
or
receive
Lethal Autonomous Weapons Systems
In 2018, the Belgian Parliament passed the “Resolution to prohibit
use, by the Belgian Defense, of killer robots and armed drone.”341 In this
resolution the Parliament states that Belgium should:
1) Participate in international working groups within the
framework of the United Nations and the Convention on Certain
Conventional Weapons (CCW) in particular to work towards an
internationally recognized definition of killer robots and to
determine which types of weapons will fall into this category in
the future;
2) Advocate in international fora, together with like-minded
countries, for a global ban on the use of killer robots and fully
automated armed drones;
3) Ensure that the Belgian Defense never deploys killer robots in
military operations; and
4) Support the development and use of robotic technology for
civilian purposes.
However, on an international level, the Belgian government has opposed a
ban on killer robots and the creation of new international law on killer
robots. 342
OECD/G20 AI Principles
Belgium has endorsed the OECD/G20 AI Principles. In the 2021
survey, the OECD noted several example of implementation of the AI
Principles by Belgium, including the establishment of an AI Observatory,
providing financial and non-financial support to retrain and attract top AI
Malgieri, Gianclaudio, Automated decision-making in the EU Member States: The
right to explanation and other “suitable safeguards” in the national legislations,
Computer Law & Security Review, 35(5), October 2019,
https://www.sciencedirect.com/science/article/pii/S0267364918303753#sec0005
341
Chambre des représentants de Belgique [Belgian Chamber of Representatives],
Proposition de resolution relative à la création d’un agenda robonumérique inclusif et
durable [Proposal for a Resolution Regarding the Creation of an Inclusive and
Sustainable Robo-Digital Agenda] (July 27, 2017) [DT],
http://www.lachambre.be/doc/flwb/pdf/54/ 2643/54k2643001.pdf,
342
US Library of Congress, Regulation of Artificial Intelligence in Selected Jurisdictions
(Jan. 2019), https://www.loc.gov/law/help/artificial-intelligence/regulation-artificialintelligence.pdf) (Campaign to Stop Killer Robots, Report on Activities, April
2018, https://perma.cc/2M7K-SLGD)
340
87
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
talent, development of an AI self-assessment tool, and the resolution to
prohibit the use of lethal autonomous weapons by local armed forces.343
Human Rights
Belgium is a signatory to many international human rights treaties
and conventions. Belgium typically ranks among the top nations in the
world for the protection of human rights and transparency. For 2021,
Freedom House gave Belgium a rating of 96/100, and reported “Belgium is
a stable electoral democracy with a long record of peaceful transfers of
power. Political rights and civil liberties are legally guaranteed and largely
respected.”344
Evaluation
Belgium does not yet have a full-fledged official national AI
strategy and AI ethics is not a central topic in any other national strategy.
However, the regions and communities work in this area and the
AI4Belgium recommendation is a promising start. There is, at the moment,
no express support for the Universal Guidelines for AI.
OECD, State of Implementation of the OECD AI Principles: Insights from National AI
Policies 10, 14, 29, 30 (June 2021), https://www.oecd.org/digital/state-ofimplementation-of-the-oecd-ai-principles-1cd40c44-en.htm
344
Freedom House, Freedom in the World 2021 – Belgium,
https://freedomhouse.org/country/belgium/freedom-world/2021
343
88
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Brazil
National AI Strategy
Brazil is “open for the development of state-of-the-art technology
and innovation efforts, such as 4.0 Industry, artificial intelligence,
nanotechnology and 5G technology, with all partners who respect our
sovereignty and cherish freedom and data protection” said President Jair
Bolsonaro before the United Nations General Assembly in September
2020.345
In April 2021, following on the Digital Transformation Strategy (EDigital),346 the Brazilian government adopted a national AI strategy,
“Estratégia Brasileira de Inteligência Artificial” (EBIA).347 The EBIA was
devised following a public participation process which gathered over 1,000
contributions for consultation between December 2019 and March 2020.348
The EBIA sets out six key objectives: develop ethical principles that
guide responsible use of AI; remove barriers to innovation; improve
collaboration between government, the private sector and researchers;
develop AI skills; promote investment in technologies; and advance
Brazilian technological innovation and involvement at the international
level.349
President Jair Bolsonaro, Remarks at the General Debate of the 75th Session of the
United Nations General Assembly (Sept. 22, 2020),
http://www.itamaraty.gov.br/en/speeches-articles-and-interviews/president-of-thefederative-republic-of-brazil-speeches/21770-remarks-by-president-jair-bolsonaro-at-thegeneral-debate-of-the-75th-session-of-the-united-nations-general-assembly-september22-2020
346
The 2018 Estratégia Brasileira para a Transformação Digital (E-Digital) includes a
specific action “to evaluate potential economic and social impact of (...) artificial
intelligence and big data, and to propose policies that mitigate negative effects and
maximize positive results”.
https://oecd.ai/dashboards/policyinitiatives?conceptUris=http:%2F%2Fkim.oecd.org%2F
Taxonomy%2FGeographicalAre as%23Brazil
347
Government of Brazil, Estratégia Brasileira para a Transformação Digital, 2021,
https://www.gov.br/mcti/pt-br/acompanhe-omcti/transformacaodigital/arquivosinteligenciaartificial/ia_estrategia_portaria_mcti_4979_2021_anexo1.pdf
348
Government of Brazil, Estratégia Brasileira para a Transformação Digital CONTRIBUIÇÕES ADICIONAIS RECEBIDAS NA CONSULTA PÚBLICA, 2020,
http://participa.br/estrategia-brasileira-de-inteligencia-artificial/estrategia-brasileira-deinteligencia-artificial-aplicacao-nos-setores-produtivos
349
Government of Brazil, Estratégia Brasileira para a Transformação Digital (pgs. 3-4),
2021, https://www.gov.br/mcti/pt-br/acompanhe-omcti/transformacaodigital/arquivosinteligenciaartificial/ia_estrategia_portaria_mcti_4979_2021_anexo1.pdf
345
89
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Brazil suggests concrete policies can enable the development of an
AI ecosystem, including opening government data, establishing regulatory
sandboxes, fostering startups in this field, as well as directing R&D
investment funds to this area. Additionally, Brazil has said it is essential that
nations cooperate in relevant international organizations to achieve a
common understanding and develop principles of ethics and responsibility
in the use of AI.350
The development of the Artificial Intelligence strategy in Brazil was
delayed due to the Ministerial change in Brazil - the Ministry of Science,
Technology, Information and Communications (MCTIC) was split into
two: a Ministry of Science, Technology and Information (MCTI) and a
Ministry of Communication (MCom). Artificial Intelligence is now the
responsibility of a broad Directorate on Science, and Digital Innovation
(under the Secretary of Entrepreneurship and Innovation of MCTI).351”
In February 2022 the Brazilian Senate is scheduled to vote on a
Legal Framework for Artificial Intelligence (Marco Legal da Inteligência
Artificial, PL 21/2020). The bill creates a legal framework for the
development and use of Artificial Intelligence (AI) by the government,
companies, various entities and individuals.352 AI agents, those who
develop, deploy or use an AI system, will have a series of duties, such as
answering legally for decisions made by an artificial intelligence system and
ensuring that the data used respects the General Data Protection Law
(LGPD). The standard regulates the processing of personal data of
customers and users of companies in the public and private sector.
However, academics and NGOs have criticized Bill 21/2020,
warning that the bill “may help perpetuate recent cases of algorithmic
discrimination through provisions that hinder the accountability for AIinduced errors and restrict the scope of rights established in Brazil’s General
Data Protection Legislation and in the Brazilian Constitution.”353 A group
of jurists wrote that proposal privileges the regime of subjective
OECD G20 Digital Economy Task Force, Examples of AI National Policies 10
(2020), https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
351
Ministério da Ciência, Tecnologia e Inovações, Organization Chart,
https://www.gov.br/mcti/pt-br/imagens/organograma/sempi.pdf
352
Câmara dos Deputados, Projeto cria marco legal para uso de inteligência artificial no
Brasil: Texto determina que a inteligência artificial deverá respeitar os direitos humanos
e os valores democráticos (Mar. 4, 2020), https://www.camara.leg.br/noticias/641927projeto-cria-marco-legal-para-uso-de-inteligencia-artificial-no-brasil/
353
José Renato Laranjeira de Pereira and Thiago Guimarães Moraes, Promoting
irresponsible AI: lessons from a Brazilian bill, Heinrich Böll Stiftung (Feb. 14, 2022),
https://eu.boell.org/en/2022/02/14/promoting-irresponsible-ai-lessons-brazilian-bill.
350
90
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
responsibility [requiring proof of fault] not only imposing the costs of
developing Artificial Intelligence applications to the citizen – “in a patent
inversion of the constitutional values” - but also does not establish the
necessary incentives for the appropriate measures of precautions for Ai.354
The advocates also warned that the non-discrimination principle “merely
mitigates the possibility of applying systems for illicit or abusive
discriminatory purposes.” A related point is that the “pursuit of neutrality
principle” creates no binding obligation. According to the NGOs, “These
provisions reduce the application scope of the non-discrimination principle
in the Brazilian Data Protection Legislation (LGPD), which prohibits
personal data processing for illicit or abusive discriminatory purposes.”
They warn that the Brazilian AI Bill, as it is currently drafted, “gravely
undermines the exercise of fundamental rights such as data protection,
freedom of expression and equality.”
Public Participation
The Ministry of Science, Technology, Innovations and
Communications (MCTIC) organized an online public consultation
between December 2019 and February 2020 to gather inputs for “a National
Artificial Intelligence Strategy that allows to enhance the benefits of AI for
the country, mitigating any negative impacts.”355 According to the terms of
the public consultation, “the objective of the strategy is to solve concrete
problems in the country, identifying priority areas in the development and
use of AI-related technologies in which there is greater potential for
obtaining benefits. It is envisaged that AI can bring gains in promoting
competitiveness and increasing Brazilian productivity, in providing public
services, in improving people's quality of life and in reducing social
inequalities, among others.”
The consultation presented discussion keys in thematic areas related
to AI, focusing on the government's role regarding the impact of such
technologies in society. Relevant documents to artificial intelligence were
made available on the consultation website. The consultation collected
354
Open letter from jurists to the Federal Senate against article 6, item VI of PL 21A/2020, change.org, https://www.change.org/p/senado-federal-carta-aberta-de-juristasao-senado-federal-contra-o-artigo-6o-inciso-vi-do-pl-21-a-2020
355
Participate Brazil, Ministério da Ciência, Tecnologia, Inovações e Comunicações,
Brazilian Artificial Intelligence Strategy - Qualifications for a Digital Future,
http://participa.br/estrategia-brasileira-de-inteligencia-artificial/blog/apresentacao-einstrucoes
91
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
about 1,000 contributions in total, which were taken into account for the
development of the strategy proposal.356
However, academics and NGOs have stated that the debate on the
Legal Framework for Artificial Intelligence lacked public participation.
“The debate on the bill ignored the claims of experts and civil society
organisations to address the high risks of the technology regarding
fundamental rights. In contrast, Members of Congress delivered fervorous
speeches on the positive impacts of AI in society, especially as a tool for
efficiency and innovation,” wrote José Renato Laranjeira de Pereira and
Thiago Guimarães Moraes.357
Research & Development
Brazil plans to establish eight AI research centres in 2020 in four
focus areas: health, agriculture, industry, and smart cities. Aimed to conduct
research, to foster an AI ecosystem and stimulate start-ups, and to build
human capacity in related technologies, these centers will bring together
governmental, academic, and private sector entities to benefit the private
and public sectors and the workforce.358
Brazil's largest public/private AI research facility, the Artificial
Intelligence Center (C4AI), was launched in October 2020 to tackle five
major challenges related to health, the environment, the food production
chain, the future of work and the development of Natural Language
Processing technologies in Portuguese, as well as projects relating to human
wellbeing improvement as well as initiatives focused on diversity and
inclusion.359
In May 2021, the State of São Paulo Research Foundation
(FAPESP), the MCTI, and the Brazilian Internet Steering Committee
(CGI.br) announced the results of a call for proposals to establish Applied
Research Centers (ARCs) on artificial intelligence focusing on health,
agriculture, manufacturing and smart cities. During the launch, FAPESP,
OECD AI Policy Observatory, Policy Initiatives for Brazil,
https://oecd.ai/dashboards/policyinitiatives?conceptUris=http:%2F%2Fkim.oecd.org%2FTaxonomy%2FGeographicalAre
as%23Brazil
357
José Renato Laranjeira de Pereira and Thiago Guimarães Moraes, Promoting
irresponsible AI: lessons from a Brazilian bill, Heinrich Böll Stiftung (Feb. 14, 2022),
https://eu.boell.org/en/2022/02/14/promoting-irresponsible-ai-lessons-brazilian-bill.
358
OECD G20 Digital Economy Task Force, Examples of AI National Policies 10 (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
359
Angelica Mari, Brazil launches artificial intelligence center, Brazil Tech (Oct. 14,
2020)
https://www.zdnet.com/article/brazil-launches-artificial-intelligence-center/
356
92
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
MCTI and CGI.br announced that they will collectively invest BRL 1
million per year in each of the new ARCs for a period of up to ten years.
This investment will be matched by partner firms, totaling BRL 20 million
per ARC.360
Privacy
In September 2020, Brazil’s President signed the new Brazilian data
protection law, Lei Geral de Proteção de Dados Pessoais (LGPD).361 The
LGPD is the first comprehensive data protection law in Brazil and mirrors
the European Union’s GDPR.362 Before the LGPD, data privacy regulations
in Brazil consisted of various provisions spread across Brazilian
legislation.363
Seven principles underpin the protection of personal data in the
LGPD: (1) respect for privacy; (2) informative self-determination; (3)
freedom of expression, information, communication and opinion; (4) the
inviolability of intimacy, honor and image; (5) economic and technological
development and innovation; (6) free enterprise, free competition and
consumer protection; and (7) human rights, the free development of
personality, dignity and the exercise of citizenship by natural persons.
The LGPD is relevant to the processing of personal data in relation
to AI applications.364 In October 2021, the Brazilian Senate Plenary
approved a constitutional amendment which makes personal data protection
(including digital data) a fundamental human right. This amendment
strengthens the foundations of the LGPD and the data protection authority
which it established by grounding it in the constitution.365
Ministério da Ciência, Tecnologia e Inovações, Lançamento dos Centros de
Inteligência Artificial do MCTI com a FAPESP, May 4 2021,
https://www.youtube.com/watch?v=nrqargMxmX8
361
Presidency of the Republic Sub -General Secretariat for Legal Affairs, General Law
on Protection of Personal Data (LGPD) (Aug. 14, 2020) (GT)
http://www.planalto.gov.br/ccivil_03/_ato2015-2018/2018/Lei/L13709.htm; Katitza
Rodriguez, Veridiana Alimonti, A Look-Back and Ahead on Data Protection in Latin
America and Spain (Sept. 21, 2020), https://www.eff.org/deeplinks/2020/09/look-backand-ahead-data-protection-latin-america-and-spain
362
Hogan Lovells Engage, Brazil creates a Data Protection Authority (Jan. 11, 2019),
https://www.engage.hoganlovells.com/knowledgeservices/news/brazil-creates-a-dataprotection-authority
363
DLA Piper, Data Protection Laws of the World: Brazil,
https://www.dlapiperdataprotection.com/index.html?t=law&c=BR
364
Lexology, An interview with Demarest Advogados discussing artificial intelligence in
Brazil (Nov. 27, 2020), https://www.lexology.com/library/detail.aspx?g=70705701-b4c64aa7-8a8a-344dd757f578
365
Proposta de Emenda à Constituição n° 17, de 2019 (fase 2), 2021,
https://www25.senado.leg.br/web/atividade/materias/-/materia/149723
360
93
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection Authority
The LGPD establishes a national data protection authority in Brazil
Autoridade Nacional de Proteção de Dados (ANPD) as an agency of the
federal government linked to the office of the President of Brazil.366 From a
subject matter perspective, the ANPD is guaranteed technical and decisionmaking autonomy,367 and is given important attributions related to the
LGPD interpretation, application and enforcement.368
Among other powers, the National Data Protection Authority (1)
regulates the General Data Protection Law; (2) supervises compliance with
personal data protection legislation, with a view to protecting the
fundamental rights of freedom, privacy and the free development of the
natural person's personality; (3) develops the guidelines of the National
Data Protection Plan in order to protect the fundamental rights of freedom,
privacy and the free development of the personality of the natural person;
and (4) applies administrative sanctions, after the respective provisions
come into force in August 2021 and the matter is regulated, considering the
public consultation contributions.369
In September 2020, the Federal Government published the
regulatory structure of the ANPD with the objective of giving effect to the
LGPD and enabling sanctions for non-compliance.370
There is concern that the ANPD lacks independent authority. Of the
five members of the ANPD Board of Directors appointed by the President,
three were military, including the ANPD's president.371 The OECD stated in
LGPD, Art. 55-A.
LGPD, Art. 55-B.
368
LGPD, Art. 55-J. Centre for Information Policy Leadership (CIPL) and Centro de
Direito, Internet e Sociedade of Instituto Brasiliense de Direito Público (CEDIS-IDP),
The Role of the Brazilian Data Protection Authority (ANPD) under Brazil’s New Data
Protection Law (LGPD) (Apr. 17, 2020), https://www.huntonprivacyblog.com/wpcontent/uploads/sites/28/2020/08/en_ciplidp_paper_on_the_role_of_the_anpd_under_the_lgpd__04.16.pdf
369
LGPD, Art. 55-J [GT], https://www.gov.br/secretariageral/ptbr/noticias/2020/agosto/governo-federal-publica-a-estrutura-regimental-da-autoridadenacional-de-protecao-de-dados.
370
Government of Brazil, Federal Government publishes the regulatory structure of the
National Data Protection Authority: Measure complies with the General Personal Data
Protection Law and provides conditions for the operationalization of personal data
protection in Brazil (Sept. 2, 2020) [GT], https://www.gov.br/secretariageral/ptbr/noticias/2020/agosto/governo-federal-publica-a-estrutura-regimental-da-autoridadenacional-de-protecao-de-dados.
371
Paula Pagani, Rafael Szmid, Brazil’s Senate approves Presidential appointees for
Brazilian Data Protection Authority (Oct. 23, 2020),
https://www.jdsupra.com/legalnews/brazil-s-senate-approves-presidential-63220/
366
367
94
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
October 2020, “administrative and legal frameworks that leave open even a
small possibility of a privacy enforcement authority being instructed by
another administrative body on how to exercise its functions do not satisfy
the independence criterion."372 The OECD recommended that Brazil amend
the law establishing the National Data Protection Authority; ensure that the
rules for appointing the ANPD’s Board of Directors and the National
Council for the Protection of Personal Data are transparent, fair and based
on technical expertise; and guarantee an adequate and predictable budget to
the ANPD through a transparent process"
Medical Data
According to another OECD report for the G20, Brazil is in the
process of establishing regulation in the area of privacy and personal data
protection in health systems, consistent with existing legislation, including
the LGPD. To this end, the country is developing a national electronic
health records system, which aims to provide a robust database for current
medical use, as well as for technology development and innovation.373
Algorithmic Transparency
Article 20 of the LGPD establishes the right of individuals “to
request the review of decisions taken solely on the basis of automated
processing of personal data that affect his interests, including decisions
designed to define his personal, professional, consumer and credit profile or
aspects of your personality.”
As a result, “the controller must provide, whenever requested, clear
and adequate information regarding the criteria and procedures used for the
automated decision, observing the commercial and industrial secrets.”
Where the information is not provided due to the observance of commercial
and industrial secrecy, the national data protection authority “may perform
an audit to verify discriminatory aspects in automated processing of
personal data.”374
Brazilian researchers, such as Prof. Renato Leite Monteiro,
understand that a comprehensive interpretation of LGPD, in conjunction
with the Constitution, consumer law and other legal provisions, guarantees
OECD, Going Digital in Brazil 127 (Oct. 26, 2020), https://www.oecdilibrary.org/docserver/e9bf7f8a-en.pdf
373
OECD G20 Digital Economy Task Force, Examples of AI National Policies 10 (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
374
LGPD, Art. 20 [GT].
372
95
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the existence of a right to explanation in Brazil. However, this position
demands greater jurisprudential consolidation.375
The EBIA heavily features algorithmic transparency as a goal for
the development of AI capabilities and policies in Brazil. One of the critical
strategic actions delineated in the EBIA is ‘Encouraging transparency and
responsible disclosure actions regarding the use of AI systems, and promote
compliance by such systems with human rights, democratic values and
diversity.’376 The EBIA also outlined algorithmic transparency as a critical
theme to be pursued in AI research. The strategy outlined transparency as a
critical element of AI governance both regarding explainability of decisions
taken by autonomous systems and the transparency of methodologies used
in the development of AI systems, including data sources and project
procedures.377
AI and the Judiciary
With a current backlog of 78 million lawsuits, the Brazilian judicial
system operates with substantial challenges in case flow management and a
lack of resources to meet this demand378 has led to numerous initiatives379
involving Artificial Intelligence.
Against this background, the President of the National Council of
Justice, Conselho Nacional de Justiça (CNJ), a judicial agency responsible
for the administrative and financial control of the judiciary and the
Institute for Research on Internet and Society, Automated decisions and algorithmic
transparency (Nov. 16, 2019), https://irisbh.com.br/en/automated-decisions-andalgorithmic-transparency/
376
Ministério da Ciência, Tecnologia e Inovações, Estratégia Brasileira de Inteligência
Artificial pg.23, 2021, https://www.gov.br/mcti/pt-br/acompanhe-omcti/transformacaodigital/arquivosinteligenciaartificial/ia_estrategia_diagramacao_4979_2021.pdf
377
Ministério da Ciência, Tecnologia e Inovações, Estratégia Brasileira de Inteligência
Artificial pg.25, 2021, https://www.gov.br/mcti/pt-br/acompanhe-omcti/transformacaodigital/arquivosinteligenciaartificial/ia_estrategia_diagramacao_4979_2021.pdf
378
SIPA, The Future of AI in the Brazlian Judicial System: AI Mapping, Integration and
Governance, https://itsrio.org/wp-content/uploads/2020/06/SIPA-Capstone-The-Futureof-AI-in-the-Brazilian-Judicial-System-1.pdf. (The study presents an overview of the
current uses of AI in the Brazilian Judiciary and suggests the adoption of a collaborative
governance structure that allows courts to achieve greater collaboration and cooperation
using the Electronic Judicial Process (PJE). In addition, it presents an analysis of the
principles, processes, incentives and internal regulations that govern the PJE and suggests
ways of improving and expanding the current management model, in accordance with
international best practices).
379
AI devices (called “robots”), tested in the Brazilian Judiciary include Leia, Poti,
Jerimun, Clara, Radar, Elis, Sinapse, Victor, each with a specific function.
375
96
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
supervision of judges,380 has published in August 2020 a Resolution on
ethics, transparency and governance in the production and use of Artificial
Intelligence in the Judiciary. 381 The National Council of Justice Resolution
addresses AI related requirements such as respect for human rights,
preservation of equality, non-discrimination, plurality and solidarity,
transparency (from disclosure to explainability), data security, user control
and accountability.
The Public Prosecutor’s Office382 of the State of Rio de Janeiro has
reportedly invested in data science and AI to expedite investigations and
prevent crimes.383 The system allowed information from different sources
and bodies to be collected and also real-time data to be collected from
suspected criminals.384 Likewise, Brazil’s federal and state police are using
AI applications such as military drones385 and crime prediction software.386
It is worth recalling that, like the EU GDPR, the LGPD (Art. 4)
excludes “the processing of data for the purposes of public security” from
US Law Library of Congress, Brazil, Legal Research Guide – The Judicial Branch
(2011), https://www.loc.gov/law/help/legal-research-guide/brazil-judicial-branch2_2011005662_RPT.pdf
381
National Council of Justice, Resolution No. 332, Provides for ethics, transparency
and governance in the production and use of Artificial Intelligence in the Judiciary and
provides other measures (Aug. 21, 2020),
https://www.jusbrasil.com.br/diarios/documentos/917269827/resolucao-n-332-25-082020-do-cnj.
382
In Brazil, the Prosecution Service is not part of the Executive, Legislative or Judicial
branches, being totally independent. It cannot be terminated and its duties cannot be
transferred to other government agencies. Prosecutors have their independence
guaranteed by the Brazilian Constitution. Therefore, they are subordinated to an authority
for administrative purposes only, but each member of the Prosecution Service is free to
act according to their conscience and convictions under the law. Brazilian Prosecution
Service, http://www.prrj.mpf.mp.br/english
383
MPRJ Aposta em Inteligência Artificial para Agilizar Investigações no Rio, G1 (Oct.
1, 2018), https://g1.globo.com/rj/rio-de-janeiro/noticia/2018/10/01/mp-aposta-eminteligencia-artificial-para-agilizar-investigacoes-no-rj.ghtml, archived
at https://perma.cc/MYB5-99TW
384
https://www.loc.gov/law/help/artificial-intelligence/americas.php#_ftnref2 385
ISTOE, Against organized crime, PF puts unmanned aerial vehicle in the Amazon
(Aug, 20, 2016), https://istoe.com.br/contra-o-crime-organizado-pf-poe-veiculo-aereonao-tripulado-na-amazonia/
386
Sarah Griffiths, CrimeRadar is using machine learning to predict crime in Rio, Wired
UK (Aug. 18, 2016), https://www.wired.co.uk/article/crimeradar-rio-app-predict-crime.
See also United for Smart Sustainable Cities, Crime prediction for more agile policing in
cities –Rio de Janeiro, Brazil: Case study of the U4SSC City Science Application
Framework (Oct. 2019), https://igarape.org.br/wpcontent/uploads/2019/10/460154_Case-study-Crime-prediction-for-more-agile-policingin-cities.pdf
380
97
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
its scope and states that such processing “shall be governed by specific
legislation, which shall provide proportional and strictly necessary
measures in order to serve the public interest.” However, such specific
legislation does not yet exist in Brazil.387
Facial recognition
Facial Recognition is implemented by both the public and
private sectors in Brazil. According to Instituto Igarapé, a Brazilian think
tank, there were at least 48 facial recognition applications throughout 16
Federal States between 2011 to 2019.388 The main use sectors are (i) public
security, (ii) border control, (iii) transportation and (iv) education.389
In August 2018, the Brazilian Institute of Consumer Protection
(IDEC) filed a public civil action390 for breach of privacy and consumer
legislation against the São Paulo Metro operator, regarding an AI crowd
analytics system that claimed to predict the emotion, age, and gender of
metro passengers without processing personal data.391 The operator was
ordered to stop collecting data and remove the cameras, but the case moved
forward, and a decision is now expected to be made soon.
Another monitoring system with facial recognition to be installed in
the São Paulo subway network is being challenged in Court. Early 2020, the
operating company was requested to provide clarifications on risk and
impact assessment expected with the implementation of the new
technology, on how personal data will be processed, on technical databases
and security systems issues, and on actions to mitigate the potential risk of
a data breach.392
Mariana Canto, Submission to the UN Special Rapporteur on the promotion and
protection of the right to freedom of opinion and expression: The Surveillance Industry
(Feb. 2019),
https://ohchr.org/Documents/Issues/Opinion/Surveillance/MARIANA%20CANTO.pdf
388
Instituto Igarapé, Facial Recognition in Brazil, https://igarape.org.br/infograficoreconhecimento-facial-no-brasil/ (“Facial recognition became especially popular in 2019.
The year began with the announcement of a PSL delegation to China to acquire the
technology.”)
389
Thiago Moraes, Facial Recognition in Brazil, Wired (Nov. 20, 2019),
https://medium.com/@lapinbr/face-recognition-in-brazil-f2a23217f5f7
390
Instituto Brasileiro de Defesa do Consumidor (Aug. 30, 2018),
https://idec.org.br/sites/default/files/acp_viaquatro.pdf.
391
AccessNow, Facial recognition on trial: emotion and gender “detection” under
scrutiny in a court case in Brazil (June 29, 2020), https://www.accessnow.org/facialrecognition-on-trial-emotion-and-gender-detection-under-scrutiny-in-a-court-case-inbrazil/
392
Tozzini Freire, Facia Recognition is Disputed in Court (Feb. 14, 2020),
https://tozzinifreire.com.br/en/boletins/facial-recognition-is-disputed-in-court
387
98
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Brazilian police has also been using live facial recognition for
Carnival with now plans to use the technology in events involving crowds
to find wanted criminals. In 2020, police forces rolled out facial recognition
in six capitals across the country. When announcing the use of live facial
recognition, the São Paulo police said a "situation room" would monitor the
images from the cameras, which are then compared with a database
managed by a biometrics lab. According to the police, the aim is to reduce
the likelihood of mistakes, such as wrongly arresting people.393
In 2021, Brazil rolled out end-to-end biometric identification
technologies by IDEMIA for use in passenger identification at several
airports, including domestic airports in São Paulo and Rio de Janeiro.394
Many have voiced concern at the government’s embrace of facial
recognition technology, especially surrounding issues of racial bias, given
that the LGPD does not address these technologies. 395
OECD/G20 AI Principles
Brazil has endorsed the OECD and the G20 AI Principles and
referred to the OECD Principles as important guidance for the development
of its national AI strategy. Brazil has also joined the Global Partnership on
AI.396
Human Rights
Brazil is a signatory to many international human rights treaties and
conventions and is considered as a free country in the world for the
Angelica Mari, Brazilian police introduces live facial recognition for Carnival, Brazil
Tech (Feb. 25, 2020), https://www.zdnet.com/article/brazilian-police-introduces-livefacial-recognition-for-carnival/
394
Angelica Mari, Brazilian airports expand facial recognition trials, ZDNet (Nov. 22,
2021), https://www.zdnet.com/article/brazilian-airports-expand-facial-recognition-trials/;
Chris Burt, Brazil’s Pilot of IDEMIA Face Biomentrics Advances to Simultaneous
Operation at Capital Airports, Biometric Update (June 16, 2021),
https://www.biometricupdate.com/202106/brazils-pilot-of-idemia-face-biometricsadvances-to-simultaneous-operation-at-capital-airports
395
Charlotte Peet, Brazil’s embrace of facial recognition worries Black communities, Rest
of World (Oct. 22, 2021), https://restofworld.org/2021/brazil-facial-recognitionsurveillance-black-communities/; Leaders League The controversial use of facial
recognition in Brazil and Europe, Aug. 12, 2021,
https://www.leadersleague.com/fr/news/the-controversial-use-of-facial-recognition-inbrazil-and-europe
396
Government of Canada, Canada concludes inaugural plenary of the Global
Partnership on Artificial Intelligence with international counterparts in Montreal (Dec.
4, 2020), https://www.globalprivacyblog.com/legislative-regulatory-developments/uaepublishes-first-federal-data-protection-law/
393
99
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
protection of human rights and transparency.397 398 Freedom House gives
Brazil a “free” (74/100) rating for political rights and civil liberties.
According to Freedom House, Brazil is ‘a democracy that holds competitive
elections, and the political arena, though polarized, is characterized by
vibrant public debate. However, independent journalists and civil society
activists risk harassment and violent attack, and the government has
struggled to address high rates of violent crime and disproportionate
violence against and economic exclusion of minorities.’399
Lethal Autonomous Weapons
During the 2018 discussions of the Group of Governmental Experts
(GGE) on lethal autonomous weapons (LAWS),400 Brazil issued a joint
statement along with Austria and Chile, which proposed to establish an
open-ended GGE to negotiate a legally binding instrument to ensure
meaningful human control over critical functions in LAWS.401
Evaluation
Brazil has developed a robust national strategy for AI. Brazil has
endorsed the OECD/G20 AI Principles and has promoted public
participation in the development of AI policy. Brazil has established a
comprehensive law for data protection and has a fairly good record on
human rights. But the growing use of facial recognition and the absence of
new safeguards for AI systems are matters of concern. Consumer groups
have objected to the use of AI crowd analytics on metro passengers.
Freedom House, Freedom in the World 2020 – Brazil (2020),
https://freedomhouse.org/country/brazil/freedom-world/2020
398
Human Rights Watch, World Report 2020: Brazil (2020), https://www.hrw.org/worldreport/2020/country-chapters/brazil
399
Freedom House, Freedom in the World 2021 – Brazil (2021),
https://freedomhouse.org/country/brazil/freedom-world/2021
400
Group of Governmental Experts on emerging technologies in the area of lethal
autonomous weapons systems (GGE LAWS) of the High Contracting Parties to the
Convention on Prohibitions or Restrictions on the Use of Certain Conventional Weapons
Which May Be Deemed to Be Excessively Injurious or to Have Indiscriminate Effects.
401
Proposal for a Mandate to Negotiate a Legally-binding Instrument that Addresses the
Legal, Humanitarian and Ethical Concerns Posed by Emerging Technologies in the Area
of Lethal Autonomous Weapons Systems (LAWS), U.N. Doc. CCW/ GGE.2/2018/WP.7
(Aug. 30, 2018)
https://www.unog.ch/80256EDD006B8954/(httpAssets)/3BDD5F681113EECEC12582F
E0038B22F/$file/2018_GGE+LAWS_August_Working+paper_Austria_Brazil_Chile.pd
f,
397
100
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Canada
National AI Strategy
The Canadian government has stated “Artificial intelligence (AI)
technologies offer promise for improving how the Government of Canada
serves Canadians. As we explore the use of AI in government programs and
services, we are ensuring it is governed by clear values, ethics, and laws.”402
Canada has set out five Guiding Principles to “ensure the effective and
ethical use of AI.” The government has committed to “understand and
measure” impacts, be transparent about use, “provide meaningful
explanations” for AI decision-making, “be as open as we can be,” and
provide sufficient training.”
The government of Canada and the government of Quebec have
announced a joint undertaking to “advance the responsible development of
AI.”403 The Center of Excellence, established in Montreal, will “will enable
Quebec to highlight the important role of its AI ecosystem, specifically in
the area of responsible development of AI, and to take its place
internationally as an essential partner and subject-matter expert.”
In 2017, the Canadian Institute for Advanced Research (CIFAR)
launched the Pan-Canadian Artificial Intelligence Strategy that includes the
AI and Society Program and AI Policy Initiatives.404 The Pan-Canadian AI
Strategy hosted the AICan Symposium virtually in March 2021. The work
is funded by the Government of Canada, Facebook, and the RBC
Foundation.405 Canada’s federal and provincial governments have dedicated
over USD 227 million (CAD 300 million) to AI research over 2017- 22,
anchored in the three AI institutes created under the CIFAR Pan-Canadian
AI Strategy.406
Government of Canada, Responsible use of artificial intelligence (AI),
https://www.canada.ca/en/government/system/digital-government/digital-governmentinnovations/responsible-use-ai.html
403
Government of Canada, The governments of Canada and Quebec and the
international community join forces to advance the responsible development of artificial
intelligence (June 15, 2020), https://www.canada.ca/en/innovation-science-economicdevelopment/news/2020/06/the-governments-of-canada-and-quebec-and-theinternational-community-join-forces-to-advance-the-responsible-development-ofartificial-intelligence.html
404
CIFAR, Pan-Canadian Artificial Intelligence Strategy, https://www.cifar.ca/ai/pancanadian-artificial-intelligence-strategy
405
CIFAR, AICan2019: Annual Report of the CIFAR Pan-Canadian AI Strategy,
https://www.cifar.ca/docs/default-source/ai-reports/ai_annualreport2019_web.pdf\
406
OECD (2021), State of implementation of the OECD AI Principles: Insights from
national AI policies (Jun. 18, 2021), https://doi.org/10.1787/1cd40c44-en.
402
101
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In October 2020, CIFAR released the Pan-Canadian AI Strategy
Impact Assessment Report (CIFAR, 2020[65]). This report highlights the
strategy’s impact on the following domains and Canadian regions:
Commercialization and adoption of AI: 50% growth in foreign
direct investment in ICTs from 2017 to 2019.
Research & development: 109 leading researchers recruited and
retained in Canada through the Canada CIFAR AI Chairs program.
In 2019, Canadians Yoshua Bengio and Geoffrey Hinton (along
with their colleague Yann LeCun), won the ACM A.M. Turing
Award, widely considered the “Nobel Prize of Computing”.
Talent and job creation: The strategy helped create a Canadian
ecosystem that attracts and retains highly skilled talent.
Education: enrolment in math, computer and information science
postsecondary programmes grew by 26% since 2015/16,
compared to 3% growth for all topics. Social: Canadian research
institutes CIFAR, Amii, Mila and the Vector Institute prioritize AI
for Good across societal causes including health, education, and
the environment through a portfolio of programs.
Responsible AI: Canada and France founded the Global
Partnership on AI (GPAI) focusing on responsible AI.
Regional Impact Evaluation: The establishment of three AI
institutes, Amii, Mila, and the Vector Institute, created a
collaborative network across Canada, enabling regions to deepen
their respective specialized strengths while building cross-regional
synergies.407
Directive on Automated Decision-making
Canada’s Treasury Board Secretariat (TBS) has established a
Directive on Automated Decision-making and the Pre-qualified AI Vendor
Procurement Program to ensure that administrative decisions are
“compatible with core administrative law principles such as transparency,
accountability, legality, and procedural fairness.”408 Canada has developed
a questionnaire for an Algorithmic Impact Assessment to “assess and
mitigate the risks associated with deploying an automated decision system”
CIFAR (2020), Pan-Canadian AI Strategy Impact Assessment Report,
https://cifar.ca/wp-content/uploads/2020/11/Pan-Canadian-AI-Strategy-ImpactAssessment-Report.pdf.
408
Government of Canada, Directive on Automated Decision-Making, May 2, 2019,
https://www.tbs-sct.gc.ca/pol/doc-eng.aspx?id=32592
407
102
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and to comply with the Directive on Automated Decision-making.409 A
timeline indicates progress from an initial White Paper on AI in October
2016 through an AI Day in early 2019. No subsequent information is posted.
Canada has also developed a questionnaire for an Algorithmic
Impact Assessment to “assess and mitigate the risks associated with
deploying an automated decision system” and to comply with the Directive
on Automated Decision-making.224 A timeline indicates progress from an
initial White Paper on AI in October 2016 through an AI Day in early 2019.
No subsequent information is posted.
In a parallel effort to support the Directive, TBS worked with Public
Services and Procurement Canada to establish a Pre-qualified AI Vendor
Procurement Program to streamline the procurement of AI solutions and
services in the government. This new AI public procurement programme
was used to help government departments and agencies build awareness of
the solutions offered by AI. It also provided small and medium AI
companies with an opportunity to provide their services to the government.
In practice, the initiative did not yet gain traction.410
Predicting Homelessness
A new AI project in the city of London, Canada proposes to predict
and prevent homelessness. According to a news report, “the Chronic
Homelessness Artificial Intelligence (CHAI) model uses machine learning
to forecast the probability of an individual in the city’s shelter system
becoming chronically homeless within the next six months – that is,
remaining in the shelter system for more than 180 days in a year.”411
According to the development team, ‘Explainable AI’ is an important aspect
of the CHAI system. The team designed the model around the principles of
the General Data Protection Regulation (GDPR), as well as the Canadian
government’s Directive on Automated Decision-Making.
Public Participation
In 2019 Canada established an Advisory Council on Artificial
Intelligence to “inform the long-term vision for Canada on AI both
409
Government of Canada, Algorithmic Impact Assessment (AIA), July 28, 2020,
https://www.canada.ca/en/government/system/digital-government/digital-governmentinnovations/responsible-use-ai/algorithmic-impact-assessment.html
410
OECD (2021), State of implementation of the OECD AI Principles: Insights from
national AI policies (Jun 18, 2021), https://doi.org/10.1787/1cd40c44-en
411
CitiesToday, ‘Explainable AI’ predicts homelessness in Ontario city (Aug, 25, 2020),
https://cities-today.com/explainable-ai-predicts-homelessness-in-ontario-city/
103
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
domestically and internationally.”412 It is unclear whether the Advisory
Council has held meetings or issued reports. Comprised of researchers,
academics, and business leaders, the Council advises the Government of
Canada on how to build on Canada’s AI strengths to support
entrepreneurship, drive economic growth and job creation and build public
trust in AI. The Council has created two working groups to date, one on
Commercialization and another on Public Awareness (OECD.AI
forthcoming). Public awareness is a key area for the Council that
emphasized that policy design, including sectoral priorities, require the trust
and support of the public to succeed.
Canada’s AI Advisory Council created its public engagement and
consultation processes using both consultation and deliberation. The
national survey elicited an array of citizens’ input on AI use in different
sectors. The results will shape deliberative workshops that take place online
due to the pandemic. The workshops aim to find ways to address ethical
concerns raised by citizens via the survey. Among the goals of the
deliberative process is to shape a new set of guidelines and
recommendations for the development of AI.413
Data Protection
The Office of the Privacy Commissioner of Canada provides advice
and information for individuals about protecting personal information.414
The agency also enforces two federal privacy laws that set out the rules for
how federal government institutions and certain businesses must handle
personal information. The Privacy Act regulates the collection and use of
personal data by the federal government.415 The Personal Information
Protection and Electronic Documents Act (PIPEDA) applies to personal
data collected by private companies.416
Government of Canada, Protecting and Promoting Privacy Rights,
https://www.priv.gc.ca/en
413
OECD (2021), State of implementation of the OECD AI Principles: Insights from
national AI policies (Jun 18, 2021), https://doi.org/10.1787/1cd40c44-en; OECD, AI
Policy Observatory (2020), https://oecd.ai/dashboards/countries/Canada.
414
Office of the Privacy Commissioner of Canada, The Privacy Act in brief (Aug. 2019),
https://www.priv.gc.ca/en/privacy-topics/privacy-laws-in-canada/the-privacyact/pa_brief/
415
Office of the Privacy Commissioner of Canada, The Privacy Act in brief (Aug. 2019),
https://www.priv.gc.ca/en/privacy-topics/privacy-laws-in-canada/the-privacyact/pa_brief/
416
Office of the Privacy Commissioner of Canada, PIPEDA in brief (May 2019),
https://www.priv.gc.ca/en/privacy-topics/privacy-laws-in-canada/the-personalinformation-protection-and-electronic-documents-act-pipeda/pipeda_brief/
412
104
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In November 2020, the Privacy Commissioner issued proposals on
regulating artificial intelligence.417 The recommendations “aim to allow for
responsible AI innovation and socially beneficial uses while protecting
human rights.” The Commissioner recommend amending PIPEDA to:
• allow personal information to be used for new purposes towards
responsible AI innovation and for societal benefits
• authorize these uses within a rights-based framework that would
entrench privacy as a human right and a necessary element for the
exercise of other fundamental rights
• create a right to meaningful explanation for automated decisions
and a right to contest those decisions to ensure they are made fairly
and accurately
• strengthen accountability by requiring a demonstration of privacy
compliance upon request by the regulator
• empower the OPC to issue binding orders and proportional
financial penalties to incentivize compliance with the law
• require organizations to design AI systems from their conception
in a way that protects privacy and human rights
The Commissioner also highlighted a public consultation, initiated
by the OPC, that received 86 comments from industry, academia, civil
society, and the legal community, among others. Those inputs were
incorporated in separate report which informs the recommendations for law
reform.418
Algorithmic Transparency
The PIPEDA includes strong rights for individual access concerning
automated decisions.419 The PIPEDA Reform Report for AI build on public
consultations and propose to “Provide individuals with a right to
explanation and increased transparency when they interact with, or are
Office of the Privacy Commissioner of Canada, Commissioner issues proposals on
regulating artificial intelligence (Nov. 2020), Commissioner issues proposals on
regulating artificial intelligence
418
Ignacio Cofone, Office of the Privacy Commissioner of Canada, Policy Proposals for
PIPEDA Reform to Address Artificial Intelligence Report (Nov. 2020),
https://www.priv.gc.ca/en/about-the-opc/what-we-do/consultations/completedconsultations/consultation-ai/pol-ai_202011/
419
Office of the Privacy Commissioner, Canada, PIPEDA Fair Information Principle 9 –
Individual Access (Aug. 2020), https://www.priv.gc.ca/en/privacy-topics/privacy-lawsin-canada/the-personal-information-protection-and-electronic-documents-actpipeda/p_principle/principles/p_access/
417
105
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
subject to, automated processing.”420 The Cofone Report also explains that
“the right to explanation is connected to the principles of privacy,
accountability, fairness, non-discrimination, safety, security, and
transparency. The effort to guarantee these rights supports the need for a
right to explanation.”
Canada was a signatory to both the 2018 GPA Declaration on Ethics
and Data Protection in Artificial Intelligence or the 2020 Resolution on
Accountability in the Development and Use of Artificial Intelligence.421
Facial Recognition
In early 2021, an investigation found that Clearview AI’s scraping
of billions of images of people from across the Internet represented mass
surveillance and was a clear violation of the privacy rights of Canadians.422
Clearview AI’s scraping of billions of images of people from across the
Internet represented mass surveillance and was a clear violation of the
privacy rights of Canadians. “Clearview AI’s technology allowed law
enforcement and commercial organizations to match photographs of
unknown people against the company’s databank of more than 3 billion
images, including of Canadians and children, for investigation purposes.
Commissioners found that this creates the risk of significant harm to
individuals, the vast majority of whom have never been and will never be
implicated in a crime.” Clearview AI’s technology allowed law
enforcement and commercial organizations to match photographs of
unknown people against the company’s databank of more than 3 billion
images, including of Canadians and children, for investigation purposes.
Commissioners found that this creates the risk of significant harm to
individuals, the vast majority of whom have never been and will never be
implicated in a crime. A related investigation by the Office of the Privacy
Professor Ignacio Cofone, Policy Proposals for PIPEDA Reform to Address Artificial
Intelligence Report (Nov. 2020), https://www.priv.gc.ca/en/about-the-opc/what-wedo/consultations/completed-consultations/consultation-ai/pol-ai_202011/
421
International Conference on Data Protection and Privacy Commissioners, Declaration
on Ethics and Data Protection in Artificial Intelligence (Oct. 23, 2018),
https://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf; Global Privacy Assembly, Resolution on
Accountability in the Development and Use of Artificial Intelligence (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/FINAL-GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN-1.pdf
422
Government of Canada, office of the Privacy Commissioner, Clearview AI’s unlawful
practices represented mass surveillance of Canadians, commissioners say (Feb. 3, 2021),
https://www.priv.gc.ca/en/opc-news/news-and-announcements/2021/nrc_210203/?=february-2-2021
420
106
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Commissioner of Canada into the RCMP’s use of Clearview AI’s facial
recognition technology remains ongoing.423
Global Partnership on AI
In 2020, Canada and France, and a dozen other countries announced
the Global Partnership on Artificial Intelligence to support “support the
responsible and human-centric development and use of AI in a manner
consistent with human rights, fundamental freedoms, and our shared
democratic values . . .”424 According to the statement, the “GPAI will be
supported by a Secretariat, to be hosted by the OECD in Paris, as well as by
two Centres of Expertise – one each in Montréal and Paris.” As the 20202021 GPAI Chair, Canada hosted the inaugural GPAI Summit in December
2020.
Canada and the European Union recently announced that they are
collaborating to leverage artificial intelligence (AI) to help the international
community respond to COVID-19. The initiatives include the GPAI’s group
on AI and Pandemic Response and the annual EU-Canada Digital
Dialogue.425
Canada and Germany are working together to advance AI
industrialization by organising joint R&D projects on the application of AI
technologies in manufacturing, as applied to manufacturing, supply
chain and other fields. 426
OECD/G20 AI Principles
Canada endorsed the OECD and the G20 AI Principles.
Government of Canada, office of the Privacy Commissioner, OPC launches
investigation into RCMP’s use of facial recognition technology
(Feb. 28, 2020), https://www.priv.gc.ca/en/opc-news/news-andannouncements/2020/an_200228/
424
Government of Canada, Joint Statement from founding members of the Global
Partnership on Artificial Intelligence (June 15, 2020),
https://www.canada.ca/en/innovation-science-economicdevelopment/news/2020/06/joint-statement-from-founding-members-of-the-globalpartnership-on-artificial-intelligence.html
425
European Union, Joint press release following the European Union-Canada
Ministerial Meeting (Sept. 9, 2020), https://eeas.europa.eu/headquarters/headquartershomepage/84921/joint-press-release-following-european-union-canada-ministerialmeeting_en
426
Canada (2020), Canada – Germany 3+2 collaborative call for proposals on
Innovative Artificial Intelligence Solutions for Industrial Production, National Research
Council Canada, https://nrc.canada.ca/en/irap/about/international/?action=view&id=62.
423
107
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Lethal Autonomous Weapons
In 2017 Canadian academics urged Prime Minister Trudeau to
oppose Autonomous Weapon Systems, as part of the #BanKillerAI
campaign.427
Human Rights
Canada is a signatory to many international human rights treaties
and conventions. Canada typically ranks among the top ten nations in the
world for the protection of human rights and transparency (98/100 in
2021).428 Freedom House reported that, “Canada has a strong history of
respect for political rights and civil liberties, though in recent years citizens
have been concerned about fair elections and transparent governance;
humane treatment of prisoners; citizens’ right to privacy; and religious and
journalistic freedom. While Indigenous peoples and other vulnerable
populations still face discrimination and other economic, social, and
political challenges, the federal government has acknowledged and made
some moves to address these issues.”429
Evaluation
Canada is among the leaders in national AI policies. In addition to
endorsing the OECD/G20 AI Principles and establishing the GPAI with
France, Canada has also taken steps to establish model practices for the use
of AI across government agencies. Canada has an admirable record on
human rights and is now working to update its national privacy law to
address the challenges of AI. In 2021, Canada took decisive action against
ClearviewAI and continued to lead, with France, the Global Partnership on
AI. There is, at the moment, no express support for the Universal Guidelines
for AI, but Canada’s I policies are similar to those recommended in the
UGAI.
Ian Kerr, Weaponized AI would have deadly, catastrophic consequences. Where will
Canada side? The Globe and Mail, Nov. 6, 2017,
https://www.theglobeandmail.com/opinion/weaponized-ai-would-have-deadlycatastrophic-consequences-where-will-canada-side/article36841036/
428
Freedom House, Freedom in the World 2021 – Canada (2021),
https://freedomhouse.org/country/canada/freedom-world/2021
429
Freedom House, Freedom in the World 2020 – Canada (2020),
https://freedomhouse.org/country/canada/freedom-world/2020; Freedom House, Freedom
in the World 2021 –Canada (2021), https://freedomhouse.org/country/canada/freedomworld/2021.
427
108
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
China
National AI Strategy
Since 2013, the Chinese government has published several nationallevel policies, guidelines, and action plans, which reflect the intention to
develop, deploy, and integrate AI in various sectors. In 2015, Prime
Minister Li Keqiang launched the “Made in China” (MIC 2025) initiative
aimed at turning the country into a production hub for high-tech products
within the next few decades. In the same year, the State Council released
guidelines on China’s Internet +Action plan. It sought to integrate the
internet into all elements of the economy and society. The document
emphasized the importance of cultivating emerging AI industries and
investing in research and development. The Central Committee of the
Communist Party of China’s 13th 5-year plan is another notable example.
The document mentioned AI as one of the six critical areas for developing
the country’s emerging industries and as an essential factor in stimulating
economic growth. Robot Industry Development Plan,430 Special Action of
Innovation and Development of Smart Hardware Industry,431 and Artificial
Intelligence Innovation Action Plan for Higher Institutions432 illustrate
detailed action plans and guidelines concerning specific sectors.
Most notable of all is the New Generation Artificial Intelligence
Development Plan (AIDP) – an ambitious strategy to make China the world
leader in AI by 2030 and the most transparent and influential indication of
China's AI strategy’s driving forces. China’s State Council issued the AIDP
in 2017. According to the plan, AI should be used in a broad range of
sectors, including defense and social welfare. The AIDP also indicates the
need to develop standards and ethical norms for the use of AI. Remarkably,
the actual innovation and transformation are expected to be driven by the
private sector and local governments.433 The Chinese government has
handpicked three major tech giants to focus on developing specific sectors
机器人产业发展规划（2016-2020年）
https://www.ndrc.gov.cn/xxgk/zcfb/ghwb/201604/t20160427_962181.html
431
智能硬件产业创新发展专项行动 (2016-2018 年) http://www.gov.cn/xinwen/201609/21/content_5110439.htm
432
高等学校人工智能创新行动计划
http://www.moe.gov.cn/jyb_xwfb/xw_fbh/moe_2069/xwfbh_2018n/xwfb_20180608/201
806/t20180608_338911.html
433
3-year plan promoting the AIDP (2018–2020) emphasizes coordination between
provinces and local governments.
430
109
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of AI, Baidu, Alibaba and Tencent.434 In return, these companies receive
preferential contract bidding, preferential contract bidding, more
convenient access to finance, and sometimes market share protection.
With regard to local governments, there is a system of incentives for
fulfilling national government policy aims. For this reason, local
governments often become a testing ground for the central government’s
policies. A clear example of this are the surveillance technologies that were
first tested in Xinjiang435 to research into “ethnic” aspects of AI-enabled
facial recognition templates distinguishing “Uyghur” features.436 Chinese
cities and provinces, regional administrations compete for the new AI
incentives. While large metropolises, such as Tianjin and Shanghai, have
already launched multi-billion-dollar AI city Venture Capital funds and
converted entire districts and islands for new AI companies. Other
provinces are still in the process of learning and development.
AI Core Values
International Competition & National Security
The AIDP strategy document states that “the development of AI [is]
… a major strategy to enhance national competitiveness and protect national
security” and that China will “[p]romote all kinds of AI technology to
become quickly embedded in the field of national defense innovation.”
At the 8th Beijing Xiangshan Forum (BXF),437 China’s major
platform for international security and defense dialogue, Major General
Ding Xiangrong, Deputy Director of the General Office of China’s Central
Military Commission, gave a major speech in which he stated that China’s
military goals are to use AI to advance Chinese military.438 Another speaker
Meng Jing and Sarah Dai, China recruits Baidu, Alibaba and Tencent to AI ‘national
team,’ South China Morning Post (Nov. 21, 2017), https://www.scmp.com/tech/chinatech/article/2120913/china-recruits-baidu-alibaba-and-tencent-ai-national-team.
435
Angela Dely, Algorithmic oppression with Chinese characteristics: AI against
Xinjiang’s Uyghurs, Global Information Society Watch (2019),
https://www.giswatch.org/node/6165#_ftn33
436
Zuo, H., Wang, L., & Qin, J. (2017). XJU1: A Chinese Ethnic Minorities Face
Database. Paper presented at IEEE International Conference on Machine Vision and
Information Technology
(CMVIT). https://ieeexplore.ieee.org/abstract/document/7878646
437
Rajeev Ranjan Chaturvedy, Beijing Xiangshan Forum and the new global security
landscape, EastAsiaForum (Dec. 1, 2018),
https://www.eastasiaforum.org/2018/12/01/beijing-xiangshan-forum-and-the-new-globalsecurity-landscape/
438 Elsa Kania, "AlphaGo and Beyond: The Chinese Military Looks to Future
‘Intelligentized’ Warfare." Lawfare (June 5,
434
110
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Zeng Yi, a senior executive at China’s third largest defense company,
predicted that by 2025 lethal autonomous weapons, military command
decision-making would be commonplace and said that ever-increasing
military use of AI is “inevitable.” Notably, he emphasized that military AI
would replace the human brain and exercise independent judgment by
stating that “AI may completely change the current command structure,
which is dominated by humans” to one that is dominated by an “AI cluster.”
These sentiments are shared by academics from the People’s Liberation
Army (PLA) who believe that AI will be used to predict battlefield
situations and outpace human decision-making.439
China’s Ministry of National Defense has established two major
new research organizations focused on AI and unmanned systems: the
Unmanned Systems Research Center (USRC) and the Artificial Intelligence
Research Center (AIRC).440 According to some experts, China is pursuing
the most aggressive strategy for developing AI for military uses among the
major military powers.441 In the spring of 2017, a civilian Chinese university
with ties to the military demonstrated an AI-enabled swarm of 1,000
uninhabited aerial vehicles at an airshow. A media report released after the
fact showed a computer simulation of a similar swarm formation finding
and destroying a missile launcher.[1] Open-source publications indicate that
China is also developing a suite of AI tools for cyber operations.[1] [12]
Economic Development
The AIDP promotes and highlights the reconstruction of economic
activities using AI as the driving force behind a new round of industrial
transformation, which will “inject new kinetic energy into China’s
economic development.”442 Guiding Opinions on Promoting on Promoting
2017), https://www.lawfareblog.com/alphago-and-beyond-chinese-military-looks-futureintelligentized-warfare.
439
Kania EB (2017a) 杀手锏 and 跨越发展: trump cards and leapfrogging. Strategy
Bridge. https ://thestrategybridge.org/the-bridge/2017/9/5/-and-trump-cards-andleapfrogging
440
Gregory C. Allen, Understanding China’s AI Strategy: Clues to Chinese Strategic Thinking
on Artificial Intelligence and National Security 4-9, Center for a New American Security (Feb.
6, 2019), https://www.cnas.org/publications/reports/understanding-chinas-ai-strategy
441
Adrian Pecotic, Whoever Predicts the Future Will Win the AI Arms Race, Foreign
Policy (Mar. 5, 2019), https://foreignpolicy.com/2019/03/05/whoever-predicts-the-futurecorrectly-will-win-the-ai-arms-race-russia-china-united-states-artificial-intelligencedefense/
442
New America, China's 'New Generation Artificial Intelligence Development Plan'
(English translation) (2017), https://www.newamerica.org/cybersecurityinitiative/digichina/blog/full-translation-chinas-new-generation-artificial-intelligencedevelopment-plan-2017/
111
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Integration of AI and Real Economy further specifies that with high
integration and strong empowerment, AI is expected to boost the transition
of China’s economy from high-speed development to high-quality
development.443 Moreover, President Xi has frequently spoken of the
centrality of AI to the country’s overall economic development.444
Notably, the Chinese government is better prepared than many other
countries when it comes to the longer-term challenges of automation.445 For
instance, there are higher education courses that address the shortage in AI
skills and support the skilled labor required in the information age.446 China
has oriented its education system to prioritize high-proficiency in science,
technology, and engineering447 and has issued several policy directives
toward this end.448 According to China’s New Generation of AI
Development Report 2020, in 2019, 180 Chinese universities added AI,
undergraduate majors. Among them, 11 universities, including Peking
University, established new academic institutes designated for AI research.
Social Governance and Welfare
Social governance is another area in which AI is promoted as a
strategic opportunity for China. The Chinese authorities focus on AI as a
way of overcoming social problems and improving the welfare of citizens.449
Xi Jinping presided over the seventh meeting of the Central Committee for deepening
reform in an all-round way.
Keep a stable direction, highlight actual results, make all efforts to tackle difficulties, and
unswervingly promote the implementation of major reform measures People’s Daily,
http://paper.people.com.cn/rmrb/html/2019-03/20/nw.D110000renmrb_20190320_201.htm
444
Jeffrey Ding, Deciphering China’s AI dream. Centre for Governance of AI, Future of
Humanity Institute, University of Oxford, Oxford, https://www.fhi.ox.ac.uk/wpcontent/uploads/Deciphering_Chinas_AI-Dream.pdf. Elsa B Kania, China’s embrace of
AI: Enthusiasm and challenges, European Council on Foreign Relations (Nov. 6, 2018),
https://ecfr.eu/article/commentary_chinas_embrace_of_ai_enthusiasm_and_challenges/
445
The Automation Readiness Index: Who is Ready for the Coming Wave of Automation?
(2018) The Economist Intelligence Unit.
https://www.automationreadiness.eiu.com/static/download/PDF.pdf
446
Fang A (2019) Chinese colleges to offer AI major in challenge to US. Nikkei Asian
Review. https ://asia.nikkei.com/Business/China
-tech/Chinese-colleges-to-offer-AI-major-in-challenge-to-US
447
Is China ready for intelligent automation? (2018) China Power, Center for Strategic
and International Studies. https://chinapower.csis.org/china-intelligent-automation/
448
the National Medium- and Long-term Education Reform and Development Plan
(2010-2020)
449
Heilmann S (2017) Big data reshapes China’s approach to governance. Financial
Times https://www.ft.com/content/43170fd2-a46d-11e7-b797-b61809486fe2
443
112
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Specifically, in the healthcare reform,450 environmental protection451, the
administration of justice,452 and Social Credit System or Social Score.453
Another concrete example of how China is using AI in social governance
can be seen in the sphere of internal security and policing. China has been
at the forefront of the development of smart cities equipped with
surveillance technologies, such as facial recognition and cloud computing.
A recent proposal for the southwestern Chinese city of Chongqing would
put “AI in charge.”454 Today’s half of the world’s smart cities are located
within China. Thus, these ambitious goals exemplify the Chinese
government’s intent to rely on AI technology for social governance and also
for control of the behavior of its citizens.
Facial Recognition
There are many reports on China’s use of facial recognition
technology against ethnic minorities.455 The discriminatory ways in which
state organs, companies and academics have researched, developed and
implemented facial recognition in China would seem not to comply with the
OECD AI Principles or as the Governance Principles for the New
Generation Artificial Intelligence. The deployment of facial recognition has
also provoked opposition within China.456 This gap between stated ethical
principles and on-the-ground applications of AI demonstrate the weakness
Ho A (2018) AI can solve China’s doctor shortage. Here’s how. World Economic
Forum. https ://www.wefor um.org/agend a/2018/09/ai-can-solve -china -s-docto r-short
age-here-s-how/.
451
Kostka G, Zhang C (2018) Tightening the grip: environmental governance under Xi
Jinping. Environ Politics 27(5):769–781. https ://doi.org/10.1080/09644
016.2018.1491116; AI-powered waste management underway in China (2019) People’s
Daily Online. https ://en.peopl e.cn/n3/2019/0226/c9864 9-9549956.html
452
Finder S (2015) China’s master plan for remaking its courts. The Diplomat.
https://thediploma t.com/2015/03/chinas-master-plan-forremaking-its-courts/; Li A
(2016) Centralization of power in the pursuit of law-based governance: legal reform in
China under the Xi Administration. China Prospect 2016:2
453
Severine Arsene, China’s Social Credit System: A Chimera with Real Claws, Asie
Visions, 2019
https://www.ifri.org/sites/default/files/atoms/files/arsene_china_social_credit_system_20
19.pdf
454
Umberto Bacchi, 'I know your favorite drink': Chinese smart city to put AI in charge,
Reuters (Dec. 5, 2020), https://news.trust.org/item/20201203131328-4n7on
455
Joi Ito, My talk at the MIT-Harvard Conference on the Uyghur Human Rights Crisis
(May 2, 2019), https://joi.ito.com/weblog/2019/05/02/my-talk-at-the-.html.
456
Seungha Lee, Coming into Focus: China’s Facial Recognition Regulations, Center for
Strategic and International Studies (May 4, 2020), https://www.csis.org/blogs/trusteechina-hand/coming-focus-chinas-facial-recognition-regulations
450
113
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of unenforceable ethics statements. (See section below regarding AI and
Surveillance).
Medical AI
In China, the ultimate ambition of AI is to liberate data for public
health purposes. The AIDP, outlines the ambition to use AI to “strengthen
epidemic intelligence monitoring, prevention and control,” and to “achieve
breakthroughs in big data analysis, Internet of Things, and other key
technologies” for the purpose of strengthening intelligent health
management. The State Council’s 2016 official notice on the development
and use of big data in the healthcare sector, also explicitly states that health
and medical big data sets are a national resource and that their development
should be seen as a national priority to improve the nation’s health.457
However, there is a rising concern that relaxed privacy rules and the transfer
of personal data between government bodies will promote the collection and
aggregation of health data without the need for individual consent.458 Some
experts warn that this concept of public health and social welfare in China
will diminish already weak safeguards for personal data.
Use of AI in Covid-19 Response
In June 2020, the State Council released a White Paper, entitled
“Fighting COVID-19: China in Action,” which provides that China has
“fully utilized” artificial intelligence to not only research, analyze, and
forecast COVID-19 trends and developments, but also to track infected
persons, identify risk groups, and facilitate the resumption of normal
business operations.”459 During the pandemic, China has used AI for
surveillance of infected individuals and medical imaging. China also sought
to reduce human interaction by using computers and robots for various
purposes and have proven to be very effective in reducing exposure,
providing necessary services such as assistance for healthcare professionals,
Zhang Zhihao, China to focus on innovation to boost economy, lives, China Daily
(Sept. 1, 2018),
https://www.chinadaily.com.cn/a/201801/09/WS5a543bd5a31008cf16da5fa9.html
458
Huw Roberts, Josh Cowls, Jessica Morley, Mariarosaria Taddeo, Vincent Wang,
Luciano Floridi, The Chinese approach to artificial intelligence: an analysis of policy,
ethics, and regulation, AI and Society (June 17, 2020),
https://link.springer.com/article/10.1007/s00146-020-00992-2
459
“Full Text: Fighting COVID-19: China in Action,” Xinhua News via the State
Council, June 7, 2020, https://archive.vn/NYJQg.
457
114
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
improving efficiency in hospitals, and precautionary measures for returning
to normal business operations.460
AI Ethics
Despite widely reported cases of unethical use of AI in China, the
Chinese authorities, private companies and academia have been active in
the global trend towards formulating and issuing statements on AI ethics.
The AIDP goes as far as to outline a specific desire for China to become a
world leader in defining ethical norms and standards for AI.461 There has
been a recent wave of attempts to define ethical standards by both
government bodies and private companies.
In 2017, China’s Artificial Intelligence Industry Alliance (AIIA),
released a draft “joint pledge” on self-discipline in the artificial
intelligence (AI) industry - emphasizing AI ethics, safety, standardization,
and international engagement.462
In 2019, the Beijing Academy of Artificial Intelligence (BAAI)
released the Beijing AI Principles463 to be followed for the research and
development, use, and governance of AI. The Beijing Principles are
centered around doing good for humanity, using AI “properly,” and having
the foresight to predict and adapt to future threats. But just like other
principles presented, they are still very vague.
In line with these principles, Governance Principles for Developing
Responsible Artificial Intelligence464 prepared in 2019, by the National New
Generation Artificial Intelligence Governance Expert Committee that was
established by China’s Ministry of Science and Technology. This document
outlines eight principles for the governance of AI: harmony and
friendliness, fairness and justice, inclusivity and sharing, respect for human
rights and privacy, security, shared responsibility, open collaboration and
Emily Weinstain, China’s use of AI in its Covid-19 Response, the Center for Security
and Emerging Technology (Aug. 2020), https://cset.georgetown.edu/research/chinas-useof-ai-in-its-covid-19-response/
461
China's 'New Generation Artificial Intelligence Development Plan' (July 20, 2017)
(English translation), https://www.newamerica.org/cybersecurityinitiative/digichina/blog/full-translation-chinas-new-generation-artificial-intelligencedevelopment-plan-2017/
462
Chinese AI Alliance Drafts Self-Discipline 'Joint Pledge' (June 17, 2019) (English
translation) https://www.newamerica.org/cybersecurityinitiative/digichina/blog/translation-chinese-ai-alliance-drafts-self-discipline-joint-pledge/
463
Beijing Principles, https://www.baai.ac.cn/news/beijing-ai-principles-en.html
464
Chinese Expert Group Offers 'Governance Principles' for 'Responsible AI' (June 17,
2019) (English translation), https://www.newamerica.org/cybersecurityinitiative/digichina/blog/translation-chinese-expert-group-offers-governance-principlesresponsible-ai/
460
115
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
agility to deal with new and emerging risks. Above all else, AI development
should begin from enhancing the common well-being of humanity, states
the document.
Another important document is a white paper on AI standards465
released in 2018 by the Standardization Administration of the People’s
Republic of China, the national level body responsible for developing
technical standards. Three key principles for setting the ethical requirements
of AI technologies are (1) the ultimate goal of AI is to benefit human
welfare; (2) transparency and the need to establish accountability as a
requirement for both the development and the deployment of AI systems
and solutions; (3) protection of intellectual property.
It is apparent that these principles bear some similarity to the OECD
AI Principles. Nevertheless, the principles established in China place a
greater emphasis on social responsibility, community relations, national
security and economic growth, with relatively less focus on individual
rights. However, establishing ethical AI principles can be viewed as a first
step and a signal that China wishes to become engaged in a dialogue with
international partners.
AI and Surveillance
As early as the 2008 Beijing Olympics, China began to deploy new
technologies for mass surveillance.466 China put in place more than two
million CCTV cameras in Shenzen, making it the most watched city in the
world.467 In recent years the techniques for mass surveillance have expanded
rapidly, most notably in Shenzen, also to oversee the Muslim minority
group the Uyghurs, and in Hong Kong. Modern systems for mass
surveillance rely on AI techniques for such as activities as facial
recognition, communications analysis and location tracking. As one
industry publication has reported, “In the world of surveillance, no country
invests more in its AI-fueled startups and growth-stage businesses than
China. And no technology epitomises this investment more than facial
Jeffrey Ding and Paul Triolo, Translation: excerpts from China’s ‘White Paper on
Artificial Intelligence Standardization,’ New America (June 20, 2018), https
://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-excerptschinas-white -paper-artificial-intelligence-standardization /
466
EPIC/Privacy International, Privacy and Human Rights: An International Survey of
Privacy Laws and Developments (2006) (Report on People’s Republic of China),
http://www.worldlii.org/int/journals/EPICPrivHR/2006/PHR2006-People_s.html;
467
Naomi Wolf, China's All-Seeing Eye With the help of U.S. defense contractors, China
is building the prototype for a high-tech police state. It is ready for export, Rolling Stone
(May 15, 2018), https://www.commondreams.org/views/2008/05/15/chinas-all-seeingeye.
465
116
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
recognition—a technology that courts more controversy than almost any
other.”468 Forbes continues, “But a thriving domestic tech base has done
nothing to quell the concerns of citizens. China is held up as a Big Brother
example of what should be avoided by campaigners in the West, but that
doesn't help people living in China.”
In September 2019, China’s information-technology ministry
announced that telecom carriers must scan the face of anyone applying for
mobile and internet service.469 There are over 850 million mobile Internet
users in China. Meanwhile, the Hong Kong government invoked emergency
powers in October 2019 to ban demonstrators from wearing face masks.470
Protests in Hong Kong over the use of facial surveillance are
widespread. Umbrellas once used to deflect pepper spray, are now deployed
to shield protester activities from the digital eyes of cameras.471 It is notable
that the battle over the use of facial surveillance in Hong Kong began with
widespread public protests about a national security law that extended
police authority over the semi-autonomous region.472 According to the AP,
“Young Hong Kong residents protesting a proposed extradition law that
would allow suspects to be sent to China for trial are seeking to safeguard
their identities from potential retaliation by authorities employing mass data
collection and sophisticated facial recognition technology.”473
China is also exporting the model of mass surveillance by facial
recognition to other parts of the world. A detailed report, published in The
Atlantic in September 2020, stated that “Xi Jinping is using artificial
intelligence to enhance his government’s totalitarian control—and he’s
Zak Doffman, Hong Kong Exposes Both Sides Of China's Relentless Facial
Recognition Machine (Aug. 26, 2019),
https://www.forbes.com/sites/zakdoffman/2019/08/26/hong-kong-exposes-both-sides-ofchinas-relentless-facial-recognition-machine/
469
Jane Li, Getting a new mobile number in China will involve a facial-recognition test,
Quartz (Oct. 3, 2019), https://qz.com/1720832/china-introduces-facial-recognition-stepto-get-new-mobile-number/
470
Ilara Maria Sala, Hong Kong is turning to a 1922 law that was used to quell a
seamen’s strike to ban face masks, Quartz (Oct. 4, 2019), https://qz.com/1721951/antimask-law-the-1922-origins-of-hong-kongs-emergency-powers/
471
Paul Mozur and Lin Qiqing, Hong Kong Takes Symbolic Stand Against China’s HighTech Controls, New York Times (Oct. 3, 2019),
https://www.nytimes.com/2019/10/03/technology/hong-kong-china-techsurveillance.html
472
BBC, Hong Kong security law: What is it and is it worrying? (June 30, 2020),
https://www.bbc.com/news/world-asia-china-52765838
473
Christopher Bodeen, Hong Kong protesters wary of Chinese surveillance technology
(June 13, 2019), https://apnews.com/article/028636932a874675a3a5749b7a533969
468
117
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
exporting this technology to regimes around the globe.”474 According to The
Atlantic, “Xi’s pronouncements on AI have a sinister edge. Artificial
intelligence has applications in nearly every human domain, from the
instant translation of spoken language to early viral-outbreak detection. But
Xi also wants to use AI’s awesome analytical powers to push China to the
cutting edge of surveillance. He wants to build an all-seeing digital system
of social control, patrolled by precog algorithms that identify potential
dissenters in real time.”
In September 2020, the United States State Department issued
voluntary guidelines for American companies “to prevent their products or
services . . . from being misused by government end-users to commit human
rights abuses.”475 The report comes amid growing concern that China is
rapidly exporting its own surveillance capabilities to authoritarian regimes
around the world, as part of its Belt and Road Initiative (BRI).476 But the
Washington Post recently highlighted the ongoing role of US-made
technology in the sweeping surveillance of China, and notably the Uighur
Muslim minority.477 The Washington Post explained that “the aim is to
monitor cars, phones and faces — putting together patterns of behavior for
‘predictive policing’ that justifies snatching people off the street for
imprisonment or so-called reeducation. This complex opened four years
ago, and it operates on the power of chips manufactured by U.S.
supercomputer companies Intel and Nvidia.”
Ross Anderson, The Panopticon is Already Here, The Atlantic (Sept. 2020),
https://www.theatlantic.com/magazine/archive/2020/09/china-ai-surveillance/614197/
475
U.S. Department of State, Bureau of Democracy, Human Rights, and Labor, U.S.
Department of State Guidance on Implementing the "UN Guiding Principles" for
Transactions Linked to Foreign Government End-Users for Products or Services with
Surveillance Capabilities (Sept. 30, 2020), https://www.state.gov/key-topics-bureau-ofdemocracy-human-rights-and-labor/due-diligence-guidance/
476
Abhijnan Rej, US Issues Human Rights Guidelines for Exporters of Surveillance Tech:
The directions to American businesses come amid growing concern around China’s
export of advanced mass-surveillance capabilities to more than 60 countries, The
Diplomat (Oct. 2, 2020), https://thediplomat.com/2020/10/us-issues-human-rightsguidelines-for-exporters-of-surveillance-tech/
477
The Washington Post, Editorial, U.S.-made technologies are aiding China’s
surveillance of Uighurs. How should Washington respond? (Nov. 28, 2020),
https://www.washingtonpost.com/opinions/us-made-technologies-are-aiding-chinassurveillance-of-uighurs-how-should-washington-respond/2020/11/26/0218bbb4-2dc911eb-bae0-50bb17126614_story.html
474
118
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Post editorial followed a New York Times investigation which
found extensive involvement by U.S. firms in the Chinese surveillance
industry.478
Public Opinion
There is growing concern in China about the misuse of personal data
and the risk of data breaches. In a 2018 survey by the Internet Society of
China, 54% of respondents stating that they considered the problem of
personal data breaches as ‘severe.’479 The World Economic Forum suggest
that 2018-2019 “could be viewed as the time when the Chinese public woke
up to privacy.” According to the WEF, a controversy arose in 2019 when
the Zao app, using AI and machine learning techniques, allowed users to
swap faces with celebrities in movies or TV shows.480 “It went viral as a tool
for creating deepfakes, but concerns soon arose as people noticed that Zao’s
user agreement gave the app the global rights to use any image or video
created on the platform for free.” The company later clarified that the app
would not store any user’s facial information. Chinese consumers also
challenged Alibaba when they learned that they had been enrolled in a credit
scoring system by default and without consent. “Under pressure, Alibaba
apologized.”
Data Protection
China adopted the Personal Information Protection Law (PIPL) in
August, 2021.481 It took effect in November 2021. The aims of the PIPL are
to
1. to protect the rights and interests of individuals (为了保护个人信
息权益),
Paul Mazur and Don Clark, China’s Surveillance State Sucks Up Data. U.S. Tech Is
Key to Sorting It: Intel and Nvidia chips power a supercomputing center that tracks
people in a place where government suppresses minorities, raising questions about the
tech industry’s responsibility (Nov. 22, 2020),
https://www.nytimes.com/2020/11/22/technology/china-intel-nvidia-xinjiang.html
479
Technology Review, China’s citizens do care about their data privacy, actually, (Mar.
28, 2018), https://www.technologyreview.com/2018/03/28/67113/chinas-citizens-docare-about-their-data-privacy-actually/
480
World Economic Forum, China is waking up to data protection and privacy. Here's
why that matters (Nov. 12, 2019), https://www.weforum.org/agenda/2019/11/china-dataprivacy-laws-guideline/
481
Stanford University, Digichina, Translation: Personal Information Protection Law of
the People’s Republic of China – Effective Nov. 1, 2021(Sept. 7, 2021),
https://digichina.stanford.edu/work/translation-personal-information-protection-law-ofthe-peoples-republic-of-china-effective-nov-1-2021/
478
119
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
2. to regulate personal information processing activities (规范个人信
息处理活动),
3. to safeguard the lawful and “orderly flow” of data (保障个人信息
依法有序自由流动),
4. to facilitate reasonable use of personal information (促进个人信息
合理利用) (Art. 1).482
Key provisions include Article 7 (principle of transparency), Article
24 (algorithmic transparency including explanation, with a new clause to
limit price discrimination) Article 28 (establishing the use of personal
identity recognition equipment in public venues), Articles 55 and 56 (the
renamed “personal information protection impact assessments”), and
Article 62 (coordination of AI by state cybersecurity authority). Other
provisions align with the GDPR, such as the definition of personal data,
deidentification, and anonymization, as well as the need for a legal basis to
process personal data. Article 34 extends the PIPL obligations to state
authorities. And provisions on data minimization and purpose specification
limite personal data available to the state for public security purposes. Other
provisions limit the export of personal data out of China (Article 36). And
other provisions limit the use of children’s data (Articles 28 and 31).
Fundamental Rights & OECD AI Principles
China has endorsed Universal Declaration of Human Rights and
G20 AI Principles. As a party to the UDHR, China shall recognize “the
inherent dignity” of all human beings and to secure their fundamental rights
to “privacy.” Privacy rights are guaranteed to Chinese citizens under the
Constitution. However, Article 40 of the Chinese constitution justifies the
invasion of privacy “to meet the needs of State security.” Furthermore, the
Constitution is regarded as irrelevant, as there is neither a constitutional
court nor any possibility to assert constitutional rights.483 Relatedly,
problematic exemptions for the collection and use of data, when it is related
Future of Privacy Forum, China’s New Comprehensive Data Protection Law: Context,
Stated Objectives, Key Provisions (Aug. 20, 2021), https://fpf.org/blog/chinas-newcomprehensive-data-protection-law-context-stated-objectives-key-provisions/
483
Greenleaf, Data Privacy (n 3) 196 f; Wang, ‘Redefining Privacy’ (n 11) 110; Ma and
Roth (n 2) 355
482
120
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
to security, health, or the flexibly interpretable “significant public interests”
484
contribute to weak data protection in China.
These exemptions are also behind the big data collection and mass
surveillance system, the Integrated Joint Operations Platform (IJOP),485
used in Xinjiang for monitoring minorities. Another example is Social
Credit System, a system that collects all kinds of data about citizens and
companies, sorts, analyses, evaluates, interprets and implements actions
based on it. Thus, the strength of privacy protection in China is likely to be
determined by the government’s decisions surrounding data collection and
usage, rather than legal and practical constraints.486 Moreover, policies and
administrative decisions on both central and provincial levels often
contradict the legal protection487 as administrative agencies may ignore the
law on the basis of party policy, morality, public opinion, or other political
considerations.488
Freedom House is extremely critical of China’s failue to protect
rights and civil liberties in the most recent survey of country practices.489
China score 9/100 and was designated “no free.” According to Freedom
House, “China’s authoritarian regime has become increasingly repressive
in recent years. The ruling Chinese Communist Party (CCP) is tightening
its control over the state bureaucracy, the media, online speech, religious
groups, universities, businesses, and civil society associations, and it has
undermined its own already modest rule-of-law reforms.”
Evaluation
China has emerged as one of the first AI superpowers and has an
ambitious plan of leading the world in AI by 2030. In addition to the G20 AI
Principles, China has endorsed important principles on AI and ethics and
recently announced a new law on data protection and a regulation of
Sacks S (2018) New China Data Privacy Standard Looks More Far-Reaching than
GDPR. Center for Strategic and International Studies. https://www.csis.org/analysis/newchina-data-privacystandard-looks-more-far-reaching-gdpr.
485
China’s Big Brother App, Human Rights Watch
https://www.hrw.org/news/2019/05/01/interview-chinas-big-brother-app
486
Huw Roberts, Josh Cowls, Jessica Morley, Mariarosaria Taddeo, Vincent Wang,
Luciano Floridi, The Chinese approach to artificial intelligence: an analysis of policy,
ethics, and regulation, May 2020
487
Aktas, I. (2015). Uighur Separatism and Human Rights: A Contextual Analysis. In M.
Kosmala-Mozlowska (Ed.), Democracy and Human Rights in East Asia and Beyond –
Critical Essays. Warsaw: Collegium Civitas Press.
488
Wang J, Liu S (2019) Ordering power under the party: a relational approach to law
and politics in China. Asian J Law Soc 6(1):1–18. https ://doi.org/10.1017/als.2018.40
489
Freedom House, Freedom in the World 2021 – China (2021),
https://freedomhouse.org/country/china/freedom-world/2021
484
121
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
recommendation algorithms. However, China’s use of its AI against ethnic
minorities and protesters in Hong Kong, as well as a means to score citizens
for their alliance with the state (a practice recently banned by the UNESCO
Recommendation on AI Ethics), is the source of widespread fear and
skepticism. There is also a concern about the development of lethal
autonomous weapons. As China is now rapidly deploying AI systems, there is
an urgent need to assess China’s actual practices against global standards for
human-centric AI.
122
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Colombia
National AI Strategy
At the launch of Colombia’s draft Ethical Framework for AI,
President Iván Duque Márquez highlighted the importance of an ethical
framework on AI with ‘tools that strengthen the principle of democracy,
free competition and equity.’490 It is with these considerations along with
ethics, social aspects, economic development and technological concepts
that Colombia developed its AI policy.
The National Planning Department of Colombia, through the
National Development Plan for 2018-2022 was the first to encourage the
inclusion of emerging technologies of the Fourth Industrial Revolution,
such as AI, the Internet of Things (IoT) and robotics in the digital
transformation of national entities and strategies across all sectors.491
Following this, the National Planning Department, the Ministry of
Information and Communications Technology (MinTIC), and the Office of
the President launched the country’s National AI Strategy, titled the
National Policy for Digital Transformation and AI (CONPES No. 3975).492
This strategy introduced enabling social and economic conditions for the
development of AI, with a framework of flexible principles and guidelines
instead of a rule-based structure. While the strategy acknowledges its
adoption of the OECD AI principles, it sets out 14 additional principles with
a focus on four aspects namely inclusive growth, sustainable development,
and well-being; building human capacity and preparing for labor market
transition; fostering a digital ecosystem for AI; and providing an enabling
policy environment for AI. The strategy also endorses the adoption of an
ethical framework for the development of responsible and inclusive AI,
utilization of data for the development of AI, and establishment of a market
that uses AI productively and competitively.
To ensure the sustainable execution and continuity of AI public
policy, various entities were created to coordinate the development and
Office of the President of the Republic of Colombia, With the Ethical Framework for
Artificial Intelligence, Colombia is at the forefront in Latin America: Duque, (Nov. 25,
2020), https://idm.presidencia.gov.co/prensa/con-el-marco-etico-de-inteligenciaartificial-colombia-se-pone-a-la-201125
491
Congress of the Republic of Colombia, National Development Plan 2018-2022 ‘Pact
for Colombia, Pact for Equity’, (2019), https://www.dnp.gov.co/DNPN/Paginas/PlanNacional-de-Desarrollo.aspx
492
Ministry of Information and Communications Technology (MinTIC), National
Planning Department and the Office of the President, National Policy for Digital
Transformation and Artificial Intelligence (CONPES No. 3975), (Nov. 8, 2019),
https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/3975.pdf
490
123
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
implementation of National AI Strategy and other AI policies in the country.
This includes the Presidential Advisory for Economic Affairs and Digital
Transformation (DAPRE), the AI Expert Mission and the International
Council for AI. DAPRE coordinates the work of government functionaries
in implementing digital transformation through various systems including
AI based ones, while advising the government on the development of a
digital ecosystem, along with the formulation and implementation of related
policy.493 The AI expert mission or task force, based on the entity in the
United States and United Kingdom serves as a bridge between regulators
and experts. It includes experts from various professions who advise the
government on policy formulation and assist them in developing a
prospective roadmap for the implementation of AI policy, combining their
technical and comprehensive vision.494 The AI expert mission was launched
with 10 experts. The International Council of AI was proposed to integrate
international experts in the implementation and deployment of Colombia’s
National AI systems.495 The Council consists of six government officials
and nine international experts as permanent guests to collectively analyze
and present policy proposals that will impact the development and
deployment of AI. The Council will also review and guide the
implementation of AI policy, while studying Colombia’s position in
international AI indices to determine points of improvement that can be
integrated into a roadmap for the future of AI.
There are various policy intelligence tools in place to monitor the
implementation of AI policy. For instance, SisCONPES monitors the
implementation of each action line of the National AI Strategy, by reporting
to implementing authorities on progress made and obstacles that arise.496
Office of the President of the Republic of Colombia, Decree No. 1784 by which the
structure of the Administrative Department of the Presidency of the Republic is modified,
(Dec. 2019), https://bit.ly/3EWfUDk
494
Office of the President of the Republic of Colombia and The Development Bank of
Latin America (CAF), Task force for the development and implementation of Artificial
Intelligence in Colombia, (Nov. 2020),
https://dapre.presidencia.gov.co/AtencionCiudadana/Documents/TASK-FORCE-paradesarrollo-implementacion-Colombia-propuesta-201120.pdf
495
Office of the President of the Republic of Colombia and The Development Bank of
Latin America (CAF), International Council of Artificial Intelligence for Colombia,
(March 2021), https://dapre.presidencia.gov.co/TD/INTERNATIONAL-COUNCIL-OFARTIFICIAL-INTELLIGENCE-FOR-COLOMBIA.pdf
496
Office of the President of the Republic of Colombia, Preparation and monitoring of
CONPES documents, https://bit.ly/3klbcqX
493
124
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Ethical Framework for Artificial Intelligence
In Colombia, it was believed that concerns arising from the
implementation of emerging technology, such as AI can only be answered
with an ethical framework.497 Thus, an ethical framework was developed
around ethical principles that can serve as a criterion for evaluating the
different uses and challenges that arise in this respect.
The Ethical Framework for AI was developed as a tool, that can be
applied to different sectors taking into consideration the diversity of
interests and opinions around the use of AI. The ten principles provided by
the framework to guide the design, development, implementation and
evaluation of AI systems include transparency, explainability, privacy,
human oversight over AI decisions, security, responsibility, nondiscrimination, inclusion, prevalence of the rights of children and
adolescents, and social benefit.498
The framework also proposes an ethical algorithm register in which
entities periodically report what their AI project is about, how they are
implementing the AI ethics principles, and the ethical risks to the use of AI
in their project. The register allows for the monitoring of progress in the
implementation of AI principles and reinforces citizen participation by
inviting their comments or questions on policies, good practices and
projects related to AI.499 The ethical algorithm register of Colombia is based
on the models of Amsterdam and Helsinki.500
The Development Bank of Latin America (CAF), Why has Colombia positioned itself
as a regional leader on Artificial Intelligence, (Sept. 14, 2021),
https://www.caf.com/es/conocimiento/visiones/2021/09/por-que-colombia-se-haposicionado-como-lider-regional-en-inteligencia-artificial/
498
Office of the President of the Republic of Colombia and The Development Bank of
Latin America (CAF), Ethical Framework for Artificial Intelligence in Colombia,
(Aug.2020), https://cyber.harvard.edu/sites/default/files/202012/Colombia_AI_Ethical_Framework.pdf
499
Office of the President of the Republic of Colombia and The Development Bank of
Latin America (CAF), Ethical Framework for Artificial Intelligence in Colombia, (May
2021), https://bit.ly/3bRiXAm
500
Center for Technology and Society Studies (CETyS) of the University of San Andrés,
The Colombian Case: Adopting collaborative governance as a path for implementing
ethical artificial intelligence, (2021),
https://repositorio.udesa.edu.ar/jspui/bitstream/10908/18743/1/The%20Colombian%20ca
se%20adopting%20collaborative%20governance%20as%20a%20path%20for%20imple
menting%20ethical%20artificial%20intelligence%20.pdf
497
125
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Regulatory Sandboxes and Beaches
Colombia has adopted a smart regulation approach to AI policy
through regulatory sandboxes and beaches.501 This controlled environment
was set up to experiment and test AI systems in the local context, to identify
technical and governance flaws while promoting innovation.502
At first, a Draft Model Concept for the Design of Regulatory
Sandboxes and Beaches in AI was published in August 2020 with public
comment received from various stakeholders.503 The purpose of this policy
was to understand technology before trying to regulate it, by balancing
precaution with experimentation and learning. The document suggests a
process of implementation that includes (1) defining a policy leadership to
implement public policy; (2) defining emerging technologies and
preliminary problems to be addressed; (3) completing a regulatory mapping
of the impacted sector; (4) selecting a public entity to perform inspection
and surveillance functions; (5) capacity building and training; (6) creating
working groups; (7) designing a risk model and defining possible risks; (8)
setting out a selection criteria for the risks; (9) designing the sandbox; (10)
sharing the project for comment; (11) publishing and implementing; and
(12) reporting on the findings and evidence.504
Later in 2020, a policy proposing a regulatory sandbox for Privacy
and AI was open for public comments until November 30 by DAPRE along
with the Superintendence of Industry and Commerce. The regulatory
sandbox here is meant to be preventive, so AI systems related to ecommerce, advertising and marketing protect personal data from the stage
of design to execution, using tools like privacy impact assessments and
Regulatory sandboxes are a testbed for selected AI projects, where the regulatory
framework is relaxed with some laws and regulations set aside while entities test their
projects. Regulatory beaches are similar to regulatory sandboxes, but they are wider in
scope. They allow a larger number of companies and sectors to participate in regulatory
experimentation during extended amounts of time even longer than a year, with the goal
of resolving industry problems.
502
The Global Policy Journal, The Colombia Case: A New Path for Developing
Countries Addressing the Risks of Artificial Intelligence, (May 2021),
https://bit.ly/3jVF2Cm
503
Presidential Advisory for Economic Affairs and Digital Transformation (DAPRE),
Draft Model Concept for the Design of Regulatory Sandboxes & Beaches in AI:
Document for discussion, https://bit.ly/3Gbd7XD
504
Armando Guío Español, Model Concept for the Design of Regulatory Sandboxes &
Beaches in AI, (August 2020),
https://dapre.presidencia.gov.co/AtencionCiudadana/DocumentosConsulta/consulta200820-MODELO-CONCEPTUAL-DISENO-REGULATORY-SANDBOXESBEACHES-IA.pdf
501
126
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
privacy by design. This document proposes (1) criteria to ensure
compliance with regulation on data processing in AI; (2) proper processing
of personal data in all stages of an AI project; (3) creation of AI products
that respects individual rights to personal data; (4) advice to companies on
the protection of personal data in AI systems; (5) adoption of a preventive
approach to protect human rights in AI projects; (6) suggestion of
amendments or modifications to Colombian regulations on technological
advances.505
Public Participation
The Development Bank of Latin America (CAF) with the authorship
of international expert, Armando Guío Español has developed AI policy and
implementation documents, including the ethical framework for AI, a
model concept for the design of regulatory sandboxes and beaches in AI, a
data governance model, a task force on the development and
implementation of AI, and the outline of an international council for the
implementation of AI policy.506
Draft AI policies and legislations of Colombia have been opened for
public comment from academia, national, regional and international civil
society actors, intergovernmental organizations and the private sector.
These consultations have taken on various forms. For example, an Expert
Roundtable on Colombia’s Draft AI Ethical Framework was organized by
the Berkman Klein Center for Internet and Society at Harvard University.507
Additionally, the Ethical Framework for AI and its ethical algorithm
register promote public participation in the implementation of AI ethics
principles, as the registry is publicly accessible and includes an interactive
channel where citizens can ask questions or post comments on the ethical
implementation of AI.508 The presence of regulatory sandboxes and beaches
also allow participation of the private sector and academic institutions in the
development AI technology.
Presidential Advisory for Economic Affairs and Digital Transformation (DAPRE),
and the Superintendence of Industry and Commerce, Sandbox on privacy by design and
by default in Artificial Intelligence projects, (2020),
https://www.sic.gov.co/sites/default/files/normatividad/112020/031120_Sandbox-sobreprivacidad-desde-el-diseno-y-por-defecto.pdf
506
Development Bank of Latin America, Experience AI: Data and Artificial Intelligence
in the Public Sector, (2021), https://scioteca.caf.com/handle/123456789/1793
507
Berkman Klein Center for Internet and Society at Harvard University, Summary
Report of Expert Roundtable on Colombia’s Draft AI Ethical Framework, (Jan. 2021),
https://cyber.harvard.edu/sites/default/files/2021-01/Colombia_Roundtable_Report.pdf
508
Office of the President of the Republic of Colombia and The Development Bank of
Latin America (CAF), Ethical Framework for Artificial Intelligence in Colombia, (May
2021), https://bit.ly/3bRiXAm
505
127
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Research & Development
On 26 July 2019, MinTIC released an Information and
Communication Technology (ICT) Plan for 2018-2022 titled ‘The Digital
Future is for Everyone’. Highlighting the need to develop human capital,
the ICT Plan proposes an AI Center of Excellence & Appropriation to
generate innovative proposals that benefit the national system and serve as
reference internationally.509
The ICT plan also emphasizes the importance of removing obstacles
to the use of technology for digital businesses. In this respect, on 29 April
2019, the World Economic Centre for the Fourth Industrial Revolution was
launched in Medellín, bringing together governments, the private sector,
civil society organizations, academia, and tech experts from across the
globe.510 Together, these actors collaborate in designing, testing, and
developing projects that prioritize policy and tech innovation on AI, the
Internet of Things, blockchain, and robotics.511
To boost innovative AI research, Colombia launched a start-up
incubator and accelerator, C-Emprende. In addition to scaling enterprises
and mobilizing resources, C-Emprende facilitates the exchange of
knowledge between national and international academia, private sector
actors, investors, and government representatives.512
Privacy and Data Protection
Data Protection in Colombia is governed by Article 15 of the
Political Constitution.513 Additionally, Colombia regulates financial, credit,
commercial and services information514 and personal data processing and
databases.515
Ministry of Information and Communications Technology (MinTIC), ICT Plan 20182022 ‘The Digital Future is For Everyone’, (2019),
https://micrositios.mintic.gov.co/plan_tic_2018_2022/pdf/plan_tic_2018_2022_2019112
1.pdf
510
World Economic Forum, Centre for the Fourth Industrial Revolution,
https://www.weforum.org/centre-for-the-fourth-industrial-revolution/affiliate-centres
511
Centre for the Fourth Industrial Revolution Colombia, Homepage, https://c4ir.co
512
C-Emprende, Homepage, https://innpulsacolombia.com/cemprende/quienes-somos
513
Congress of the Republic of Colombia, Political Constitution of Colombia, (1991),
https://www.constituteproject.org/constitution/Colombia_2015.pdf?lang=en
514
Congress of the Republic of Colombia, Law 1266 of 2008 on the processing of
financial data, credit records, and commercial information collected in Colombia or
abroad, (Dec. 31, 2008)
515
Congress of the Republic of Colombia, Law 1581 of 2012 on the protection of
personal data (Oct. 17, 2012),
https://www.funcionpublica.gov.co/eva/gestornormativo/norma.php?i=49981#:~:text=La
%20presente%20ley%20tiene%20por,el%20art%C3%ADculo%2015%20de%20la
509
128
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
For the implementation and monitoring of regulations around
privacy, the Personal Data Authority (DPA) was established under the
Division of Data Protection of the Superintendence of Industry and
Commerce, according to Article 19 of Law 1581. This authority functions
as an oversight body, providing instructions and setting mandates, along
with receiving complaints on the handling of data.
On automated decision-making, Law 1581 establishes that for
personal data processing must be for a legitimate purpose under the
Constitution and other laws, it must be notified to the subject, and the
purpose must be specific. As a result, if automated decision-making is the
purpose of processing data then (1) legitimate as per the Political
Constitution and other laws of Colombia, (2) specific in purpose, and (3)
informed to the data subject.516
Secondary decrees, decisions and regulations provide a better
understanding of data protection, AI, particularly the application of
automated mechanisms to databases. Databases have been defined under
Article 3 of Law 1581 as an organized set of personal data which is treated
the same. Decree 886 that regulates Article 25 of Law 1581 explains this
further by stating that, when automation is applied to databases containing
personal data, it should be registered in the public directory of databases
called the National Register of Databases.517 Drawing from the necessity to
register information from databases and the protection of the right to habeas
data,518 the Constitutional Court concluded that the administrator of a
database has specific obligations regarding the quality of data being
transmitted and allows data subjects to authorize how their information in
an automated system is handled.519 Additionally, Article 26 of Decree 1377
establishes the principle of proven liability, according to which those
responsible for handling of personal data have an obligation to prove that
Dejusticia, Accountability of Google and other businesses in Colombia: Personal
Data Protection in the Digital Age, (2019), https://www.dejusticia.org/wpcontent/uploads/2019/01/Accountability-of-Google-and-other-Businesses-inColombia.pdf
517
Office of the President of the Republic of Colombia, Decree 886 of 2014, (2014),
https://www.suin-juriscol.gov.co/viewDocument.asp?id=1184150
518
Habeas data is a fundamental right and tool to provide legal protection to owners of
personal data, particularly when faced with undue or illegal processing of their personal
data by databases, or public or private registries
519
Constitutional Court of Colombia, Sentence C-1011/08: Habeas Data in statutory law
and the handling of information contained in personal databases (2008),
https://www.corteconstitucional.gov.co/relatoria/2008/C-1011-08.htm
516
129
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
they have taken sufficient and effective measures to abide by regulations,
even when the data is processed by an automated method.520
In July 2020, through Resolution 38281, the Superintendence of
Industry and Commerce concluded that Law 1581 is thematically and
technologically neutral.521 Thus, the provisions of Law 1581 apply to
processing of any data regardless of techniques or technologies used.
Protection of personal data extends to all techniques and tools, including AI
in its use for predictive dialing, robocalls and nuisance calls. Nelson
Remolina, the Superintendent for the Protection of Personal Data elaborated
on this by stating that, while Colombian law allows for the creation, design,
and use of technological innovations to process data, it must be done in a
way that respects the legal system by complying with all rules to process
personal data.522
Colombian privacy law differs in scope from GDPR, since it only
applies to data processing carried out by data processors and data controllers
within the country or to those who have a legal obligation under
international laws and treaties. Unlike GDPR, Colombian privacy law also
does not regulate the right to be forgotten and does not set out conditions
under which data profiling is allowed. This could allow for exploitation of
gaps in the law, providing lesser protection to individuals and their personal
data.
Colombia was a primary sponsor of the 2020 GPA Resolution on
Accountability in the Development and Use of Artificial Intelligence.523
Data Infrastructure
Colombia facilitates data access for those designing and developing
AI systems, achieved by removing unnecessary and unjustified barriers to
access information. To facilitate such data access and to generate social and
Office of the President of the Republic of Colombia, Decree 1377 of 2013, (2013),
http://www.suin-juriscol.gov.co/viewDocument.asp?ruta=Decretos/1276081
521
Superintendence of Industry and Commerce, Resolution 38281, (July 14, 2020),
https://www.sic.gov.co/sites/default/files/files/Proteccion_Datos/Res%2038281%20del%
2014VII2020%20Mervicol%20marcadores%20predictivos%20robocalls%20IA.pdf
522
Ibero-American Data Protection Network (RIPD), Colombian data protection
authority concluded that predictive dialing, robocalls and artificial intelligence must
comply with regulation regarding the processing of personal data, (20 July 2020),
https://www.redipd.org/en/news/colombian-data-protection-authority-concludedpredictive-dialing-robocalls-and-artificial
523
Global Privacy Assembly, Resolution on Accountability in the Development and Use
of Artificial Intelligence (Oct. 2020), https://globalprivacyassembly.org/wpcontent/uploads/2020/10/FINAL-GPA-Resolution-on-Accountability-in-theDevelopment-and-Use-of-AI-EN-1.pdf
520
130
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
economic well-being, the Colombian government has developed data
infrastructure policies with a shared dynamic and standardized resources
across different actors. Thus, data infrastructure is used to strengthen
institutional capacity to provide better quality services to citizens, to include
citizens and the private sector in data governance, to drive innovation in
governance, and to guide decision-making.
The first policy document on data infrastructure is the National
Policy on Data Exploitation or CONPES No. 3920 of 2018, developed by
the National Council on Economic and Social Policy, the National Planning
Department, and the Office of the President.524 This policy uses data within
a legal, ethical, and institutional framework to generate social and economic
value; to increase the availability and interoperability of government data;
to promote data culture in public entities, academia and the private sector;
to promote data ethics and AI; and to provide test environments through
data sandboxes, sandboxes on privacy and AI, and conceptual models for
regulatory sandboxes and beaches in AI. To achieve this target, CONPES
3920 sets out 45 action steps with indicators, responsible parties, budgets
and a timeline.
The second policy document is the National Data Infrastructure Plan
(PNID) developed by MinTIC, the National Planning Department, and the
Office of the President, with the support of the World Economic Centre for
the Fourth Industrial Revolution.525 The draft of PNID was shared for public
comment until 17 September 2021. PNID presents an approach to data as
infrastructure, defines the components of data infrastructure, and provides
a roadmap with concrete actions to implement data infrastructure in the
country. This roadmap identifies 6 elements including governance, data,
data leveraging, infrastructure interoperability, data security and privacy, as
well as technical and technological input for data management.
For the successful integration of PNID into the data regulation
ecosystem, between 2022 and 2025, the government intends to create
guidelines for PNID; to identify priority data and create guidelines to ensure
data quality; to develop a data infrastructure governance model; to identify
National Council on Economic and Social Policy, the National Planning Department,
and the Office of the President, National Policy on Data Exploitation (Big Data), (Apr.
17, 2018),
https://www.mindeporte.gov.co/recursos_user/2020/Jur%C3%ADdica/Julio/Conpes_392
0.pdf
525
Ministry of Information and Communications Technology (MinTIC), National
Planning Department and the Office of the President, National Data Infrastructure Plan
(PNID), (Sept. 2021), https://mintic.gov.co/portal/715/articles-179710_recurso_2.pdf
524
131
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
indicators for monitoring; and to draw up a collaborative participation
strategy for different actors in the data ecosystem.
The third policy document is the Data Infrastructure Governance
Model for the Development of Emerging Technologies that was created by
DAPRE, the National Planning Department, and CAF.526 CONPES 3920
and PNID both emphasized the need to develop an institutional framework
to accompany the development of data infrastructure. The governance
model responds to this need. The governance model outlines five objectives
to guide its design, including institutional coordination, private sector
participation, confidence building, technical modelling, and international
impact. Under each of these objectives, responsible parties or entities and
specific tasks have been provided.
AI and the Judiciary
From 1996, the Colombian government introduced the use of
technology in the administration of justice through Article 95 of Law 270,
while mandating protection of confidentiality, privacy, and security of
personal data.527 As a result, several government entities use AI in judicial
aspects of their work. This includes the Constitutional Court, the Office of
the Attorney General, and the Superintendence of Companies.
In the Constitutional Court of Colombia, where thousands of case
documents are received daily, their processing has been expedited and
improved using an AI system called Promotea. Applying machine learning
abilities, this system investigates, analyses, identifies and suggests priority
cases on health-related aspects within a few minutes.528 Additionally, it
produces statistical reports, automates documentation, systematizes and
synthesizes case law across the country, and improves security by
Presidential Advisory for Economic Affairs and Digital Transformation (DAPRE), the
National Planning Department and the Development Bank of Latin America (CAF), Data
Infrastructure Governance Model for the Development of Emerging Technologies,
(August 2020), https://bit.ly/306wVvn
527
Congress of Colombia, Law 270 of 1996, (Mar. 7, 1996),
https://www.funcionpublica.gov.co/eva/gestornormativo/norma.php?i=6548#:~:text=Exp
ide%20la%20Ley%20Estatutaria%20de,las%20jurisdicciones%20y%20altas%20Cortes.
528
Juan Camilo Rivandeniera, Prometea, artificial intelligence for the revision of
guardianships in the Constitutional Court, (22 March 2019),
https://www.ambitojuridico.com/noticias/informe/constitucional-y-derechoshumanos/prometea-inteligencia-artificial-para-la
526
132
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
integrating blockchain technology. This system that includes human
oversight, improved the efficiency of case processing by 937%.529
The Office of the Attorney General makes use of an AI system
called Fiscal Watson, which consolidates criminal cases across different
databases and regional offices to analyze similar evidentiary elements such
as modus operandi, physical attributes, types of weapon and other
aspects.530 By accelerating and improving the processing of case
information, Fiscal Watson has helped connect and solve similar cases
across the country. The Attorney General has also suggested that Fiscal
Watson can be used to identify irregularities in government contracts made
during the COVID-19 pandemic, since all data and documentation is
available online.531
The Superintendence of Companies, an administrative body
employs a robot assistant called Siarelis (System based on AI for the
Resolution of Company Litigation) to exercise its discretionary judicial
powers in corporate cases related to the piercing of the corporate veil,
insolvency and so on.532 Using Case Based Reasoning (CBR), Siarelis helps
officials identify relevant case law that applies to a legal case and also
provides users with possible decisions that could be reached in their case.533
The outcome reached by the system is decided based on the judicial history
and precedent relevant to a specific case.
Facial Recognition
In Colombia, facial recognition technology is used extensively by
the State. The Ministry of Transport is integrating a network of cameras
with facial recognition technology throughout Bogota. This system that is
meant to prevent and reduce road accidents, will become operational by
Laboratory of Innovation and Artificial Intelligence Buenos Aires (IA Lab), Analyse
2016 sentences in 2 minutes? Prometea in the Constitutional Court of Colombia, (Aug. 6,
2019), https://ialab.com.ar/prometeacolombia/
530
Pablo Medina Uribe and Luisa Fernanda Gomez, ‘Watson,’ the intelligent investigator
with which the Prosecutor’s Office seeks to block crime, (25 July 2020),
https://www.elpais.com.co/judicial/watson-el-investigador-inteligente-con-el-que-lafiscalia-busca-cerrarle-el-paso-al-crimen.html
531
Vanguardia, Fight against corruption in Santander will be done with Artificial
Intelligence, (24 June 2020), https://www.vanguardia.com/politica/lucha-contra-lacorrupcion-en-santander-se-hara-con-inteligencia-artificial-XC2532257
532
Center for Technology and Society Studies (CETyS) of the University of San Andrés,
Readiness of the judicial sector for Artificial Intelligence in Latin America – Analytical
and Exploratory Framework, Republic of Colombia, (2021), https://cetys.lat/wpcontent/uploads/2021/09/colombia-ENG.pdf
533
Superintendence of Companies, Siarelis,
https://www.supersociedades.gov.co/delegatura_mercantiles/Paginas/siarelis.aspx
529
133
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
December 2021 and provides the location of cameras for transparency.534 A
facial recognition system to improve surveillance was also introduced by
local authorities in September 2021, at a stadium in Barranquilla. This
system combines cameras with access to unlimited databases to record and
track individuals.535 The system will be used to identify and detain anyone
with cases pending before the judicial system or any other relevant
authority. The Atanasio Girardot stadium in Medellín also has 170 smart
cameras installed for surveillance, since 2016.536 Expanding the reach of
surveillance in the city, Medellín introduced 40 security robots with facial
recognition capability and an integrated AI security system to patrol the
city.537 The Border Control Agency located at El Dorado International
Airport in Bogota uses the Iris recognition system, with the system expected
to reach other airports in the country within the next few years.538 539
Fundoción Karisma, a civil society organization dedicated to
supporting the responsible use of tech highlights the pitfalls of these
systems. In their report titled Discreet Cameras, they point out that
surveillance technology and biometric identification systems in Colombia
only take into consideration the technical and impact considerations while
assessing systems. There is no analysis using necessity, proportionality or
the possible effect of the technology on human rights. Although the
government tries to ensure transparency by sharing the location of video
surveillance systems that use facial recognition technology, the right to
privacy and other fundamental rights of individuals are still ignored.540
El Tiempo, Do you agree with life-saving cameras recognizing your face?, (Aug. 2,
2021), https://www.eltiempo.com/bogota/camaras-salvavidas-tendrian-reconocimientofacial-en-bogota-movilidad-607508
535
El Tiempo, Colombia vs Chile match will have facial recognition system, (Sept. 9,
2021), https://www.eltiempo.com/colombia/barranquilla/el-metropolitano-cuenta-consistema-de-reconocimiento-facial-616845
536
NEC Corporation, Integrated Surveillance and Security System for Atanasio Girardot
Stadium - Medellín, (2017), https://www.nec.com/en/case/medellin/es/pdf/brochure.pdf
537
El Tiempo, Mayor Daniel Quintero revealed robots to track crime in Medellín, (Aug.
11, 2021), https://www.eltiempo.com/colombia/medellin/daniel-quintero-presento-robotpara-predecir-delitos-en-medellin-609912
538
El Tiempo, How is facial recognition done in Colombia?, (May 17, 2019),
https://www.eltiempo.com/tecnosfera/dispositivos/colombia-que-usos-dereconocimiento-facial-hay-en-el-pais-362220
539
Bloomberg, Gemalto’s Biometric Authentication Technology revolutionizes automated
border control in Colombia, (3 Mat 2018), https://www.bloomberg.com/pressreleases/2018-05-03/gemalto-s-biometric-authentication-technology-revolutionizesautomated-border-control-in-colombia
540
Fundación Karisma, Discreet Cameras, (2 February 2018),
https://web.karisma.org.co/camaras-indiscretas/
534
134
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD/G20 AI Principles
Colombia has endorsed the OECD and G20 AI Principles.
Colombia’s Ethical Framework for AI introduced as a guideline for
trustworthy AI, that provides standards for the ethical use and governance
of AI and is aligned with the OECD AI principles.541 Colombia has not
joined the Global Partnership on AI (GPAI).542 Additionally, Colombia has
developed policy intelligence tools and a follow up plan to monitor the
implementation of the OECD AI Principles, while identifying good
practices to determine if OECD’s recommendations to Colombia have been
implemented.
Human Rights
Colombia is a signatory to many international human rights treaties
and conventions. However, it is only considered to be partly free for the
protection of human rights and transparency, due to the illegal surveillance
operations by the state security forces.543 According to Article 93 of
Colombia’s Political Constitution, rights and duties in the national system
are interpreted according to international treaties and conventions that have
been ratified by its Congress.544 Thus, Colombia has made powerful
commitments backed by strong action that encourages legal certainty, with
an entire implementation, regulatory and monitoring ecosystem for AI. This
is strengthened by expert contribution and public participation at the
national, regional and international level. According to Freedom House,
Colombia scores 65/100 for protection of political rights and civil liberties
and is therefore designated “partly free.” 545
Lethal Autonomous Weapons
During the meeting of the Group of Governmental Experts (GGE)
on lethal autonomous weapons (LAWS), Colombia issued a statement
calling for multilateral regulation to ensure human control over autonomous
Presidential Advisory for Economic Affairs and Digital Transformation (DAPRE),
Ethical Framework for Artificial Intelligence, (Aug. 2020), https://bit.ly/3jTbLIa
542
The Global Partnership on Artificial Intelligence (GPAI), Members,
https://www.gpai.ai/community/
543
Semana, Strikes without quarter: The Persecution of Semana, (Jan. 12, 2020),
https://www.semana.com/nacion/articulo/persecucion-espionaje-y-amenazas-aperiodistas-de-la-revista-semana/647890/
544
Congress of the Republic of Colombia, Political Constitution of Colombia (1991),
https://www.constituteproject.org/constitution/Colombia_2015.pdf?lang=en
545
Human Rights Watch, World Report 2021: Colombia, (2021),
https://www.hrw.org/world-report/2021/country-chapters/colombia
541
135
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
weapons at all times, so no machine makes an autonomous decision.546
Colombia has called for a pre-emptive ban on all LAWS547 and for an
international treaty that will ensure meaningful human control over any use
of force.548
Evaluation
Colombia has anchored its AI policy in the Ethical Framework for
Artificial Intelligence, which has influenced AI policies across Latin
America. Colombia has also endorsed the OECD AI Principles. The
Constitution of Colombia established a right to data protection, and the
country has comprehensive national privacy law and a data protection
agency. However, Colombia’s laws do not include the algorithmic
transparency provisions of the GDPR. There also growing concerns about
the use of AI-enabled facial surveillance systems.
Government of Colombia, Statement at the Convention on Certain Convention
Weapons – Informal meeting of experts on Lethal Autonomous Weapons Systems,
(Apr.2015), https://www.reachingcriticalwill.org/images/documents/Disarmamentfora/ccw/2015/meeting-experts-laws/statements/17April_Colombia.pdf
547
World Summit of Nobel Peace Laureates, Final Declaration of the 16th World Summit
of Nobel Peace Laureates, (4 February 2017), http://www.nobelpeacesummit.com/finaldeclaration-of-the-16th-world-summit-of-nobel-peace-laureates/
548
Government of Colombia, Statement at the Convention on Certain Conventional
Weapons – Group of Government al Experts on Lethal Autonomous Weapons Systems,
(13 April 2018), https://reachingcriticalwill.org/images/documents/Disarmamentfora/ccw/2018/gge/statements/13April_Colombia.pdf
546
136
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Denmark
National AI Strategy
The Danish government unveiled their National AI strategy in March
2019.549 The Danish strategy on AI development outlines the issues that
must be tackled, and defines specific policy efforts and key initiatives. The
National AI strategy intends to establish Denmark as a leader in responsible
AI development. There are four objectives to accomplish this goal:
● Establish a consistent ethical and human-centred foundation for
artificial intelligence;
● Prioritize and promote research in artificial intelligence;
● Encourage the growth of Danish firms through the development
and use of artificial intelligence;
● Ascertain that the public sector utilizes AI to provide world-class
services to citizens and society.
In addition, the strategy is divided into seven major initiatives which
includes the following:
● Digital Hub Denmark;
● SME:Digital;
● The Technology Pact;
● Strengthened Computational Thinking in Elementary Schools;
● Data as a Growth Driver;
● Agile Regulation for New Business Models; and
● Strengthened Cyber Security in Businesses.
The Danish strategy, which featured a budget of 1 billion DKK for
initiatives through 2025,550 is based on proposals from a Digital Growth
Panel551 and the Danish Government's Disruption Committee.552
The Danish Government, National Strategy for Artificial Intelligence (2019)
https://en.digst.dk/media/19337/305755_gb_version_final-a.pdf
550
Ministry of Industry, Business and Financial Affairs, New Strategy to make Denmark
the New Digital Frontrunner (2018), https://eng.em.dk/news/2018/januar/new-strategyto-make-denmark-the-new-digital-frontrunner/
551
European Commission, Directorate-General for Communications Networks, Content
and Technology, Digital Growth Strategy 2025 (2021) https://digital-skillsjobs.europa.eu/en/actions/national-initiatives/national-strategies/denmark-digital-growthstrategy-2025
552
The Danish Government, Ministry for Economic Affairs and the Interior, Denmark’s
National Reform Programme (2019), https://ec.europa.eu/info/sites/default/files/2019european-semester-national-reform-programme-denmark-en.pdf
549
137
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Denmark issued "Towards a Digital Growth Strategy – MADE" in
October 2017,553 identifying AI as a significant growth sector and
identifying a Danish center for artificial intelligence (DCKAI) as one of the
targeted strategies.554 The Digital Strategy titled "A Strong and more Secure
Digital Denmark" was established for the period of 2016-2020,555 which
was issued in May 2016 and includes a quick mention of AI. The plan
established three objectives:
● to create digital solutions simple to use and of high quality,
● to provide favorable conditions for growth, and
● to always promote security and confidence.
The OECD AI Principles
The OECD AI Principles were launched in May of 2019. Since then,
as part of the Danish National Strategy for AI,556 the Danish government
has initiated several aimed at laying a responsible foundation for AI,
including a data ethics council and a data ethics toolbox. 557 By promoting
transparency and establishing data ethics standards, these initiatives hope to
increase the ethical use of data and artificial intelligence in both the public
and private sectors. The Danish government intends to launch a new set of
policies aimed at ensuring a responsible digital economy that is both
trustworthy, ethical, and safe in its use of data and artificial intelligence, as
well as capable of creating innovative solutions.558
Universal Declaration of Human Rights
The 2030 Agenda is based on international human rights principles,
and the Sustainable Development Goals (SDGs) aim to ensure that
Ministry of Foreign Affairs of Denmark, The Danish Government Present the Digital
Growth Strategy (2018) https://investindk.com/insights/the-danish-government-presentsdigital-growth-strategy
554
2021.AI CVR, Partnering with the Danish Center for Artificial Intelligence (2017)
https://2021.ai/partnering-danish-center-artificial-intelligence/
555
Danish Ministry of Finance, Local Government Denmark and Danish Regions, A
Strong and more Secure Digital Denmark, Digital Strategy 2016-2020 (2016),
https://digst.dk/media/16165/ds_singlepage_uk_web.pdf
556
OECD.AI (2021), powered by EC/OECD (2021), database of national AI policies,
accessed on 8/12/2021, https://oecd.ai/en/dashboards/countries/Denmark
557
Frederik Weiergang Larsen, Denmark: an independent council and a labelling scheme
to promote the ethical use of data (2020), https://oecd.ai/en/wonk/an-independentcouncil-and-seal-of-approval-among-denmarks-measures-to-promote-the-ethical-use-ofdata
558
Ministry of foreign Affairs Denmark, Denmark Paves the way for the implementation
of trust by design (2021) https://investindk.com/insights/denmark-paves-the-way-forimplementation-of-trust-by-design
553
138
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
everyone's human rights are met.559 Human rights are a top priority for
Denmark, both nationally and internationally. Human rights is the
foundation upon which all other advancements are built. Human rights are
crucial to Denmark as they are the necessary foundation for a society to
progress.560 The government pledged in January to respond to the CESCR
2019 recommendation that Denmark develop a legal framework requiring
corporate entities to perform due diligence with respect to human rights in
their operations.561 Additionally, the CESCR urged that enterprises be held
accountable for human rights breaches and that victims be given the
opportunity to seek redress.562 By year's end, the government had taken no
efforts toward establishing the necessary legislative framework. Denmark's
largest corporations, for example, are required by law to consider human
rights and report on what they have done to defend them on a yearly basis.
The Danish Parliament established the Danish Centre for Human
Rights in 1987, which was renamed the Danish Institute for Human Rights
in 2002 (Institut for Menneskerettigheder). As Denmark's NHRI (National
Human Rights Institute), the Institute has counterparts in other countries.563
Denmark, as a member of the United Nations, has ratified several human
rights treaties, including those prohibiting torture and strengthening the
rights of persons with disabilities.
Denmark has also ratified several European human rights
instruments, including the European Convention on Human Rights
(ECHR)564 and the establishment of the European Court of Human Rights
(ECtHR or ECHR) in Strasbourg.565 Denmark, as an EU member, supports
the European Parliament's human rights efforts. Denmark has a variety of
mechanisms and authorities in place to safeguard civil rights. According to
Office of the High Commissioner for Human Rights, Human Rights, the SDGs and the
2030 Agenda for Sustainable Development (2019)
https://www.ohchr.org/Documents/HRBodies/UPR/SDGs_2030_Agenda.pdf
560
Amnesty International, Denmark: Human Rights must be ensured for all (2020)
https://www.amnesty.org/en/documents/eur18/3229/2020/en/
561
The Danish Institute for Human Rights, Documenting Business Respect for Human
Rights (2020)
https://menneskeret.dk/sites/menneskeret.dk/files/media/document/~%2020_0034560%20Documenting%20Business%20Respect%20for%20Human%20Rights%202020%
20504132_1_1.PDF
562
Amnesty International, Denmark 2021 (2020)
https://www.amnesty.org/en/location/europe-and-central-asia/denmark/report-denmark/
563
https://www.humanrights.dk/
564
European Court of Human Rights, Council of Europe, European Convention on
Human Rights (n.d) https://www.echr.coe.int/documents/convention_eng.pdf
565
https://www.echr.coe.int/pages/home.aspx?p=home
559
139
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Freedom House, Denmark ranks very highly (97/100) for defense of
political rights and civil liberties.566
Public Participation
Denmark provides programs that enable non-governmental actors
(e.g., the academic community, business, civil society, and regional and
local governments) to express their perspectives or provide expert advice
that informs policy-making processes. These policy initiatives enable
stakeholders or experts to engage in public discussions to share information
and foster collaboration. Public awareness campaigns and civic engagement
activities include informing and consulting with members of the public.567
Privacy and Data Protection
Denmark, like other European countries, has enacted laws to
supplement the EU General Data Protection Regulation (GDPR).568 In
Denmark, the GDPR and its Danish supplementary act, the Data Protection
Act569 are the primary regulations governing the processing of personal
data.
In addition to the GDPR's regulations, the Data Protection Act and
national practice provide for certain exceptions to the GDPR's law
governing the processing of personal data, most notably regarding the
processing of personal data in the employment sector and the processing of
national registration numbers. In 2002, the Danish Act on Personal Data
Processing came into force, implementing Directive 95/46 EC. However,
despite the fact that the Danish data protection regulation is approximately
two decades old, until the GDPR was implemented in 2016, little attention
was paid to data protection in Denmark. Prior to 2017–2018, the term 'data
protection' was almost unknown to the broader Danish population and many
businesses. Thus GDPR compliance has been a hot topic in recent years.
Since the GDPR's adoption, Danish businesses have invested in data
protection compliance, mostly to mitigate economic and legal risks.
Freedom House, Freedom in the World 2021-Denmark (2021),
https://freedomhouse.org/country/denmark/freedom-world/2021
567
OECD, Public awareness campaigns and civic participation activities (n.d)
https://oecd.ai/en/dashboards/policyinstruments/Public_awareness_campaigns_and_other_outreach_activities
568
Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April
2016 on the protection of natural persons with regard to the processing of personal data
and on the free movement of such data, and repealing Directive 95/46/EC (General Data
Protection Regulation).
569
Act No. 502 of 23 May 2018 on supplementary provisions to the regulation on the
protection of natural persons with regard to the processing of personal data and on the
free movement of such data.
566
140
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Danish and other European regulatory agencies have released
several recommendations and decisions interpreting the GDPR and national
additional legislation, allowing Danish businesses to conduct significantly
more targeted and resource-efficient compliance operations. Denmark has
lagged behind the majority of other EU Member States, most notably
Germany and France, when it comes to data protection knowledge and
compliance. This situation is largely explained by the comparatively high
level of trust in Danish society. On the other hand, Denmark appears to have
caught up with the majority of Member States in recent years, owing largely
to increasing public and company knowledge of data protection
requirements, as well as huge countrywide corporate resource investments
since 2016.
Denmark considered developing guidelines to assign specific
initiatives aimed at bolstering government and business efforts in the areas
of information technology security and data protection. Denmark has
maintained a high level of trust in the public sector among both citizens and
business.
According to Statistics Denmark, 83 percent of Danes have
confidence in the way public authorities manage personal information.570
The government has set a goal of increasing this to 90% by 2024 in worldclass digital services. To accomplish this goal and fully exploit the potential
of artificial intelligence, a common ethical framework for the development
and use of artificial intelligence is required. This instills confidence in the
public sector's and private sector's work with data and new technologies. As
a result, the government has established six ethical principles that will serve
as the foundation for future development and application of artificial
intelligence. Additionally, the government will launch several initiatives
promoting a strong emphasis on data ethics. Artificial intelligence
development and use must adhere to applicable legal frameworks. Personal
data should always be used in accordance with the General Data Protection
Regulation's fundamental principles. The legislative framework is found in
the General Data Protection Regulation, administrative law, and other
pieces of legislation that regulate, among other things, work with artificial
intelligence.
AI Policies and Practices in the Public Sector
Denmark aims to support more effective deployment and use of new
technologies, including AI, across the public sector. Among Europeans,
Denmark has the most favorable attitudes about robotics and artificial
Danish Government, National Strategy for Artificial Intelligence 25 (2019)
https://en.digst.dk/media/19337/305755_gb_version_final-a.pdf
570
141
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
intelligence, and they rank second in Europe in terms of learning the skills
necessary for the future workforce, such as social cooperation, creativity,
and digital literacy. Furthermore, Denmark is a leader in various subfields
of artificial intelligence research, with notably strong ecosystems in
algorithmics, machine learning for audio, robotics and computer vision, and
natural language processing.
Oversight Mechanisms
The Danish government established an independent labelling
scheme in collaboration with a consortium of the Confederation of Danish
Industry, the Danish Chamber of Commerce, SMEdenmark, and the Danish
Consumer Council. The seal is an independent labelling scheme awarded to
businesses that meet the organization's requirements for cybersecurity and
safe data handling connected to artificial intelligence.571 The seal will tell
consumers which companies handle data and AI in a trustworthy, ethical
and secure way. As a seal of approval, it will hopefully create a market
incentive for actors to be more data ethical. The consortium plans to launch
the seal in the second half of 2020.
Denmark hopes that these initiatives will normalize the ethical use
of data and create transparency and sustained awareness about data ethics
in business, both in Denmark and globally. In 2019, the Danish government
launched a program whose mission is to advise the public and private
sectors on ethical matters related to data. In 2020, the council will define
data ethics as well as investigate the ethical dimensions of data combination
in the public sector. More generally, the council plans to increase awareness
about ethical dilemmas, in part through public debate.
Universal Guidelines for AI
Denmark is one of the countries that have targeted AI strategies that
include AI-related actions within broader plans of the Universal Guidelines
for Artificial Intelligence. In supporting these guidelines, Denmark aims to
increase AI researchers and skilled graduates; to strengthen national AI
research capacity; and to translate AI research into public- and privatesector applications. In considering the economic, social, ethical, policy and
legal implications of AI advances, the national initiatives reflect differences
in national cultures, legal systems, country size and level of AI adoption,
although policy implementation is at an early stage. This chapter also
Larsen, F. (2020), “Denmark: An independent council and a labelling scheme to
promote the ethical use of data”, The AI Wonk, OECD.AI Policy Observatory,
https://oecd.ai/wonk/an-independent-council-and-seal-of-approval-among-denmarksmeasures-to-promote-the-ethical-use-of-data
571
142
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
examines recent developments in regulations and policies related to AI;
however, it does not analyse or assess the realisation of the aims and goals
of national initiatives, or the success of different approaches.
Independent Council for Ethical use of Data
The Danish government has launched several initiatives to establish
a responsible foundation for AI as part of the Danish National Strategy for
AI. A critical component of these initiatives is to increase the accountability
of both the public and private sectors' use of data and AI by ensuring
transparency and adhering to data ethical guidelines. The Danish
government established an Independent Data Ethics Council in 2019 with
the mission of advising the public and private sectors on data-related ethical
issues. By 2020, the council would have defined data ethics and examined
the ethical dimensions of data fusion in the public sector.
Evaluation
Denmark’s National AI strategy, released in 2019, sets out an
ambitious agenda for the country. Denmark has emphasized responsible AI
development, established an independent Data Ethics Council, endorsed the
OECD AI Principles, and promoted opportunities for public participation in
the development of AI policy. Denmark has also introduced certification
seals to promote trustworthy AI.
143
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Dominican Republic
National AI Strategy
The Dominican Republic does not currently have a formal National
Strategy on Artificial Intelligence (AI), but there have been steps toward
one. The Dominican Republic’s Ministry of Economy, Planning and
Development (MEPyD) detailed goals for "the pillars of the fourth
industrial revolution, the internet of things, Big Data and big data analysis,
cloud computing, artificial intelligence, among other aspects of the
technology.”572 MEPyD Minister Juan Ariel Jiménez recognized that "large
volumes of data, artificial intelligence and blockchain are transforming
economic activities around the world. This technological trend has
repercussions on production, social interaction, planning and, of course, on
public management.” The presentation detailed opportunities for citizens to
participate in public debates to help further the MEPyD’s commitment to
become an "open, close, collaborative and transformative" institution.
The Dominican Republic has focused on using AI for public
management and economic development. In 2021, President Luis Abinader
issued Decree 71-21, creating the Dominican Republic’s Digital Agenda
2030 and establishing the overseeing Digital Transformation Cabinet. The
Cabinet "responds to the government's vision of making information and
communication technologies a strategic tool for sustainable development.
[sic] and inclusive for Dominican society." 573 Lisandro Macarrulla,
Minister of the Presidency, stressed that the Agenda "will raise national
productivity and competitiveness levels, placing us in a better position in
global markets. . . . [and] will improve the quality of life of citizens because
they will be able to receive more and better services from the State and they
will develop new skills that will allow them to access better jobs.” However,
none of the legal provisions specifically concern transparency.
The Vice Minister of Digital Agenda, José David Montilla, assured
that working groups and thematic committees have been created that
Gobierno de la República Economía, Planificación, MEPYD uses artificial intelligence
to improve decision-making in public management (Nov. 28, 2019),
https://mepyd.gob.do/mepyd-utiliza-inteligencia-artificial-mejorar-la-toma-decisionesla-gestion-publica
573
Presidencia de la República Dominicana, President Abinader creates the Digital
Transformation Cabinet, 26 Aug. 2021, https://presidencia.gob.do/noticias/presidenteabinader-crea-el-gabinete-de-transformacion-digital; Luis Abinader, Decree 571-21
(Aug. 26, 2021). https://presidencia.gob.do/sites/default/files/decree/202108/Decreto%20527-21%20Agenda%20Digital%202030.pdf
572
144
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
“include members of the cabinet, representatives of public and private
institutions involved, specialists and volunteers, who will develop each of
the pillars of the Agenda." However, there is no independent oversight, as
the Digital Transformation Cabinet will administratively depend on the
Ministry of the Presidency. 574
There are also initiatives among civil society focused on Artificial
Intelligence policy in the Dominican Republic, notably the Sociedad
Dominicana de IA (SODIA) (Dominican Society of AI).575 Founded in 2014
by a group of professors, students and collaborators, SODIA is a non-profit
dedicated to the study, research and dissemination of AI. This society seeks
to contribute to the creation of an ecosystem of AI in the Dominican
Republic, organizing local events and collaborating with international
associations such as the Association for the Advancement of Artificial
Intelligence, the Canadian Artificial Intelligence Association, the Spanish
Association of AI, the Mexican Society of AI and the Argentine Association
of AI.
Privacy and Data Protection
The Dominican Republic has started to draft a new privacy and data
protection law to become compliant with the Council of Europe’s
Convention 108+, which has been modernized to cover artificially
intelligent systems. Toward this effort, the Dominican Republic requested
and is receiving the support of the Council of Europe.576
Human Rights
According to Freedom House, the Dominican Republic is "partly
free," with a score in 2020 of 67/100 for political rights and civil liberties,
unchanged from 2020.577 According to the recent report, "The Dominican
Republic holds regular elections that are relatively free, though recent years
have been characterized by controversies around implementing a new
electoral framework."
Presidencia de la República Dominicana, President Abinader creates the Digital
Transformation Cabinet, 26 Aug. 2021, https://presidencia.gob.do/noticias/presidenteabinader-crea-el-gabinete-de-transformacion-digital
575
Sociedad Dominicana de Inteligencia Artificial (SODIA), http://www.socdia.org/
576
GLACY+: The Dominican Republic works on new data protection law, Dec. 2019,
https://www.coe.int/en/web/cybercrime/glacyplusactivities//asset_publisher/uKE6ShlCfApw/content/glacy-the-dominican-republic-works-on-newdata-protection-law
577
Freedom House, Freedom in the World 2021 – Dominican Republic (2021),
https://freedomhouse.org/country/dominican-republic/freedom-world/2021
574
145
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Facial Recognition
In 2021, the Dominican Republic’s General Director of
Immigrantion, Enrique García, reported that facial recognition technology
will be used to enforce security in airports and border entry to combat drug
trafficking and international crime.578
Lethal Autonomous Weapons
The Dominican Republic also consented to the 11 Principles on
Lethal Autonomous Weapons Systems (LAWS). These principles respond
to challenges relating to the development of LAWS within the UN’s
Convention of Certain Conventional Weapons (CCW). 579 These priniciples
affirm “international humanitarian law applies to these systems; a human
must always be responsible for the decision to use these systems; [and]
States must examine the legality of these new weapons that they are
developing or requiring at the design stage.
Evaluation
The Dominican Republic has much ground to cover, having yet to
endorse or implement the OECD AI Principles,580 the Universal Guidelines
for AI, or the 2018 GPA Resolution on AI and Ethics, or the 2020 GPA
Resolution on AI and Accountability. New data protection laws guided by
the Council of Europe’s standards on human rights, democracy and the rule
of law are positive steps, as are endorsements to international principles of
transparency. As the country focuses on capacity building and digitization,
how the Dominican Republic uses AI technology should be closely
monitored, especially as the country adopts facial recognition technology in
law enforcement.
Arecoa, DR airports will strengthen security with facial recognition technology (Apr.
20, 2021), https://www.arecoa.com/aeropuertos/2021/04/20/aeropuertos-rd-reforzaranseguridad-tecnologia-reconocimiento-facial/
579
France Diplomacy, 11 Principles on Lethal Autonomous Weapons Systems (LAWS),
https://www.diplomatie.gouv.fr/en/french-foreign-policy/united-nations/multilateralisma-principle-of-action-for-france/alliance-for-multilateralism/article/11-principles-onlethal-autonomous-weapons-systems-laws#sommaire_2.
578
580
https://www.oecd.org/digital/artificial-intelligence/ai-principles/
146
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Egypt
National AI Strategy
In November 2019581 the Egyptian Cabinet approved the formation
of a National Council582 for Artificial Intelligence (NCAI) made up of
representatives from all relevant government entities, as well as
independent experts in the field of Artificial Intelligence (AI).583 The
Technical Committee of the National Council later published Egypt’s
National AI Strategy, building on the previous work of the Ministry of
Communications and Information Technology and the Ministry of Higher
Education and Scientific Research, along with input from independent
experts and private sector companies.584 The deployment of the National
Strategy is considered to be a key step toward Egypt achieving its relevant
UN Sustainable Development Goals (namely goals 4, 5, 8, 9, 10, 11),585 a
top priority for the country’s extensive development road map.586
The primary objectives of the Egyptian National AI Strategy are
to:
1) Exploit AI technologies to support the achievement of Egypt’s
sustainable development goals, to the benefit of all Egyptians.
2) Play a key role in facilitating regional cooperation within the
African and Arab regions and establish Egypt as an active
international player in AI.587”
The strategy consists of four pillars:
1) AI for government: the automation of government processes
and the embedding of AI in decision-making cycles in order to
increase efficiency and transparency.
2) AI for development: the application of AI in different
economic sectors, prioritizing agriculture/environment/water
management; healthcare; Arabic natural language processing;
581
Ministry of Communications and Information Technology, Egypt National Artificial
Intelligence Strategy, July 2021, § 10,
https://mcit.gov.eg/Upcont/Documents/Publications_672021000_Egypt-National-AIStrategy-English.pdf (hereinafter “Egypt National AI Strategy”)
582
National Council for Artificial Intelligence (NCAI), Ministry of Communications and
Information Technology (Egypt), https://mcit.gov.eg/en/Artificial_Intelligence.
583
Egypt National AI Strategy, §2.
584
Id.
585
Id., §5.
586
Id., §4.
587
Id., §2.
147
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
economic planning and development; and manufacturing and
smart infrastructure management.
3) Capacity building: increasing general awareness of AI and
providing professional training.
4) International activities: fostering cooperation at both the
regional and the international level.
In turn, these four pillars are supported by four categories of enablers:
1) Governance: including ethics, laws and regulations, tracking
and monitoring.
2) Data: including collection, management and monetization
strategies.
3) Ecosystem: including private sector, research and academia,
and civil society
4) Infrastructure: including fair access to compute, storage,
networking, and other assets.588
The National Strategy will be implemented in a phased approach.
The first phase, which started in 2020 and which will last until the end of
2022589 is focused on proving the value of AI in different domains and on
building the foundations upon which to build AI at scale.590 In the second
phase, also intended to last 3 years, the emphasis will be placed on
expanding AI into additional sectors. Simultaneously, the government
intends to roll out AI applications at scale, to establish a “paperless,
collaborative, and smart” government.591
Public Participation
The Ministry for Communications and Information Technology has
launched an AI Platform592 which allows the public to easily access the
National Strategy, in addition to AI news, details about AI events, projects,
and capacity-building programs, and information about AI partnerships
with governments, international organizations, private sector companies,
and academia. The AI Platform also includes a page where researchers can
submit academic articles. Neither the National Council for AI nor the
Ministry for Communications and Information Technology, however, has
sought public feedback on any AI policy proposals, nor have they
588
Egypt National AI Strategy, §2.
Id.
590
Id., §11.2.
591
Id.
592
Ministry of Communications and Information Technology, Egypt Artificial
Intelligence Platform, https://ai.gov.eg. (hereinafter “Egypt AI Platform”)
589
148
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
established any meaningful mechanisms by which individuals or groups can
express concerns around the use of AI.
Regional Leadership
Egypt is actively working to bring the perspective of developing
countries to international discussions, thereby helping to narrow the AI
knowledge and development gap between developed and developing
countries, as well as helping to foster the use of AI applications in the latter
group.593 In 2019, Egypt participated in the drafting of the UNESCO AI
recommendations, serving as the Ad Hoc Expert Group’s vice-chair.594
Egypt is also positioning itself as a regional leader in the AI policy world,
having helped create the African’s Union’s African Working Group on AI,
a group tasked with drafting a continent-wide AI strategy.595 In 2021,
UNESCO distributed an AI needs assessment to African countries, the
results of which would inform the African Union’s Working Group future
work.596 UNESCO distilled their survey findings into four key
recommendations: the creation of an AI policy toolkit, the development of
implementation guides and model use cases, the deployment of AI pilot
projects in areas of interest to African countries, and the establishment of
policy guidelines to tackle gender equality issues in AI.597
Egypt also chairs the Arab League’s AI Working Group.598 In
October 2021, the group held their second meeting, in which they discussed
the general outline of a unified Arab strategy for AI.599
593
Egypt National AI Strategy, §9.
Egypt AI Platform, Partnerships, https://ai.gov.eg/Partnerships.
595
François Candelon, HInd El Bedraoui, Hamid Maher, Developing an Artificial
Intelligence Strategy for Africa (Feb. 9, 2021) https://oecd-developmentmatters.org/2021/02/09/developing-an-artificial-intelligence-for-africa-strategy/.
596
UNESCO, UNESCO launches Artificial Intelligence Needs Assessment Survey in
Africa (Mar. 4, 2021), https://en.unesco.org/news/unesco-launches-artificial-intelligenceneeds-assessment-survey-africa.
597
UNESCO, Artificial intelligence needs assessment survey in Africa (2021),
https://unesdoc.unesco.org/ark:/48223/pf0000375410.
598
Ministry of Communications and Information Technology, Egypt Elected Chair of
Arab AI Working Group, (Feb. 16, 2021),
https://mcit.gov.eg/en/Media_Center/Latest_News/News/57187.
599
Ministry of Communications and Information Technology, Egypt Chairs Second
Meeting of Arab AI Working Group (Oct. 18, 2021)
https://mcit.gov.eg/en/Media_Center/Latest_News/News/63741.
594
149
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Responsible AI
The following points list major Key Performance Indicators (KPIs) for
ensuring Ethical AI in Egypt:
● The establishment of a dedicated track within NCAI for AI Ethics.
● Publish Guidelines for Responsible and Ethical development of AI.
● A set of rules and regulations for responsible AI use.
● Ethics in AI/technology courses being offered in universities as
part of computing degrees.600
Most significantly, Egypt is currently developing the Egyptian
Charter on Responsible AI,601 which the government purports will
demonstrate how the country is implementing and/or adjusting the OECD
AI Principles to suit its unique goals. The document will include guidelines
and best practices for assessing AI systems.
OECD/G20 AI Principles
Egypt has endorsed the OECD AI Principles and has taken a few
meaningful actions to begin implementing them. As reported by the OECD
in their 2021 white paper “State of implementation of the OECD
principles,” Egypt has set up a governing body (the National Council for
AI) to oversee the implementation of their AI strategy.602 This is a concrete
first step toward fulfilling the OECD recommendation of ensuring “a policy
environment that will open the way to deployment of trustworthy AI
systems.” However, this recommendation will only be satisfied if Egypt
adopts AI policies that protect human rights and that ensure the responsible,
transparent, and fair use of the technology.
Egypt has also taken steps toward fulfilling three of the four other
OECD recommendations. The creation of both the AI Platform and the new
Egyptian Center of Excellence603 a government group that will work with
private or academic partners to deliver AI projects on behalf of
beneficiaries, help to “foster accessible AI ecosystems with digital
infrastructure and technologies and mechanisms to share data and
knowledge.” Empowering “people with the skills for AI and support
workers for a fair transition” will be accomplished through enrollments in
600
Egypt National AI Strategy, §10.1.2
Egypt AI Platform, AI Strategy Info, https://ai.gov.eg/strategy/strategy-info.
602
Organisation for Economic Co-operation and Development (OECD), "State of
implementation of the OECD AI Principles: Insights from national AI policies", OECD
Digital Economy Papers, No. 311, 18 June 2021, https://doi.org/10.1787/1cd40c44-en,
pg. 10.
603
Egypt AI Platform, Egyptian AI Center of Excellence (AIEG),
https://ai.gov.eg/strategy/center-of-excellence.
601
150
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the newly created “Faculties of AI” at eight public and private Egyptian
universities.604 Egypt’s cooperation “across borders and sectors to progress
on responsible stewardship of trustworthy AI” is evidenced by its
participation in and leadership of international AI committees.
The final OECD recommendation is that governments “facilitate
public and private investment in research & development to spur innovation
in trustworthy AI.” The Egyptian government’s future plans set a public
goal of 7.7% of Egyptian Gross Domestic Product deriving from AI &
robotics by 2030.605
Human Rights
While Egypt has endorsed the Universal Declaration of Human
Rights, its human rights record has been highly criticized, with Freedom
House giving the country a freedom score of 18/100 for 2021.606 In March
2021, 31 UN member states penned a joint declaration, supported by
numerous NGOS, strongly condemning human rights abuses in Egypt. The
declaration highlighted constraints on citizens’ freedom of expression, as
well as their ability to voice political opposition and to peacefully
assemble.607
Biometric Recognition
Egypt is increasingly adopting biometric technologies for security
and surveillance. A recent deal was struck between the Arab Organization
for Industrialization and Idemia, a leading biometric company, for the latter
to produce biometric devices, including facial recognition systems, in
Egypt.608 The Egyptian government contracted with Idemia in early 2020 to
build a biometric ID system.609 In 2021, Fingo, another organization
604
Sally Radwan, Samar Sobeih, Egypt’s AI strategy is more about development than AI,
OECD.ai Policy Observatory, 26 May 2021, https://oecd.ai/en/wonk/egypt-ai-strategy.
605
Rebellion Research, Egypt AI: Egypt’s Artificial Intelligence Future (Mar. 14, 2021),
https://www.rebellionresearch.com/egypts-artificial-intelligence-future.
606
Freedom House, Freedom in the World 2021,
https://freedomhouse.org/country/egypt/freedom-world/2021.
607
Human Rights Watch, Condemnation of Egypt’s Abuses at UN Rights Body: Overdue
Action is a Step Forward (Mar. 12, 2021),
https://www.hrw.org/news/2021/03/12/condemnation-egypts-abuses-un-rights-body#.
608
Ayang MacDonald, Idemia renews Mauritania contract, signs deal with AOI for
biometric device production in Egypt, Biometric Update.COM (Nov. 9, 2020)
https://www.biometricupdate.com/202011/idemia-renews-mauritania-contract-signs-dealwith-aoi-for-biometric-device-production-in-egypt
609
Chris Burt, Idemia to build biometrics-backed digital identity service in Egypt, supply
TSA trials, joins Kantara, Biometric Update.COM (Mar. 12, 2020),
151
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
specializing in biometrics, announced it had partnered with Egypt to
develop a vein-based recognition system for the country’s national ID
program.610
Evaluation
Egypt has made progress toward implementing the OECD AI
Principles. These efforts have had the dual purpose of increasing AI
capacity within its borders, for example by fostering private sector AI
projects or by establishing AI Faculties at Egyptian universities. Similarly,
many of Egypt’s AI endeavours have been dedicated to bolstering its
leadership credentials, such as by heading international working groups.
Less has been done to enshrine AI principles and practices based in human
rights across Egypt’s public and private sectors. More information on how
Egypt plans to implement the OECD AI Principles will be available once
the country releases its Charter on Responsible AI, though this may not
alleviate all concerns, as Egypt has stated the document will expound on not
only the implementation of OECD Principles, but also their modification so
as to better conform to Egypt’s objectives. This, in combination with a lack
of public participation and a poor track record of human rights, allows us to
accord Egypt only middling scores.
https://www.biometricupdate.com/202003/idemia-to-build-biometrics-backed-digitalidentity-service-in-egypt-supply-tsa-trials-joins-kantara.
610
Fingo, “Egypt to unlock futuristic ID verification with finger-vein recognition tech”,
18 Feb. 2021, https://www.fingo.to/media/egypt-to-unlock-futuristic-id-verification-withfinger-vein-recognition-tech/
152
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Estonia
A global leader in the use of digital technologies for egovernment,611 the Estonian public sector has adopted at least 41 AI projects
and has a goal of having at least 50 AI use cases by the end of 2020.612 In
2018, the Estonian Undersecretary for Communications and State
Information Systems emphasized the importance of facilitating AI in
Estonia for investment and innovation, as well as for public
administration.613 In light of its commitment to e-government, Estonia
emphasizes the use of AI for government services.614 Indeed, KrattAI refers
to “the vision of how digital public services should work in the age of
artificial intelligence;” or more specifically, KrattAI is described as an
“interoperable network of AI applications, which enable citizens to use
public services with virtual assistants through voice-based interaction.”615
The Estonian government makes use of automated decision-making
in many different contexts.616 For example, the Tax and Customs Board uses
automated decision-making to facilitate tax refunds following the
submission of an online income tax return. Other examples include the use
of tachographs on lorries and automated speed checks on motorways to
issue cautionary fines and the use of automated decision-making for the
determination of a child’s school on the basis of their registered residence.617
There has been international coverage of Estonia’s ambitious plans for AI
e-estonia, https://e-estonia.com.
Republic of Estonia GCIO Office, Artificial Intelligence for Estonia,
https://www.kratid.ee/in-english’
613
Riigikantselei, Estonia will have an Artificial Intelligence Strategy (Mar. 27, 2018)
https://www.riigikantselei.ee/et/uudised/eesti-saab-tehisintellekti-strateegia; Tanel
Kerikmäe and Evelin Pärn-Lee, Legal Dilemmas of Estonian Artificial Intelligence
Strategy: In Between of E-Society and Global Race, AI & Society (2020).
614
It has been noted that the “Estonian public sector is highly digitalized, whereas the
private sector is not.” Tanel Kerikmäe and Evelin Pärn-Lee, Legal Dilemmas of Estonian
Artificial Intelligence Strategy: In Between of E-Society and Global Race, AI & Society
(2020).
615
KRATT Artificial Intelligence Programme of #Estonia, #KrattAI: roadmap for 2020
https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_19625e00a7b84900b99e952b1ce7d21a.pdf;
Republic of Estonia, Ministry of Economic Affairs and and Communications, Report of
Estonia’s AI Taskforce (May 2019), https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf. The
report of Estonia's AI Taskforce defined ‘kratt’ as being “a practical application that uses
artificial intelligence and that fulfils a specific function.”
616
See also #KrattAI Roadmap for 2020 https://www.kratid.ee/roadmap.
617
Council of State of the Netherlands and ACA-Europe, An Exploration of Technology
and the Law (May 14, 2018), http://www.aca-europe.eu/colloquia/2018/Estonia.pdf.
611
612
153
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
in the public sector – including on the issue of “Robot Judges.”618 The
Estonian court system embraces digitalization and started an e-File system
in 2005. The use of AI to tackle an immense backlog of cases has been
considered, including the adoption of projects that can make “autonomous
decisions within more common court procedures/tasks that would otherwise
occupy judges and lawyers alike for hours.”619
National AI Strategy
The Estonian Cabinet adopted its National AI Strategy in July
620
2019. The Government Chief Information Officer Office, based in the
Ministry of Economic Affairs and Communications, is tasked with steering
the AI Strategy. The National AI Strategy builds on a May 2019 report of
Estonia’s AI Taskforce.621 The actions detailed in the AI Strategy are
designed to advance the adoption of AI solutions in both the private and
public sectors, to increase AI capacities and research and development, and
to develop the legal environment to facilitate AI. The AI Strategy commits
to the establishment of a steering group, comprised of government
representatives and other stakeholders, in order to monitor the
implementation of the AI Strategy. In addition, the e-Estonia Council will
consider the strategy’s implementation annually. The AI Strategy is a shortterm strategy, intended to apply up until 2021. By adopting a short-term
strategy, Estonia intends to gain insight and develop a long-term strategy in
response to the experience. Estonia will monitor the development of the
short-term action plan and keep the European Union informed of
developments.
In spite of Estonia’s national digital adviser initially proposing the
adoption of a law granting legal personality to AI, Estonia’s AI Taskforce
concluded that no substantial legal changes are currently required to address
the issues presented by AI.622 The Taskforce Report maintained that: “Both
618
Eric Niller, Can AI Be a Fair Judge in Court? Estonia Thinks So, Wired (Mar. 23,
2019) https://www.wired.com/story/can-ai-be-fair-judge-court-estonia-thinks-so/.
619
Anett Numa, Artificial Intelligence as the New Reality of E-justice, e-estonia (Apr.
2020) https://e-estonia.com/artificial-intelligence-as-the-new-reality-of-e-justice/https://eestonia.com/artificial-intelligence-as-the-new-reality-of-e-justice/.
620
Estonia’s National AI Strategy 2019-2021 (July 2019) https://f98cc689-5814-47ec86b3-db505a7c3978.filesusr.com/ugd/7df26f_27a618cb80a648c38be427194affa2f3.pdf.
621
Report of Estonia’s AI Taskforce (May 2019) https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf.
622
Report of Estonia’s AI Taskforce (May 2019) https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf.; See
Astghik Grigoryan, Estonia: Government Issues Artificial Intelligence Report (July 31,
154
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
now and in the foreseeable future, kratts are and will be human tools,
meaning that they perform tasks determined by humans and express the
intention of humans directly or indirectly.” Accordingly, the AI Taskforce
Report clarifies that the “actions” of AI are attributable to the relevant state
body or private party that uses the AI solution.623 Minor changes
recommended include the removal of obsolete laws and providing
additional clarity in order to facilitate the use of AI. Estonia’s Chief
Information Officer stated that Estonia wants to “build on the EU
framework, not to start creating and arguing” for a separate Estonian
framework.624
Neither the AI Strategy nor the AI Taskforce Report provide
significant detail on questions related to the ethics of artificial intelligence.
Reference is, however, made to guidance provided by the European
Commission for the development and implementation of trustworthy
artificial intelligence.625 The Taskforce Report acknowledges that
“trustworthy artificial intelligence must be guided by the principles of
human rights, positive rights, and values, thus ensuring the ethics dimension
and objective.”626 The Report recognizes the relevance of the EU Charter of
Fundamental Rights and refers to the following rights as central according
to the Commission guidance on AI:
•
•
•
The right to human dignity.
The right to freedom.
Respect of the principles of democracy and the state, based on
the rule of law.
2019), https://www.loc.gov/law/foreign-news/article/estonia-government-issuesartificial-intelligence-report/.
623
Estonia’s National AI Strategy (July 2019) https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_27a618cb80a648c38be427194affa2f3.pdf.
624
Astghik Grigoryan, Estonia: Government Issues Artificial Intelligence Report, US
Library of Congress (July 31, 2019) https://www.loc.gov/law/foreignnews/article/estonia-government-issues-artificial-intelligence-report/; referencing Ronald
Liive, Estonian State IT Manager Siim Sikkut: If There Were 1% in the State Budget for
Science, We Could Talk More About Kratind, DigiGeenius (May 5, 2019). In 2018,
Estonia signed up to a European Union Declaration of Cooperation on Artificial
Intelligence https://ec.europa.eu/digital-single-market/en/news/eu-member-states-signcooperate-artificial-intelligence.
625
Report of Estonia’s AI Taskforce (May 2019) https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf;
European Commission, Ethics Guidelines for Trustworthy AI (Apr. 8, 2019)
https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai.
626
Report of Estonia’s AI Taskforce (May 2019), https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf.
155
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
Right to equality, non-discrimination, and acknowledgement of
minorities.
Civil rights.
To ensure that the development and use of AI is ethical, the
Taskforce Report emphasizes the importance of ensuring that AI is humancentric; that rights, ethics principles, and values are fundamental; and that
AI may bring unintended consequences. The AI Strategy references the EU
guidelines that identify the importance of the following values: human
agency, technical reliability, privacy and data management, transparency,
non-discrimination, social and environmental well-being, and
responsibility.
OECD AI Principles
In May 2019, Estonia signed the OECD Principles on Artificial
Intelligence, “agreeing to uphold international standards that aim to ensure
AI systems are designed to be robust, safe, fair and trustworthy.”627
Human Rights
Estonia is a member of the European Union and the Council of
Europe and is, accordingly, committed to the upholding of the Charter of
Fundamental Rights and the European Convention on Human Rights.
Estonia is committed to the Universal Declaration on Human Rights and
has acceded to international human rights treaties, such as the International
Covenant on Civil and Political Rights. The Estonian Constitution grants
basic rights to citizens.
In Freedom House’s 2020 and 2021 Country Reports, Estonia
ranked highly (94/100). It was reported that, ‘Democratic institutions are
strong, and political and civil rights are widely respected in Estonia.’628 On
the issue of openness and transparency, Freedom House reported that
“Estonia is well-known for its transparency and well-developed egovernance services. Recently, however, several security flaws in these
systems were revealed. While the government announced a plan to remedy
the situation, additional resources to support the maintenance and further
expansion of the e-governance program are needed.”
627
OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence
(May 22, 2019), https://www.oecd.org/science/forty-two-countries-adopt-new-oecdprinciples-on-artificial-intelligence.htm.
628
Freedom House, Freedom in the World 2021 – Estonia (2021),
https://freedomhouse.org/country/estonia/freedom-world/2021
156
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In a 2018 report of the Commissioner for Human Rights of the
Council of Europe, the Commissioner urged the Estonian authorities to give
careful consideration “to the ethical, legal and human rights implications of
using robots and artificial intelligence in the care of older persons” given
Estonia’s strong focus on digitalization, new technologies, and AI.629
Algorithmic Transparency
Estonia is also a member of the Council of Europe and was among
the first states to ratify the modernized Privacy Convention.630 Article
9(1)(c) of the Convention provides a right for algorithm transparency. As a
member of the European Union, Estonia is also committed to the protection
of personal data as required by Article 8 of the Charter of Fundamental
Rights and the data protection laws of the EU. The Personal Data Protection
Act was enacted in 2018 in order to adapt the GDPR and to implement the
Law Enforcement Directive into Estonian law.631 Accordingly, the
processing of personal data in Estonia must comply with the data protection
principles, including the principles of purpose limitation, minimization, and
fair and lawful processing. Moreover, automated processing can only be
carried out in specific circumstances and data subjects are granted specific
rights in that context. Article 17 of the Personal Data Protection Act places
limits on automated processing.632 The Estonian Supervisory Authority is
the Data Protection Inspectorate.633
The Estonian government provides a data tracker tool accessible
through the state portal (eesti.ee) that enables anyone with an eID to keep
track of which institutions have accessed their data and for what purposes.634
As pointed out on the e-estonia website, transparency is “fundamental to
foster trust in the effective functioning of the whole system.” Notably,
information is also provided regarding automated processing although
Algorithm Watch states that it “is not always clear if data is used as a part
Council of Europe, Commissioner for Human Rights, Report of the Commissioner for
Human Rights of the Council of Europe Dunja Mijatović Following her Visit to Estonia
from 11 to 15 June 2018, 21 https://rm.coe.int/report-of-the-council-of-europecommissioner-for-human-rights-dunja-mi/16808d77f4.
630
Council of Europe, Estonia, 7th State to ratify Convention 108+ (Sept. 16, 2020),
https://www.coe.int/en/web/human-rights-rule-of-law/-/estonia-7th-state-to-ratifyconvention-108631
Personal Data Protection Act
https://www.riigiteataja.ee/en/eli/523012019001/consolide.
632
Algo:Aware, State-of-the-Art Report: Algorithmic Decision-Making (Dec. 2018)
https://actuary.eu/wp-content/uploads/2019/02/AlgoAware-State-of-the-Art-Report.pdf.
633
Data Protection Inspectorate https://www.aki.ee/en.
634
Federico Plantera, ‘Data Tracker - Tool that Builds Trust in Institutions’ (e-estonia,
September 2019) https://e-estonia.com/data-tracker-build-citizen-trust/.
629
157
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of an automatic process or viewed by an official.”635 In spite of the ambition
of this tool, the Estonian Human Rights Center argue that the data provided
is variable depending on the service and at times not detailed enough. To
assist transparency and understanding, the Estonian Human Rights Center
suggests that visual depictions of data use should be provided.636 Similarly,
Algorithm Watch state that the current tool does not provide a “clear
understanding of what profiling is done by the state, which data is collected,
how it is used, and for what purpose.”637
The Ministry of Justice intends to draft legislation addressing highrisk algorithmic systems that will require the creators of AI (both public and
private) to provide transparency regarding when AI communicates with an
individual, processes an individual’s data, or makes a decision on the basis
of the individual’s data.638 A representative of the Ministry said that nontransparency of decisions is the biggest threat. When it comes to AI, based
on current knowledge, even the person who wrote the algorithm's code is
unable to explain the reasons behind a decision, as the system is selflearning and self-evolving. “An assessment or a decision made by an
algorithm may have a significant impact on fundamental rights no matter
whether we are speaking of a self-learning or a human-defined algorithm.
It is a duty of a country of rule of law to be foresightful and prevent serious
interferences with fundamental rights by means of setting out a relevant
legislative framework,’ said Kai Härmand with the Ministry of Justice.
Public Participation
In 2018, the Estonian government brought together an expert group
to participate in a cross-sectional coordination project on AI.639 The three
tasks of this expert group were to
Algorithm Watch, Automating Society Report 2020 75 (Oct. 2020),
https://automatingsociety.algorithmwatch.org/wp-content/uploads/2020/10/AutomatingSociety-Report-2020.pdf.
636
Kari Käsper and Liina Rajavee, Inimõigused, Infoühiskond Ja Eesti: Esialgne
Kaardistus (Estonian Human Rights Centre 2019)
https://humanrights.ee/app/uploads/2019/12/ EIK-kaardistamine.pdf (Estonian).
637
Algorithm Watch, Automating Society Report 2020 75 (Oct. 2020),
https://automatingsociety.algorithmwatch.org/wp-content/uploads/2020/10/AutomatingSociety-Report-2020.pdf.
638
Estonian Ministry, Use of AI must Respect Fundamental Rights (Aug. 19, 2020)
www.baltic-course.com/eng/Technology/?doc=158411&output=d.
639
Report of Estonia's AI Taskforce 42 (May 2019) https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf.
635
158
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
•
prepare draft legislation to ensure clarity in the Estonian
judicial area and organize the necessary supervision;
develop the so-called Estonian artificial intelligence action
plan;
notify the public about the implementation of kratts and
introduce possible options.
Participants in the group included representatives from state
authorities, the private sector, universities, and sectoral experts. In order to
prepare the report, interviews were conducted, including with company
representatives involved in the development of AI and ICT representatives
from universities. Working groups (in the fields of law, education, and the
public sector) were also assembled for discussion.640 There is a commitment
to the importance of diverse inputs in the AI debate. The e-estonia website
states:
In these debates, technical and legal expertise goes a long
way. But the discussion must also involve the public.
Honest, meaningful debate requires that dreamy utopias be
balanced with open discussions about AI’s controversial
attributes and threats. Only this can create user-friendly
legislation that’s equipped to reduce legal nightmares in the
long-term.641
Documents relating to the AI Strategy are accessible on the internet.
The website Kratid provides links to the National Artificial Intelligence
Strategy, the Report of Estonia's AI Taskforce, the ‘Vision Paper on
#KrattAI: The Next Stage of Digital Public Services in #eEstonia’, and the
‘#KrattAI Roadmap for 2020’.642
Evaluation
Estonia has set out a short-term AI Strategy formed from the AI
Taskforce Report. As a member of the European Union and the Council of
Europe, Estonia is committed to the protection of human rights, ethics in
AI, and algorithmic transparency. Estonia has also endorsed the OECD AI
Principles and signed the Declaration of Collaboration on AI in the NordicBaltic Region which includes a commitment “to develop ethical and
Report of Estonia's AI Taskforce May (2019) (See Annex for details on membership_,
42 https://f98cc689-5814-47ec-86b3db505a7c3978.filesusr.com/ugd/7df26f_486454c9f32340b28206e140350159cf.pdf.
641
e-estonia, AI and the Kratt Momentum (Oct. 2018) https://e-estonia.com/ai-and-thekratt-momentum/.
642
Kratid, Search for Estonia, https://www.kratid.ee/in-english.
640
159
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
transparent guidelines, standards, norms and principles that can be
employed as a steering mechanism to guide AI programmes.”643 In spite of
these commitments, neither the AI Strategy nor the AI Taskforce Report
consider the issues of ethics and human rights in significant depth. Due to
the short-term nature of the current AI Strategy, there is an opportunity –
and apparent intention – for Estonia to adopt a clear ethical framework in
practice.
Government of Sweden, Nordic Council of Ministers, AI in the Nordic-Baltic region
(May 14, 2018),
https://www.regeringen.se/49a602/globalassets/regeringen/dokument/naringsdepartement
et/20180514_nmr_deklaration-slutlig-webb.pdf.
643
160
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Finland
National Approach to Artificial Intelligence
In 2017, Finland was among the first countries to develop a national
AI strategy, establishing proposed target dates and allocating public funds
in furtherance of the country’s AI-related business objectives. At that time,
Finland explained that it implemented this AI strategy to: (1) enhance
business competitiveness using AI, (2) ensure top-level expertise and attract
top experts, (3) provide the world’s best public services, and (4) make
Finland a front runner in the age of AI.644 To that end, in May 2017,
Finland’s Minister of Economic Affairs launched “Finland's AI
Programme,”645 an operational program charged with “turning Finland into
a leading country in the application of artificial intelligence.” The
Programme focused on three areas -- an efficient public sector, a wellfunctioning society, and a competitive business and industry sector.646
Under the operational umbrella of the AI Programme, the Minister of
Economic Affairs identified five tasks necessary to achieve this objective:
1) To generate a snapshot of the current status and prospects for
AI and robotics around the world and in Finland.
2) To propose a goal state, which Finland should strive to achieve
in the application of AI in collaboration with companies,
research institutes, educational institutions and public
organizations.
3) To enter a proposal on measures the implementation of which
is necessary in order to achieve the stated objectives. Special
attention must be given to the field’s innovation activities,
preparedness for changes to working life, increasing education
and upgrading the qualifications of those in the labour market.
4) To draw up a model for the implementation of the plan that
will ensure the efficient realisation of the operational
programme.
644
OECD, State of Implementation of the OECD AI Principles 18 (June 2021),
https://read.oecd.org/10.1787/1cd40c44-en?format=pdf.
645
AI Finland Background (2018), https://www.tekoalyaika.fi/en/background/
646
Ministry of Economic Affairs and Employment, Leading the way into the age of
artificial
Intelligence: Final Report of Finland’s Artificial Intelligence Programme 44 (2019)
(“Final Report”),
https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/161688/41_19_Leading%20the
%20way%20into%20the%20age%20of%20artificial%20intelligence.pdf.
161
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
5) To prepare a proposal for the expansion of the working group’s
task description and composition, so as to allow it to develop
the measures necessary for the promotion of AI in the longterm and analyse the more broad-scoped societal change
related to digitalisation and provide proposals for solutions to
the Government.
The Minister of Economic Affairs also appointed a steering
committee, which included representatives from the public, private, and
research sectors, and charged the steering committee with publishing a
report containing the committee’s recommendations regarding the
operationalization of Finland’s AI Programme. The Minister of Economic
Affairs also established four subgroups “assigned to participate in the
compilation and implementation of the Finnish AI Programme” and focused
on four (4) key areas: (1) expertise and innovations; (2) data and platform
economy; (3) transformation of society and work; and (4) ethics.
In December 2017, the steering committee published the first of
three reports entitled, “Finland’s age of artificial intelligence647: Turning
Finland into a leading country in the application of artificial intelligence.”
(Finland, 2017). The report examined “the significance of artificial
intelligence to Finland’s well-being, revised the programme objectives, and
made eight (8) recommendations for actions, which, if implemented, would
facilitate Finland’s objective to “adopt and benefit from AI” 648:
1) Enhancement of business competitiveness through the use of
AI
2) Effective utilization of data in all sectors
3) Ensure AI can be adopted more quickly and easily
4) Ensure top-level expertise and attract top experts
5) Make bold decisions and investments
6) Build the world’s best public services
7) Establish new models for collaboration
8) Make Finland a front runner in the age of AI
In June 2018, the transformation of work and society subgroup,
working under the direction of the AI Programme steering group, published
a second report entitled, “Work in the age of artificial intelligence: Four
647
Ministry of Economic Affairs and Employment, Finland’s Age of Artificial
Intelligence: Turning Finland into a leading country in the application of artificial
intelligence, Objective and recommendations for measures (2017),
http://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/160391/TEMrap_47_2017_verk
kojulkaisu.pdf
648
https://futureoflife.org/ai-policy-finland/
162
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
perspectives on the economy, employment, skills and ethics.”649 The June
2018 report separately examined (1) impacts of artificial intelligence on
growth and employment; (2) labour market dynamics in a technological
revolution; (3) learning and skills in a transition; and (4) good application
of artificial intelligence technology and ethics. The report also offered three
(3) policy recommendations to add to the prior 8 recommended actions
previously delineated in the December 2017 report:
1) Increase the competitiveness of business and industry;
2) Provide high-quality public services and improve the efficiency
of the public sector;
3) Ensure a well-functioning society and wellbeing for its
citizens.650
In December 2019, the steering committee published Leading the
way into the age of artificial intelligence.651 The final report detailed the
steering committee’s policy recommendations for Finland’s AI
Programme. The report also set forth the contours of Finland’s “vision” of
a country that by 2025, “is competitive and able to attract talent and has
the most relevantly educated population consisting of well-informed and
independent citizens” in “the age of artificial intelligence.”
As the steering committee explained in the 2019 final report:
In another five years time, artificial intelligence will be an active
part of every Finn’s daily life. Finland will make use of artificial
intelligence boldly in all areas of society – from health care to the
manufacturing industry – ethically and openly. Finland will be a
safe and democratic society that produces the world’s best services
in the age of artificial intelligence. Finland will be a good place for
citizens to live and a rewarding place for companies to develop and
grow. Artificial intelligence will reform work as well as create
wellbeing through growth and productivity.
Finland’s Artificial Intelligence Programme has grown to include several
AI initiatives including Finland’s AI Business Programme, a 2018
programme that provided public funding for 115 AI business projects, and
649
Ministry of Economic Affairs and Employment, Work in the age of artificial
intelligence: Four perspectives on the economy, employment, skills and ethics (2018)
https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/160980/TEMjul_21_2018_Wor
k_in_the_age.pdf
650
The report cautioned that “[t]he conclusions of the report d[id] not necessarily
represent the group’s joint views,” but “d[id] represent a majority opinion.”
651
Finland, Ministry of Economic Affairs and Employment (2019),
http://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/161688/41_19_Leading%20the
%20way%20into%20the%20age%20of%20artificial%20intelligence.pdf
163
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Finland’s Artificial Intelligence Accelerator, a 6-month AI-deployment
accelerator program focused on assisting organizations with the
operationalization of AI solutions. In addition to these initiatives, in
November 2020, Finland launched an updated AI strategy entitled the
Artificial Intelligence 4.0 Programme,652 which promotes the development
and introduction of AI and other digital technologies in companies, with a
special focus on subject matter experts (“SMEs”) from several industrial
and service sectors.653
Access to Data
The 2019 final report noted that “[d]ata has become the world's most
valuable resource, but when the existing operating models are applied, it
primarily benefits a few giant corporations that collect the data from their
service users.” With respect to the adoption of the GDPR by Finland (and
other EU countries), the final report opined that although the law
“strengthened the rights of individuals and harmonised the EU regulation
related to processing of personal data” as currently formulated, “there is no
joint concept or interoperable open ecosystem for the exchange of personal
data based on consumer consent.”
The final report observes that “Finland is in a position to become a
global trendsetter and a forerunner within the EU in the creation of fair,
consumer-oriented principles” and that consumer-oriented principles
require “a visionary approach and a joint EU-level roadmap, as well as
technical proof of functional exchange of data.” According to the report,
“‘Small data’ . . . may offer new opportunities for applying AI within the
B2B field in particular . . . and “an opportunity for Finland.”
Foreign Policy and AI
Finland is a signatory of declaration, “AI in the Nordic-Baltic
region” establishing a collaborative framework on “developing ethical and
transparent guidelines, standards, principles and values to guide when and
how AI applications should be used” and “on the objective that
infrastructure, hardware, software and data, all of which are central to the
652
Ministry of Economic Affairs and Employment, Artificial Intelligence 4.0 programme
to speed up digitalisation of business (Nov. 17, 2020), https://tem.fi/en/-/artificialintelligence-4.0-programme-to-speed-up-digitalisation-of-business
653
European Commission, AI Watch, Finland AI Strategy Report (2020),
https://knowledge4policy.ec.europa.eu/ai-watch/finland-ai-strategy-report_en
164
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
use of AI, are based on standards, enabling interoperability, privacy,
security, trust, good usability, and portability.”654
Besides Ala-Pietilä, Pekka, who has been appointed as a chair for
the Finland’s AI Programme has chaired to the European Commission's
High Level Expert Group on AI which determined the Ethics Guidelines for
Trustworthy AI, the document that puts forward a human-centric approach
on AI and the key requirements for a trustworthy AI.
Within the European Union and in international settings, Finland cooperates
with other countries with a view to strengthening a market-based data
economy, access to data and cross-border data mobility, with due
consideration for privacy and national security. Finland participates in
international processes aiming to create ethical frameworks and shared core
values for making use of artificial intelligence.655
UNICEF and Finland have been collaborating to create
internationally applicable policy guidance for the use and development of
artificial intelligence (AI) for children. The Ministry for Foreign Affairs
supports the project, where practices are developed for the planning of safe
and inclusive AI solutions that take the rights of the child into account. The
second draft of the policy guidance, developed based on the results of the
piloting, will be published in November 2021.656
Also Finland designed a free online course, The Elements of AI,
with the University of Helsinki and Reaktor, a Finnish technology company.
The objective of the course is to encourage everyone, regardless of age or
educational background, to learn the basics of artificial intelligence.657
654
Nordic Cooperation, Declaration AI in the Nordic-Baltic region (May 14, 2018),
https://www.norden.org/en/declaration/ai-nordic-baltic-region.
655
Finland, Government report on information policy and artificial intelligence (Dec. 5,
2018),
https://vm.fi/documents/10623/7768305/VM_Tiepo_selonteko_070219_ENG_WEB.pdf/
89b99a8e-01a3-91e3-6ada38056451ad3f/VM_Tiepo_selonteko_070219_ENG_WEB.pdf.pdf/VM_Tiepo_selonteko
_070219_ENG_WEB.pdf
656
Ministry of Foreign Affairs, Policy Guidance on AI for Children piloted in different
parts of the world (Oct. 19, 2021), https://um.fi/current-affairs//asset_publisher/gc654PySnjTX/content/lapsiin-liittyvan-tekoalyn-pelisaantoja-pilotoitueri-puolilla-maailmaa
657
Ministry of Foreign Affairs, Finland to enhance Europeans’ digital skills – Elements
of AI online course to be launched in EU countries (May 5, 2020), https://um.fi/currentaffairs/-/asset_publisher/gc654PySnjTX/content/suomi-vahvistaa-eurooppalaistendigitaitoja-ja-osaamista-elements-of-ai-verkkokurssin-lanseeraukset-eu-maissa-alkavat
165
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation
Finland’s longstanding and broad commitment to an open
democracy has traditionally been given expression by extensive
consultation with established groups.658 Also the Finland’s Constitution
states that “democracy entails the right of the individual to participate in
and influence the development of society and his or her living conditions.”
Provisions on consultation and participation are given further weight in
various laws and guidelines including the Act on the Openness of
Government Activities. Bearing in mind this approach the steering
committee formed to make recommendations to implement Finland’s
Artificial Intelligence Programme included among its membership
members of the public.
Facial Recognition
According to recent news reports, Finland’s National Bureau of
Investigation has acknowledged using facial recognition technology in
connection with certain law enforcement activities within the last two
years.659 After initially denying that it had used facial recognition
technology in response to media questioning, the Finnish officials from the
National Bureau of Investigation acknowledged that four members of its
Child Exploitation Investigation Unit had conducted 120 searches of the
Clearview AI system during the 2019 to 2020 time period.
The Office of the Deputy Ombudsman issued a note to the National
Bureau of Investigation regarding the controversial use of Clearview AI
facial recognition technology. In September 2021, the Data Protection
Commissioner warned the National Bureau of Investigation that its police
officers had used a facial recognition technology system without first
verifying that it complied with data security or data protection laws.
Lethal Autonomous Weapons
Human Rights Watch lists Finland among the countries that
participated in discussions regarding the use of fully autonomous lethal
658
OECD, Better Regulation in Europe: Finland, Transparency through consultation
and communication 71 (2010), https://www.oecd.org/gov/regulatorypolicy/45054502.pdf
659
Nord News, The Data Protection Commissioner raped the Finnish police for a
controversial facial identity application (Sept. 29, 2021),
https://nord.news/2021/09/29/the-data-protection-commissioner-raped-the-finnish-policefor-a-controversial-facial-identity-application/
166
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
weapons.660 Finland also participated in the Convention of Conventional
Weapons meetings held each year between 2014 and 2019. According to
the August 2020 report, at the 2014 UN General Assembly, Finnish officials
noted that the issue of lethal autonomous weapons systems is “a complex
issue.” Finnish officials cautioned that the “development of weapons and
means of warfare where humans are completely out of the loop would pose
serious risks from the ethical and legal viewpoint,” noted their view that
“humans should always bear the ultimate responsibility when dealing with
questions of life and death.” Finnish officials have not supported proposals
to negotiate a new international treaty to ban or restrict killer robots.
However, in June 2019, Finland’s new government released a coalition
platform that seeks to ban weapons systems based on artificial intelligence.
OECD AI Principles
Finland is a long-time member of the OECD and has adopted OECD
AI Principles,661 committing “to uphold international standards that aim to
ensure AI systems are designed to be robust, safe, fair and trustworthy.”662
Algorithmic Transparency
As an EU Member State, the General Data Protection Regulation
(Regulation (EU) 2016/679) (GDPR) governs the use of data and the
transparency of algorithms in Finland. This means that Finnish citizens have
the right to “meaningful information about the logic involved” as well as
about “the significance and the envisaged consequences”663 of algorithms
used in their services and products. In addition to GDPR, Finland enacted
the Data Protection Act of Finland (Tietosuojalaki), a supplemental
provision to the GDPR effective as of January 1, 2019.664
In December 2018, Juha Sipilä's Government of 2015–2019
submitted to Parliament the Government report on information policy and
artificial intelligence. The report combines two aspects while paying special
660
Human Rights Reports, Stopping Killer Robots: Country Positions on Banning Fully
Autonomous Weapons and Maintaining Human Control (Aug. 10, 2020),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#_ftn1
661
OECD, Finland, https://www.oecd.org/finland/
662
OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence
(May 22, 2019), https://www.oecd.org/science/forty-two-countries-adopt-new-oecdprinciples-on-artificial-intelligence.htm
663
AIDV2020 266.
664
DLA Piper, Data Protection Laws of the World: Finland,
https://www.dlapiperdataprotection.com/index.html?t=law&c=FI#:~:text=The%20protect
ion%20of%20employees'%20privacy,concerning%20privacy%20in%20working%20life.
167
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
attention to ethical issues and 200 people from different sectors of society
were involved in working on the Report.665
In the report, information policy is studied not only from the
viewpoint of information management, but also from the perspectives of the
conditions for the use of information, value basis, ethical principles and
financial impacts. The report constitutes the knowledge basis and a policy,
upon which a roadmap with prioritized concrete actions can be built in the
future.666 Information policies discussed in the report relate to data access
rights, data ownership, copyrights, security and personal data protection. It
constitutes the knowledge basis and a policy, upon which a roadmap with
prioritized actions can be built in the future. The development and
deployment of AI raises uncertainty about the application of the current
legislation on these issues and increases the need for a reform of the
legislative and regulatory framework.667
Human Rights
As one of the signatories to the Universal Declaration of Human
Rights and several international human rights treaties and conventions,668
Finland is committed to protecting human rights, civil liberties, and political
rights. Under Finnish law, these rights are guaranteed and subject to the rule
of law as interpreted by an independent judiciary.
Freedom House gives Finland top scores (100/100) for political
rights and civil liberties, observing that “Finland’s parliamentary system
features free and fair elections and robust multiparty competition.
Corruption is not a significant problem, and freedoms of speech, religion,
665
Finland, Government report on information policy and artificial intelligence (Dec. 5,
2018),
https://vm.fi/documents/10623/7768305/VM_Tiepo_selonteko_070219_ENG_WEB.pdf/
89b99a8e-01a3-91e3-6ada38056451ad3f/VM_Tiepo_selonteko_070219_ENG_WEB.pdf.pdf/VM_Tiepo_selonteko
_070219_ENG_WEB.pdf
666
Ministry of Finance, Information policy report, https://vm.fi/en/information-policyreport
667
European Commission, AI Watch, Finland AI Strategy Report (2020),
https://knowledge4policy.ec.europa.eu/ai-watch/finland-ai-strategy-report_en#regulation
668
These include the International Covenant on Economic, Social and Cultural Rights,
International Covenant on Civil and Political Rights, European Convention on Human
Rights and Protocol amending the Convention for the Protection of Individuals with
regard to Automatic Processing of Personal Data (108+)
168
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and association are respected. The judiciary is independent under the
constitution and in practice.”669
Evaluation
Finland is among the leaders in the development of AI policies. As
one of the first countries to develop a national AI strategy, Finland has
committed to an open and inclusive process for AI with a strong emphasis
on ethics. Finland has also worked closely with UNICEF to develop
internationally applicable policy guidance for the use AI for children.
Finland receives high marks for its defense of political rights and civil
liberties. However, the use of the Clearview facial recognition system
remains controversial.
Freedom House, Freedom in the World 2021 – Finland,
https://freedomhouse.org/country/finland/freedom-world/2021
669
169
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
France
National AI Strategy
France’s national Strategy on Artificial Intelligence670 (AI) aims to
make France a world leader in AI. “AI will raise a lot of issues in ethics, in
politics, it will question our democracy and our collective preferences,”
stated French President Emmanuel Macron in 2018.671 ”If you want to
manage your own choice of society, your choice of civilization, you have
to be able to be an acting part of this AI revolution.”
France’s AI strategy sets out four objectives672: Reinforcing the AI
ecosystem to attract the very best talents, (2) Developing an open data
policy, especially in sectors where France already has the potential for
excellence, such as healthcare, (3) Creating a regulatory and financial
framework favoring the emergence of “AI champions,” and (4) Promoting
AI regulation and ethics, to ensure to high standard and acceptability for
citizens. This includes supporting human sciences research on ethics of use,
making all algorithms used by the State public, including admission to
higher education, and encouraging AI’s openness to diversity.
The national AI strategy builds on the work of France Strategy,673
the work of the Commission Nationale de l’Informatique et des Libertés674
(CNIL), and the Villani675 report For a Meaningful Artificial Intelligence:
President of France, France’s new national strategy for artificial intelligence - Speech
of Emmanuel Macron (March 29, 2018), https://www.elysee.fr/emmanuelmacron/2018/03/29/frances-new-national-strategy-for-artificial-intelligence-speech-ofemmanuel-macron.en
671
Nicholas Thompson, Emmanuel Macron Talks to WIRED About France's AI Strategy
(Mar. 31, 2018), https://www.wired.com/story/emmanuel-macron-talks-to-wired-aboutfrances-ai-strategy
672
Government of France, Artificial Intelligence: “Making France a leader”
(Mar. 30, 2018), https://www.gouvernement.fr/en/artificial-intelligence-making-francea-leader
673
France Stratégie, the strategy department attached to the French Prime Minister,
released a synthesis France intelligence artificielle report in March 2017.
https://www.enseignementsup-recherche.gouv.fr/cid114739/rapport-strategie-france-i.a.pour-le-developpement-des-technologies-d-intelligence-artificielle.html
674
The CNIL (National Commission on Computer Technology and Civil Liberties)
organized a public debate and produced a report on “the ethical stakes of algorithms and
artificial intelligence” in December 2017 which recommends six concrete actions.
https://www.cnil.fr/en/algorithms-and-artificial-intelligence-cnils-report-ethical-issues
675
Cedric Villani is a French mathematician, Fields Medal winner and Member of
Parliament. Part 5 of his report focuses on ethical considerations of AI and notably
includes proposals to open the “black box”, implement ethics by design, and set up an AI
Ethics Committee.
670
170
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Towards a French and European strategy (March 2018).676 The National
Coordinator for AI works with all administrations, centers and research
laboratories dedicated to AI.677
The Health Data Hub Controversy
In pursuit of the objective of an open data policy, France launched
in December 2019678 the Health Data Hub679(HDH) to facilitate data sharing
and foster research. Pulling together 18 public databases of patient data, the
HDH could, in the future, be connected with environmental, patient
compliance and quality of life data to enable consideration of all the data
surrounding a patient.680 The HDH’s compiled health data is hosted by
Microsoft.681
Following the Schrems II decision in July 2020 that invalidated the
Privacy Shield, France’s highest administrative court (the Conseil d’État)
considered a request for the suspension of the HDH. In October, the Judge
rejected the request. The judge observed that “personal data hosted in the
Netherlands under a contract with Microsoft cannot legally be transferred
outside the European Union. While the risk cannot be completely excluded
that the American intelligence services request access to this data, it does
not justify, in the very short term, the suspension of the Platform, but
Cedric Villani, For a Meaningful Artificial Intelligence: Toward a French and
European Strategy (Mar. 2018),
https://www.aiforhumanity.fr/pdfs/MissionVillani_Report_ENG-VF.pdf
677
Government of France, Prime Minister, Nomination de M. Renaud VEDEL. comme
Coordinateur national pour l’intelligence artificielle (Mar. 9, 2020),
https://www.gouvernement.fr/sites/default/files/document/document/2020/03/communiqu
e_de_presse_de_m._edouard_philippe_premier_ministre__nomination_de_m._renaud_vedel_comme_coordinateur_national_pour_lintelligence_art
ificielle_-_09.03.2020.pdf
678
Government of France, Ministry of Solidarity and Health, Création officielle du
Health data hub (Dec. 2, 2019), https://solidaritessante.gouv.fr/actualites/presse/communiques-de-presse/article/creation-officielle-duhealth-data-hub
679
Health Data Hub (under “reconstruction”), https://www.health-data-hub.fr/;
Government of France, Ministry of Solidarity and Health, Le Health data hub est
officiellement créé (Dec. 2, 2019), https://solidarites-sante.gouv.fr/IMG/pdf/191202__cp_-_health_data_hub.pdf
680
Opus Line, Heath Data Hub: An Ambitious French Initiative for Tomorrow’s Health
(Mar. 25, 2019), https://www.opusline.fr/health-data-hub-an-ambitious-french-initiativefor-tomorrows-health/
681
Florian Dèbes, L'Etat choisit Microsoft pour les données de santé et crée la
polémique, Les Ecos (June 4, 2020) https://www.lesechos.fr/tech-medias/hightech/letatchoisit-microsoft-pour-les-donnees-de-sante-et-cree-la-polemique-1208376
676
171
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
requires special precautions to be taken, under the supervision of the
CNIL.”682
Following the decision concerning data protection and the Health
Data Hub, the CNIL announced it will advise public authorities on the
implementation of appropriate guarantees and will ensure that use of the
HDH for research projects related to the health crisis is really necessary.683
The press reported in October 2020 that the debates are far from over
since the CNIL and the Conseil d’État do not have the same analysis of the
situation. According to the CNIL, the end of the Privacy Shield requires an
urgent change of host for the personal data. According to the Conseil d’Etat,
the risks are hypothetical and not urgent.684 A recent CNIL’s draft
determination, pending validation by a commissioner, would essentialy
prevent implementation of the HDH.685 According to Mediapart, at the end
of November, the Minister of Health and Solidarity, Olivier Véran,
responded to the President of the CNIL that he would put an end to
Microsoft's hosting of the Health Data Hub within two years.686
In 2022, it is highly likely health data will still be at the center of
concern for France’s DPA, the CNIL. Indeed, this data, known as sensitive
data in European law, has been widely collected and processed by many
different data controllers and processors in the current health context to
fulfill different purposes, such as access to the workplace for certain
professions, allowing establishment of the sanitary pass, monitoring the
evolution of the pandemic, establishing vaccination campaigns, deepening
682
Le Conseil d'Etat, Health Data Hub et protection de données personnelles: des
précautions doivent être prises dans l’attente d’une solution pérenne (Oct. 14, 2020),
https://www.conseil-etat.fr/actualites/actualites/health-data-hub-et-protection-dedonnees-personnelles-des-precautions-doivent-etre-prises-dans-l-attente-d-une-solutionperenne
683
CNIL, Le Conseil d’État demande au Health Data Hub des garanties supplémentaires
pour limiter le risque de transfert vers les États-Unis (Oct. 14, 2020),
https://www.cnil.fr/fr/le-conseil-detat-demande-au-health-data-hub-des-garantiessupplementaires
684
Informatique News, Divergences sur le Health Data Hub (Oct. 19, 2020),
https://www.informatiquenews.fr/divergences-sur-le-health-data-hub-les-annonces-dezoomtopia-cohesity-sassocie-a-aws-des-iphone-12-en-5g-le-teletravail-en-questionnetapp-insight-74042
685
Alice Vitard, Les détails de la mise en œuvre du Health Data Hub ne conviennent pas
à la Cnil, L’Usine Digitale, (Nov. 14, 2020), https://www.usine-digitale.fr/article/lesdetails-de-la-mise-en-uvre-du-health-data-hub-ne-conviennent-pas-a-la-cnil.N1024349 686
Mediapart, Health Data Hub: Véran s’engage à retirer l’hébergement à Microsoft
d’ici «deux ans» (Nov. 22, 2020),
https://www.mediapart.fr/journal/france/221120/health-data-hub-veran-s-engage-retirer-lhebergement-microsoft-d-ici-deux-ans.
172
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
research, implementing health protocols for people suffering from COVID19 and more.
In view of the numerous data breaches that have occurred in this
field and the numerous interests this type of data can arouse, the verification
of the conformity of the data processing implemented and security measures
taken should still give rise to numerous controls by CNIL agents.
Similarly, it is anticipated employee monitoring systems will be
subject to increased vigilance by the CNIL. As a result of the pandemic,
many employees are now working in a hybrid work environment, with
periods of office work and periods of working at home. This requires
companies to adapt, since they must allow them to continue carrying out
their remote missions under the same conditions as if they were in the office.
Therefore, companies must give them access to personal data, such as data
on customers, prospects, suppliers or even employees of the organization,
under appropriate security conditions, all while controlling their activity.687
Launch of National AI Research Institutes
France has established interdisciplinary institutes for AI (3IA
institutes) to bring researchers together and to focus on academic
excellence, interdisciplinary research and collaboration with industries.
Each institute has been given areas of focus: MIAI in Grenoble focuses on
health, environment and energy. 3IA Côte d’Azur in Nice focuses on health
and the development of the territories. The PRAIRIE institute in Paris
focuses on health, transport and the environment. The ANITI in Toulouse
focuses on transport, the environment and health. It is reported that EUR
225 million will be spent on 3IA research projects in total.688
AI Cloud
In April 2020, France and Germany launched Gaia-X, a platform
joining up cloud-hosting services from dozens of French and German
companies, to allow business to move their data freely under Europe's data
processing rules. "We are not China, we are not the United States — we are
European countries with our own values and our own European interests
that we want to defend” said French Economy Minister Bruno Le Maire. A
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
688
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
687
173
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
prototype of “Gaia-X” is set to be released early 2021.689 Gaia-X will play
a key role in the European data strategy, the Commission said, as its success
lies in the ability to harmonize rules on data sharing to allow for
upscaling.690
Gaia-X will be open to American, Chinese and Indian technology
companies. Digital Europe, which counts among its members Google,
Apple and Facebook, submitted his application to be a member of this
collective of providers last October.691
National Pilot Committee for Digital Ethics
With regard to AI regulation and ethics (objective 4), in July 2019,
the Prime Minister asked the French National Consultative Committee on
Bioethics (CCNE) to launch a pilot initiative dedicated to Digital
Ethics. The National Pilot Committee for Digital Ethics (NPCDE) created
in December 2019 “shall submit initial contributions on the ethics of digital
sciences, technologies, uses and innovations and determine relevant
equilibria for the organization of public debate on digital ethics and artificial
intelligence.’’ It is also tasked to maintain ethical oversight and to raise
awareness, inform and assist individuals, companies, administrations,
institutions, etc., in their decision-making process.692 A recommendation for
the formation of a permanent body is expected early 2021.
The 27-member multidisciplinary pilot Committee has started work,
at the request of the Prime Minister on the ethical issues raised by chatbots,
autonomous car and medical diagnosis and health AI. Since its creation the
NPCDE has issued three watch bulletins on digital ethical issues in the
Marion Simon Rainaurd, Gaia-X : où en est le projet de méta-cloud européen qui veut
protéger vos données? 01net (Nov. 13, 2020), https://www.01net.com/actualites/gaia-xou-en-est-le-projet-de-meta-cloud-europeen-qui-veut-proteger-vos-donnees1991857.html
690
Janosch Delcker and Melissa Heikkilä, Germany, France launch Gaia-X platform in
bid for ‘tech sovereignty,’ Politico (June 5, 2020),
https://www.politico.eu/article/germany-france-gaia-x-cloud-platform-eu-techsovereignty/
691
Alice Vitard, Le projet de cloud européen Gaia-X ouvert aux entreprises américaines,
chinoises et indienne, L’Usine Nouvelle (Oct. 16, 2020), https://www.usinedigitale.fr/article/le-projet-de-cloud-europeen-gaia-x-ouvert-aux-entreprises-americaineschinoises-et-indiennes.N1017634
692
Claude Kirchner, The French National Committee for Digital Ethics (Feb. 24, 2020),
https://ai-regulation.com/the-french-national-committee-for-digital-ethics/
689
174
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
COVID-19 health crisis.693 In July 2020, the NPCDE issued a call for public
comments on the ethical issues of chatbots.694
However, civil society groups such as Access Now have objected to
government studies that simply propose ethical guidelines rather than hard
law. As the group explains, “There is solid and creative thinking in the
advisory paper that informed the strategy around the ethical and regulatory
challenges posed by AI, but at the moment the proposed solutions largely
involve the creation of groups to study them rather than the proposal of new
or modified norms.”695 Access Now continues, “France’s AI strategy
generally cleaves to the ‘ethics’ framework and makes scant reference to
hard legal constraints on AI development.” The group does note that the
“The Villani report is considerably more detailed about the ethical and legal
challenges posed by AI.”
Fundamental Rights
On another front, the French independent administrative authority
Défenseur des droits (Defender of Rights) and the CNIL have “both, in their
own area of expertise, voiced their concerns regarding the impact of
algorithmic systems on fundamental rights.”696 Following a joint expert
seminar in May 2020, they have called in June 2020 for a collective
mobilization to prevent and address discriminatory biases of algorithms.697
Their report Algorithms: preventing automated discrimination698
stresses that bias can be introduced at every stage of the development and
deployment of AI systems, discusses how algorithms can lead to
discriminatory outcomes and includes recommendations on how to identify
Comité Consultatif National d'Ethique, Opinion (Apr. 14, 2020), https://www.ccneethique.fr/en/publications/national-pilot-committee-digital-ethics-ethics-watch-bulletinno1
694
Comité Consultatif National d'Ethique, Ethical Issues of Conversational Agents (Oct.
31, 2020), https://www.ccne-ethique.fr/sites/default/files/cnpen-chatbots-callparticipation_1.pdf
695
AccessNow, Mapping Regulatory Proposals for Artificial Intelligence in Europe 18
(Nov. 2018),
https://www.accessnow.org/cms/assets/uploads/2018/11/mapping_regulatory_proposals_
for_AI_in_EU.pdf.
696
Defender of Rights and CNIL, Algorithms: preventing automated discrimination
(2020), https://www.defenseurdesdroits.fr/sites/default/files/atoms/files/synth-algos-ennum-16.07.20.pdf
697
CNIL, Algorithms and discrimination: the Defender of Rights, with the CNIL, calls for
collective mobilization (June 2, 2020), https://www.cnil.fr/fr/algorithmes-etdiscriminations-le-defenseur-des-droits-avec-la-cnil-appelle-une-mobilisation
698
Defender of Righta, Algorithms: preventing automated discrimination n. 19 (May
2020), https://www.defenseurdesdroits.fr/sites/default/files/atoms/files/synth-algos-ennum-16.07.20.pdf.
693
175
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and minimize algorithmic biases. The Defender of Rights called on the
government and relevant actors to take appropriate measures to avoid
algorithms that replicate and amplify discrimination.699. In particular, The
Defender of Rights recommends to: i) support research to develop studies
to measure and methods to prevent bias; ii) reinforce algorithms’
information, transparency and explainability requirements; and iii) perform
impact assessments to anticipate algorithms’ discriminatory effects.
Facial Recognition
Facial recognition is a processing of sensitive personal data
prohibited in principle by the GDPR and the French data protection law,
subject to exceptions such as individual’s consent or for important public
interests. In the latter case, facial recognition can be authorized by a Decree
of the Conseil d’État informed by an opinion from the CNIL.
Facial recognition has long been used in France, on a voluntary
basis, for passport control in airports. Facial recognition is also
implemented in some banks and tested in a number of colleges. The French
government is considering the deployment of facial recognition for access
to public services. The ID program, called Alicem,700 to be deployed in
November 2019, was however put on hold following an appeal of NGOs to
the Conseil d’État requesting the annulment of the decree authorizing its
creation. Early November, the Conseil d’État dismissed the appeal.701
In November 2019, the CNIL published guidance on the use of
facial recognition.702 The document, primarily directed at public authorities
in France that want to experiment with facial recognition, presents the
technical, legal and ethical elements that need to be considered.
After recalling that facial recognition, experimental or not, must
comply with the European GDPR and the "police justice" directive, the
Inside Tech Media, French CNIL Publishes Paper on Algorithmic Discrimination
(June 9, 2020), https://www.insideprivacy.com/artificial-intelligence/french-cnilpublishes-paper-on-algorithmic-discrimination/
700
https://www.interieur.gouv.fr/Actualites/L-actu-du-Ministere/Alicem-la-premieresolution-d-identite-numerique-regalienne-securisee (in French) -; Charlotte Jee, France
plans to use facial recognition to let citizens access government services, MIT
Technology Review (Oct. 3, 2020),
https://www.technologyreview.com/2019/10/03/132776/france-plans-to-use-facialrecognition-to-let-citizens-access-government-services/
701
Marion Garreau, Le ministère de l'Intérieur va pouvoir lancer l’application Alicem,
basée sur la reconnaissance faciale, L’Usine Nouvelle (Nov. 5, 2020),
https://www.usinenouvelle.com/editorial/le-ministere-de-l-interieur-va-pouvoir-lancer-lapplication-alicem-basee-sur-la-reconnaissance-faciale.N1024754
702
CNIL, Reconnaissance faciale - pour un débat à la hauteur des enjeux (Nov. 2020),
https://www.cnil.fr/sites/default/files/atoms/files/reconnaissance_faciale.pdf (in French.
699
176
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
CNIL sets out three general requirements: (1) facial recognition can only be
used if there is an established need to implement an authentication
mechanism that ensures a high level of reliability, and there are no other
less intrusive means that would be appropriate; (2) the experimental use of
facial recognition must respect the rights of individuals (including consent
and control, transparency and security); and (3) the use of facial recognition
on an experimental basis must have a precise timeline and be based on a
rigorous methodology setting out the objectives pursued and the criteria for
success.
In
December
2019,
the Observatoire
des
Libertés
703
Numériques and 80 organisations signed an open letter calling on the
French Government and Parliament to ban any present and future use of
facial recognition for security and surveillance purposes.704
Earlier this year the administrative tribunal of Marseille rendered a
decision on facial recognition that ruled illegal a decision by the South-East
Region of France (Provence-Alpes-Côte d’Azur) to test facial recognition
at the entrance of two High schools.705 Following an analysis from the
CNIL,706 the court ruled that there was no opportunity for free and informed
consent and also that there were other, less intrusive means to manage
entrance to high schools. The French NGO La Quadrature du Net brough
the successful challenge to the regional program.707 This was the first
decision ever by a court applying the General Data Protection Regulation
(GDPR) to Facial Recognition Technologies (FRTs).708
The Observatoire des Libertés Numériques federates several French NGOs monitoring
legislation impacting digital freedoms: Le CECIL, Creis-Terminal, Globenet, La Ligue
des Droits de l’Homme (LDH), La Quadrature du Net (LQDN), Le Syndicat des Avocats
de France (SAF), Le Syndicat de la Magistrature (SM).
704
La Quadrature du Net, Joint Letter from 80 organisations: Ban Security and
Surveillance Facial Recognition (Dec. 19, 2019),
https://www.laquadrature.net/en/2019/12/19/joint-letter-from-80-organisations-bansecurity-and-surveillance-facial-recognition/
705
Tribunal Administratif de Marseille, La Quadrature du Net, No. 1901249 (27 Nov.
2020), https://forum.technopolice.fr/assets/uploads/files/15828024229301090394890_1901249.pdf
706
CNIL, Expérimentation de la reconnaissance faciale dans deux lycées : la CNIL
précise sa position (Oct. 29, 2019), https://www.cnil.fr/fr/experimentation-de-lareconnaissance-faciale-dans-deux-lycees-la-cnil-precise-sa-position
707
La Quadrature du Net, First Success Against Facial Recognition in France (Feb. 27,
2020), https://www.laquadrature.net/en/2020/02/27/first-success-against-facialrecognition/
708
AI Regulation, First Decision of a French Court Applying GDPR to Facial
Recognition (Feb. 27, 2020), https://ai-regulation.com/first-decision-ever-of-a-frenchcourt-applying-gdpr-to-facial-recognition/
703
177
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In 2020, real-time facial recognition on public roads in France is still
not authorized. However, many experiments are already taking place, and
companies are positioning themselves, with the Olympic Games in Paris in
2024 in their sights, and a market of seven billion euros at stake.709
Consumer Perspective
According to BEUC, the European consumer association, more than
80% of those polled in France are familiar with Artificial Intelligence and
over 50% respondents agreed that companies use AI to manipulate
consumer decisions.710 BEUC also reported that there is little trust over
authorities to exert effective control over organizations and companies
using AI. More than 60% of respondents in France said users should be able
to say “no” to automated decision-making.
The Global Partnership on AI
In June 2020, Canada and France, and a dozen other countries
announced the Global Partnership on Artificial Intelligence to support
“support the responsible and human-centric development and use of AI in a
manner consistent with human rights, fundamental freedoms, and our
shared democratic values . . .”711 According to the statement, the “GPAI
will be supported by a Secretariat, to be hosted by the OECD in Paris, as
well as by two Centres of Expertise – one each in Montréal and Paris.” The
first expert’s plenary session was held in Montreal December 2020. As
GPAI co-chair, France hosted the 2021 GPAI Summit.712
Algorithmic Transparency
France is subject to the General Data Protection Regulation which
established rights to “meaningful information about the logic involved” as
France Culture, Quand la reconnaissance faciale en France avance masquée (Sept. 4,
2020), https://www.franceculture.fr/societe/quand-la-reconnaissance-faciale-en-franceavance-masquee
710
BEUC, Artificial Intelligence, what consumers say: Findings and policy
recommendations of a multi-country survey on AI, (Sept. 7, 2020)
https://www.beuc.eu/publications/beuc-x-2020078_artificial_intelligence_what_consumers_say_report.pdf
711
France Diplomacy, Joint Statement from founding members of the Global Partnership
on Artificial Intelligence (June 15, 2020), https://www.diplomatie.gouv.fr/en/frenchforeign-policy/digital-diplomacy/news/article/launch-of-the-global-partnership-onartificial-intelligence-by-15-foundingdevelopment/news/2020/06/joint-statement-fromfounding-members-of-the-global-partnership-on-artificial-intelligence.html
712
GPAI, The Magazine, The GPAI Paris Summit (Nov. 11-12, 2021),
https://magazine.gpai.paris/en/
709
178
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
well as about “the significance and the envisaged consequences.”713 The
French data protection agency (CNIL) has published several papers on AI.
A 2018 report followed extensive public outreach in 2017. More than 3,000
people took part in 45 debates and events, organized by 60 partners,
including research centers, public institutions, trade unions, think tanks,
companies).714 The report set out two founding principles – fairness and
vigilance -- six recommendations, and six concerns. The work of the CNIL
also contributed to the Declaration on Ethics and Data Protection in AI,
adopted by the Global Privacy Assembly in 2018, which emphasized
fairness and accountability.715 In the 2020 paper with the Defender of
Rights, the CNIL went into more details concerning the transparency
obligations of those who are responsible for AI systems.716
Following the assassination in October 2020 of history professor
Samuel Paty, the Secretary of State for digital, Cédric O, wrote in a blog
that “the opacity of the functioning of (social media) algorithms and their
moderation is a societal and democratic aberration.” He added “it is also
essential that full transparency be observed vis a vis the public authorities
as regards the principles governing in detail the choices made by their
moderation algorithms, whether it is about online hatred or dissemination
of false information.”717
OECD/G20 AI Principles
France endorsed the OECD and the G20 AI Principles. France is
also co-hosting the Global Partnership for AI.718 France is a signatory to
many international human rights treaties and conventions.
[GDPR Art. 22, Art. 13.2.f]
CNIL, Algorithms and artificial intelligence: CNIL’s report on the ethical issues (May
25, 2018), https://www.cnil.fr/en/algorithms-and-artificial-intelligence-cnils-reportethical-issues
715
Global Privacy Assembly, Declaration on Ethics and Data Protection in AI (Oct. 23,
2018), http://globalprivacyassembly.org/wpcontent/uploads/2019/04/20180922_ICDPPC-40th_AI-Declaration_ADOPTED.pdf
716
CNIL, Algorithmes et discriminations : le Défenseur des droits, avec la CNIL, appelle
à une mobilisation collective (May 2020), https://www.cnil.fr/fr/algorithmes-etdiscriminations-le-defenseur-des-droits-avec-la-cnil-appelle-une-mobilisation
717
Cédric O, Régulations, Medium.com (Oct. 20, 2020),
https://medium.com/@cedric.o/r%C3%A9gulations-657189f5d9d2
718
The Government of France, Launch of the Global Partnership on Artificial
Intelligence (June 17, 2020), https://www.gouvernement.fr/en/launch-of-the-globalpartnership-on-artificial-intelligence
713
714
179
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Human Rights
France typically ranks among the top nations in the world for the
protection of human rights and transparency.719 Freedom House reports,
“The French political system features vibrant democratic processes and
generally strong protections for civil liberties and political rights. However,
due to a number of deadly terrorist attacks in recent years, successive
governments have been willing to curtail constitutional protections and
empower law enforcement to act in ways that impinge on personal
freedoms.”
Lethal Autonomous Weapons
President Macron declared in an interview that he is “dead against”
the deployment of lethal autonomous weapons. 720“You always need
responsibility and assertion of responsibility.” However, the French
government has only proposed the adoption of a nonbinding declaration to
curtail Lethal Autonomous Weapons (LAWS), and is opposed to the idea
of a new international treaty on the issue,721 though an earlier French
initiative led to annual international discussions on LAWS ) within the
framework of the Convention on Certain Conventional Weapons.722
Evaluation
France is among the leaders in national AI policies. France has
endorsed the OECD/G20 AI Principles and is a co-host for the Global
Partnership on AI. French authorities in charge of human rights, data
protection and ethics are actively involved in AI policy and have
published practical
guidance
regarding
facial
recognition
and algorithmic transparency. However, public information about progress
toward the national strategy on AI is not readily available. While there is,
Freedom House Report: France (2020), https://freedomhouse.org/country/france
Nicholas Thompson, Emmanuel Macron Talks to Wired About France’s AI Strategy,
Wired (Mar. 31, 2018), https://www.wired.com/story/emmanuel-macron-talks-to-wiredabout-frances-ai-strategy/
721
Armes : Il faut négocier un traité d’interdiction des armes létales
autonomes [Weapons: We Must Negotiate a Treaty to Ban Lethal Autonomous Weapons],
Human Rights Watch (Aug. 27, 2018), https://www.hrw.org/fr/ news/2018/08/27/armesil-faut-negocier-un-traite-dinterdiction-des-armes-letales-autonomes, archived
at https://perma.cc/JC23-3BFB
722
Presentation and Position of France, MISSION PERMANENTE DE LA FRANCE
AUPRÈS DE LA CONFÉRENCE DU DÉSARMEMENT À GENÈVE [PERMANENT
REPRESENTATION OF FRANCE TO THE CONFERENCE ON DISARMAMENT IN
GENEVA] (Aug. 3, 2016), https://cd-geneve.delegfrance.org/Presentation-and-positionof-France-1160, archived at https://perma.cc/6XD3-U82R.
719
720
180
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
at the moment, no express support for the Universal Guidelines, France’s
AI policies share similarities to those recommended in the UGAI.
181
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Germany
National AI Strategy
The German government published its national AI strategy in
November 2018.723 The three main goals are:
1) “to make Germany and Europe a leading centre for AI and thus
help safeguard Germany’s competitiveness in the future”
2) To ensure “a responsible development and use of AI which serves
the good of society”
3) To “integrate AI in society in ethical, legal, cultural and
institutional terms in the context of a broad societal dialogue and
active political measures”
The guiding slogan for the strategy is “AI made in Germany.” One
section of the AI Strategy states: “The Federal Government advocates using
an “ethics by, in and for design” approach throughout all development
stages and for the use of AI as the key element and hallmark of an ‘AI made
in Europe’ strategy.” The Strategy continues, “The Federal Government is
engaging in dialogue with national and international bodies, including the
Data Ethics Commission or the EU Commission’s High-Level Expert
Group on AI and will take into account the recommendations of these
bodies as it develops standards on ethical aspects at German and European
level.”
The German government further emphasizes transparency for the
development of AI to ensure civil rights as well as maintain trust in
businesses and institutions. The AI Strategy proposes “government
agencies or private-sector auditing institutions that verify algorithmic
decision-making in order to prevent improper use, discrimination and
negative impacts on society.” AI ethics is a core component of the AI
Strategy.724
There are several programs underway to implement the National AI
Strategy. According to the OECD, there are approximately 29 initiatives on
AI across several topics and institutions.725 They range from the ethical
guidelines to initiatives that foster fruitful business environments. There are
four that specifically focus on ethics.
Die Bundesregierung, Artificial Intelligence Strategy, (Nov. 2018),
https://www.bmbf.de/files/Nationale_KI-Strategie.pdf
724
The Federal Government of Germany, Artificial Intelligence Strategy (Nov. 2018),
https://www.ki-strategie-deutschland.de/home.html?file=files/downloads/Nationale_KIStrategie_engl.pdf
725
OECD.ai, AI in Germany, https://oecd.ai/dashboards/countries/Germany/
723
182
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
First, the Ethical Guidelines for Automated and Connected Driving
set out 20 ethical principles for autonomous and semi-autonomous
vehicles.726 This was among the first guidelines worldwide to establish
ethical guidelines for connected vehicular traffic. The Ethical Guidelines
led to an action plan and the “creation of ethical rules for self-driving cars”
that was adopted by the Federal Government.727
Second, the German AI Observatory forecasts and assesses AI
technologies’ impact on society. The AI Observatory also develops
regulatory frameworks that help deal with the rapidly changing labor market
in an attempt to ensure that social aspects of these changes are not
neglected.728
Third, the Ethical, Legal and Social Aspects of Modern Life
Sciences Funding Priority, launched originally in 1997, funds research with
the goal of establishing “findings regarding the opportunities and risks
presented by modern life sciences” and developing a basis for discourse
amongst involved stakeholders.729
Fourth, the Federal Ministry for Economic Cooperation and
Development launched the Development Cooperation initiative FAIR
Forward in 2019. The initiative aims to promote more “open, inclusive and
sustainable approach to AI on an international level” by “working together
with five partner countries: Ghana, Rwanda, South Africa, Uganda and
India.” The FAIR Forward goals are to: Strengthen Technical Know-How
on AI, Remove Entry Barriers to AI, and Develop Policy Frameworks ready
for AI. Several projects are underway in partner countries.730
Further, the Federal Ministry for Economic Affairs and Energy
launched a Regulatory Sandboxes initiative in 2018. This initiative focuses
Federal Ministry of Transport and Digital Infastructure, Ethics Commission:
Automated and Connected Driving (2017),
https://www.bmvi.de/SharedDocs/EN/publications/report-ethics-commission-automatedand-connected-driving.pdf
727
Federal Ministry of Transport and Digital Infrastructure, Automated and Connected
Driving, https://www.bmvi.de/EN/Topics/Digital-Matters/Automated-ConnectedDriving/automated-and-connected-driving.html
728
Denkfabrik: Digitale Arbeitsgesellschaft, Policy Lab Digital, Work & Society: Reimaging Work, https://www.denkfabrik-bmas.de/en/about-us/policy-lab-digital-worksociety-re-imagining-work
729
Federal Ministry of Education and Research, The ELSA funding initiative (June 2016),
https://www.gesundheitsforschungbmbf.de/files/bmbf_flyer_ELSA_funding_initiative_e.pdf
730
Toolkit Digitalisierung, FAIR Forward – Artificial Intelligence for All, https://toolkitdigitalisierung.de/en/fair-forward/
726
183
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
on “testing innovation and regulation which enable digital innovations to be
tested under real-life conditions and experience to be gathered.”731
In response to the White Paper on AI, Germany called for tighter
regulation of AI on the EU level. The German government stated they
welcome new regulations but want more specific definitions and tighter
requirements for data storage, more focus on information security and more
elaborate definitions of when human supervision is needed.732
In December 2020, Germany published an updated report on the
German AI strategy. With this update, the federal government is responding
to current developments and high-priority topics such as the COVID-19
pandemic, environmental and climate protection; and complementing the
strategy with further measures.733
Public Participation
One AI initiative, Plattform Lernende Systeme (Platform for
Artificial Intelligence), focuses specifically on fostering dialogue between
different stakeholders, like civil society, government and business on the
topic of self-learning systems. The Platform for AI also aims to “shape selflearning systems to ensure positive, fair and responsible social coexistence”
as well as strengthen skills for developing and using self-learning
systems.734 The IT Security, Privacy, Legal and Ethical Framework working
group has published two papers concerning AI and Discrimination as well
as AI and IT Security.735
To inform the public about AI policy, the government created a
website to provide information on AI strategy implementation and new
731
Federal Ministry for Economic Affairs and Energy, Regulatory Sandboxes – Testing
Environments for Innovation and Regulation (June 2019),
https://www.bmwi.de/Redaktion/EN/Dossier/regulatory-test-beds-testing-environmentsfor-innovation-and-regulation.html
732
Die Bundesregierung, Stellungsnahme der Bundesregierung der Bundesrepublik
Deutschland zum Weissbuch zur Künstlichen Intelligenz – ein europäisches Konzept für
Exzellenz und Vertrauen (2020), https://www.ki-strategiedeutschland.de/files/downloads/Stellungnahme_BReg_Weissbuch_KI.pdf
733
OECD, State of Implementation of the OECD AI Principles Insights from National AI
Policies 69 (June 2021), https://www.oecd.org/digital/state-of-implementation-of-theoecd-ai-principles-1cd40c44-en.htm
734
Lernende Systeme, Mission Statement, https://www.plattform-lernendesysteme.de/mission-statement.html
735
Lernende Systeme, WG 3: IT Security, Privacy, Legal and Ethical Framework,
https://www.plattform-lernende-systeme.de/wg-3.html
184
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
policy developments.736 Plattform Lernende Systeme also offers a map that
shows, by region, AI developments across Germany.737
There was a Bundestag EnqueteCommission comprising in equal
parts of parliamentary representatives and experts called the “Study
Commission on Artificial Intelligence, Social Responsibility and
Economic, Social and Ecological Potential.”738 Their aim was to develop
recommendations on AI and its potential “for example with regard to our
value systems, fundamental and human rights, and the benefits for society
and the economy.” Some of their meetings were broadcasted on
parliamentary television or could be attended in person. After two years of
work, the Enquete Commission presented its final report (19/23700) to the
Parliament (Bundestag) on October 28, 2020. The Commission findings
were debated in the Bundestag on November 5, 2020.739
Data Ethics Commission
In 2018 a Data Ethics Commission was established to “build on
scientific and technical expertise in developing ethical guidelines for the
protection of the individual, the preservation of social cohesion, and the
safeguarding and promotion of prosperity in the information age.”740 In 2020
the Commission recommended to the German parliament that
sustainability, justice and solidarity, democracy, security, privacy, selfdetermination and human dignity should be the ethical and legal principles
that guide the regulation of AI.741 The Data Ethics Commission suggested a
risk-based approach to the regulation of AI, which distinguishes five levels
Die Bundesregierung, https://www.ki-strategie-deutschland.de/home.html
Lernende Systeme, Artificial Intelligence in Germany, https://www.plattformlernende-systeme.de/map-on-ai.html
738
Deutscher Bundestag, Study Commission, Artificial Intelligence, Social Responsibility
and Economic, Social and Egological Potential,
https://www.bundestag.de/webarchiv/Ausschuesse/ausschuesse19/weitere_gremien/enqu
ete_ki
739
Unterrichtung der Enquete-Kommission Künstliche Intelligenz – Gesellschaftliche
Verantwortung und wirtschaftliche, soziale und ökologische Potenziale, Bericht der
Enquete-Kommission Künstliche Intelligenz – Gesellschaftliche Verantwortung und
wirtschaftliche, soziale und ökologische Potenziale (Oct. 28, 2020),
https://dserver.bundestag.de/btd/19/237/1923700.pdf
740
Bundesministerium der Jusitz und für Verbraucherschutz, Data Ethics Commission,
https://www.bmjv.de/DE/Themen/FokusThemen/Datenethikkommission/Datenethikkom
mission_EN_node.html
741
Datenethikkommission, Opinion of the Data Ethics Commission (Jan. 2020),
https://www.bmjv.de/SharedDocs/Downloads/DE/Themen/Fokusthemen/Gutachten_DE
K_EN_lang.pdf?__blob=publicationFile&v=3
736
737
185
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of criticality in a “criticality pyramid” and respective measures in its riskadapted regulatory system for the use of algorithmic systems.
The German consumer organization vzbv favored the creation of the
Commission and strongly supported the recommendations, as did the main
German industry body Bundesverband der Deutschen Industrie (BDI).742
The vzbv further emphasized that the aim of ADM (Automated
DecisionMaking) regulation must be to ensure compliance with existing
laws. Toward that goal, “it must be possible for supervisory authorities to
scrutinise and verify the legality of ADM systems and their compliance with
existing laws so that they can impose penalties if the law is infringed.”743
Vzbv also noted it is “important to ensure consumers’ self-determination
when making decisions, to strengthen consumers’ confidence in ADM
systems by creating transparency and to foster competition and innovation.”
Further, in 2018 the Cabinet Committee on Digitisation was founded
with the goal of advising the Federal Government on how to best implement
the National AI Strategy. The Committee is comprised of AI experts in
science and business. An exchange between politics and national as well as
international experts is also in the forefront of their activities.744745
Facial Recognition
In 2018 facial recognition technology at a large crossing in Berlin,
set up by the government, sparked opposition from civil society.746 There
was further outcry in 2020, when Der Spiegel wrote that there are plans to
set up cameras capable of identifying people at 134 train stations and 14
airports. 747 In late 2021, Germany’s incoming coalition government said it
Communication between the Editor and Isabelle Buscke, vzbz Nov. 27, 2020 (on file).
Vzbv, Artificial Intelligence: Trust is Good, Control is Better (2019),
https://www.vzbv.de/sites/default/files/2019_vzbv_factsheet_artificial_intelligence.pdf
744
Die Bundesregierung, Der Digitalrat: Experten, die uns antreiben,
https://www.bmvi.de/SharedDocs/EN/publications/report-ethics-commission-automatedand-connected-driving.pdf
745
Die Bundesregierung, Digitalisierung wird Chefsache,
https://www.bundesregierung.de/breg-de/aktuelles/digitalisierung-wird-chefsache1140420
746
Janosch Delcker, Big Brother in Berlin, Politico (Sept. 13, 2018),
https://www.politico.eu/article/berlin-big-brother-state-surveillance-facial-recognitiontechnology/
747
Phillipp Grüll, Germany’s plans for automatic facial recognition meet fierce criticism,
Euractiv (Jan. 10, 2020), https://www.euractiv.com/section/data-protection/news/germanministers-plan-to-expand-automatic-facial-recognition-meets-fierce-criticism/
742
743
186
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
would exclude biometric recognition in public spaces as well as automated
state scoring systems by AI.748
Predictive Policing
According to AlgorithmWatch, the German government is using AI
to assist in predictive policing both on the federal and state level. One state,
North Rhine-Westphalia is using AI to assist police in identifying child
pornography and preventing suicides in jails. Further, the Federal
government is using AI techniques to identify evolving international crises
in their foreign policy, check identities of immigrants and administer social
services.749
AI Oversight
The German Institute for Human Rights was founded in 2001 by the
German Bundestag (Parliament). The Institute for Human Rights is an
independent national institution, financed by the Bundestag and is
considered a civil society body. The Institute works to ensure the
observation and promotion of human rights by the German government in
Germany and abroad.750 The Institute’s responsibilities include
documentation, consulting politicians and society, human rights education
in Germany, providing a specialized scientific library on human rights,
cooperation with other human rights institutions and promoting dialogue on
human rights issues in Germany.751 The German Institute has not yet
explicitly addressed AI but might do so in the future as have human rights
commissions in other countries.752
In Germany, the data protection authority landscape is quite large.
The private sector is mainly supervised by the states with exception of the
telecommunications and postal sector which is supervised on a federal level.
POLITICO, German coalition backs ban on facial recognition in public places (Nov.
24, 2021), https://www.politico.eu/article/german-coalition-backs-ban-on-facialrecognition-in-public-places/; Alliance for Freedom, Justice, and Sustainability, Dare
More Progress: Coalition agreement 2021 – 2025 between the Social Democratic Party
of Germany (SPD), ALLIANCE 90 / THE GREENS and the Free Democrats (FDP),
https://www.welt.de/bin/Koalitionsvertrag%202021-2025.pdf_bn-235257672.pdf
748
AlgorithmWatch, Automating Society 2020 (Oct. 2020),
https://automatingsociety.algorithmwatch.org/report2020/belgium/)
750
German Institute for Human Rights, Das Institut, https://www.institut-fuermenschenrechte.de/das-institut.
751
German Institute for Human Rights, FAQ, https://www.institut-fuermenschenrechte.de/das-institut/faq
752
See, for example, the activities of the Human Rights Commission of Australia.
749
187
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Every state has a respective data protection authority dedicated to matters
involving the private sector.753
In Bavaria, there is one authority responsible for the private sector
and one for the public sector: the Data Protection Authority of Bavaria for
Private Sector (BayLDA) and the Bavarian Data Protection Commissioner,
which is responsible for enforcing data rights against public authorities and
government agencies.754 In other states, one authority is responsible for all
data protection supervision and enforcement in the state. One example of
this is Hessen where the Hessian Commissioner for Data Protection and
Freedom of Information is responsible for the public authorities,
government agencies as well as the private sector.755
At the federal level, the Federal Commissioner for Data Protection
and Freedom of Information (BfDI) is responsible for the supervision of all
public bodies that belong to the federal government and the
telecommunication and postal services companies.756
In December 2021 the Telecommunication Telemedia Data
Protection Act (TTDSG) came into effect. Amongst other things, it clarifies
application of the GDPR and ePrivacy Directive. The TTDSG
(Telekommunikation-Telemedien-Datenschutz-Gesetz) contains regulations
about cookie management and Personal Information Management
Systems.757 Further the German Civil Code Article 327q entered into effect
and is meant to deal with cases where a consumer gives their personal data
in order to access some service. It is highly regarded by privacy protection
organisations for enhancing the German Civil Code.758
Germany was a sponsor of the 2018 GPA Declaration on Ethics and
Data Protection in Artificial Intelligence and a primary sponsor of the 2020
Landesbeauftragte für Datenschutz und Informationsfreiheit NordrheinWestfalen, Datenschutzaufsichtsbehörden für den nicht-öffentlichen Bereich,
https://www.ldi.nrw.de/mainmenu_Service/submenu_Links/Inhalt2/Aufsichtsbehoerden/
Aufsichtsbehoerden.php.
754
Datenschutz Bayern, Bavarian Data Protection Commissioner,
https://www.datenschutz-bayern.de; BayLDA - Offizielle Webseite,
https://www.lda.bayern.de/de/index.html
755
Datenschutz Hessen, Zuständigkeit des Hessischen Beauftragten für Datenschutz und
Informationsfreiheit, https://datenschutz.hessen.de/ueber-uns/zuständigkeit-deshessischen-datenschutzbeauftragten.
756
Der Bundesbeauftragte für den Datenschutz und die Informationsfreiheit, Aufgaben
und Befugnisse, https://www.bfdi.bund.de/DE/BfDI/Artikel_BFDI/AufgabenBFDI.html
757
TTDSG, Telekommunikation-Telemedien-Datenschutz-Gesetz, https://gesetz-ttdsg.de
758
Federal Ministry of Justice, German Civil Code, https://www.gesetze-iminternet.de/englisch_bgb/
753
188
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Resolution on Accountability in the Development and Use of Artificial
Intelligence.759
Algorithmic Transparency
Germany is subject to the General Data Protection Regulation which
established rights to “meaningful information about the logic involved” as
well as about “the significance and the envisaged consequences.”760
According to AlgorithmWatch,761 the data protection agencies of the federal
government and eight German federal states stated that greater transparency
in the implementation of algorithms in the administration was indispensable
for the protection of fundamental rights.762 The agencies demanded that if
automated systems are used in the public sector, it is crucial that processes
are intelligible, and can be audited and controlled. In addition, public
administration officials have to be able to provide an explanation of the
logic of the systems used and the consequences of their use. Self-learning
systems must also be accompanied by technical tools to analyse and explain
their methods. An audit trail should be created, and the software code should
be made available to the administration and, if possible, to the public.
According to the position paper, there need to be mechanisms for citizens
to demand redress or reversal of decisions, and the processes must not be
discriminating. In cases where there is a high risk for citizens, there needs
to be a risk assessment done before deployment. Very sensitive systems
should require authorisation by a public agency that has yet to be created.
In 2019 the Ministry of Education and Research started a funding
priority for AI R&D projects on explainability and transparency. The
Ministry stated that improving explainability and transparency are two of
International Conference on Data Protection and Privacy Commissioners, Declaration
on Ethics and Data Protection in Artificial Intelligence (Oct. 23, 2018),
https://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf; Global Privacy Assembly, Resolution on
Accountability in the Development and Use of Artificial Intelligence (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/FINAL-GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN-1.pdf.
759
[GDPR Art. 22, Art. 13.2.f]
Algorithm Watch, Automating Society: Germany (Jan. 29, 2019),
https://algorithmwatch.org/en/automating-society-germany/
762
Freedom of Information Commissioners in Germany, “Transparenz der Verwaltung
beim Einsatz von Algorithmen für gelebten Grundrechtsschutz unabdingbar“ (Oct. 16,
2018),
https://www.datenschutzzentrum.de/uploads/informationsfreiheit/2018_PositionspapierTransparenz-von-Algorithmen.pdf
760
761
189
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the Federal government’s central research goals.763 Funding is “aimed at
collaborative projects between science and industry in an interdisciplinary
composition.”764
OECD/G20 Principles and Global Partnership on AI
Germany is a member of the OECD and endorsed the OECD and
the G20 AI Principles. In 2020, Germany joined 14 other countries to
announce the Global Partnership on Artificial Intelligence to “support the
responsible and human-centric development and use of AI in a manner
consistent with human rights, fundamental freedoms, and our shared
democratic values.”765 In 2021, the OECD noted several examples of
Germany’s implementation of the OECD AI Principles, including
guidelines for trustworthy AI that are largely in line with the OECD AI
Principles (Germany’s Data Ethics Commission ethics recommendations),
the establishment of a dedicated body to coordinate AI strategy, annual
evaluation of national AI strategies, and fostering partnership between
public and private research organization.766
Human Rights
According to Freedom House, Germany is one of the top countries
in the world for the protection of political rights and civil liberties, receiving
as score of 94/100, unchanged from the previous year.767 Freedom House
reports that, “Germany is a representative democracy with a vibrant
political culture and civil society. Political rights and civil liberties are
largely assured both in law and practice.” 768
Bundesministerium für Bildung und Forschung, KI-Erklärbarkeit und Transparenz,
https://www.softwaresysteme.pt-dlr.de/de/ki-erkl-rbarkeit-und-transparenz.php
764
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
765
Federal Ministry for Economic Affairs and Energy & Federal Ministry for Social
Affairs and Work, Joint Press Release: Germany is a founding Member of the Global
Partnership on Artificial Intelligence (June 15, 2020),
https://www.bmwi.de/Redaktion/EN/Pressemitteilungen/2020/20200615-germany-is-afounding-member-of-the-global-partnership-on-artificial-intelligence.html
766
OECD, State of Implementation of the OECD AI Principles Insights from National AI
Policies 10, 14, 15, 65 (June 2021), https://www.oecd.org/digital/state-ofimplementation-of-the-oecd-ai-principles-1cd40c44-en.htm
767
Freedom House, Freedom in the World 2021– Germany (2021),
https://freedomhouse.org/country/germany/freedom-world/2021
768
Freedom House, Freedom in the World 2020 – Germany (2020),
https://freedomhouse.org/country/germany/freedom-world/2020
763
190
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Lethal Autonomous Weapons
The German government’s 2018 coalition agreement stated that
it “rejects autonomous weapon systems devoid of human control”
and calls for a global ban.769 Further in cooperation with the French
government, the German government, published a joint statement on Lethal
Autonomous Weapons at the “Meeting of the Group of Governmental
Experts on Lethal Autonomous Weapons Systems.” They write: “At the
heart of our proposal is the recommendation for a political declaration,
which should affirm that State parties share the conviction that humans
should continue to be able to make ultimate decisions with regard to the use
of lethal force and should continue to exert sufficient control over lethal
weapons systems they use.”770
The new German government’s 2021 coalition agreement states:
“Biometric recognition in public spaces as well as automated state scoring
systems through AI are to be excluded under European law.”771 (p.18) “We
reject lethal autonomous weapon systems that are completely removed from
human control. We actively promote their international outlawing. We want
the peaceful use of space and cyberspace. For weapons technology
developments in biotech, hypersonics, space, cyber and AI, we will take
early arms control initiatives. We want to contribute to strengthening norms
for responsible state behavior in cyberspace.”
Evaluation
Germany has undertaken a broad AI program, that pushes ethical
considerations into the national discourse. Germany has recommended
regulation of AI technologies. Germany has led efforts within the European
Union to establish comprehensive regulation for AI. Further, Germany has
promoted ethical use of AI across all sectors. While there has been no
Konrad Abenauer Stiftung Europe, A New Awakening for Europe. A New Dynamic for
Germany. A New Solidarity for Our Country: Coalition Agreement between CDU, CSU,
and SPD (2018),https://www.kas.de/c/document_library/get_file?uuid=bd41f012-1a719129-8170-8189a1d06757&groupId=284153)
770
Permanent Representation of the Federal Republic of Germany to the Conference on
Disarmament in Geneva & Représentation Permanente de la France auprès de la
Conférence du Désarmement, Meeting of the Group of Governmental Experts on Lethal
Autonomous Weapons Systems, Statement by France and Germany (Apr. 2018),
http://perma.cc/2FQB-W8FX); US Library of Congress, Regulation of Artificial
Intelligence in Selected Jurisdictions (Jan. 2019), https://www.loc.gov/law/help/artificialintelligence/regulation-artificial-intelligence.pdf
771
Koalitionsvertrag 2021, Mehr Fortschritt wagen 18, 145,
https://www.bundesregierung.de/resource/blob/974430/1990812/04221173eef9a6720059
cc353d759a2b/2021-12-10-koav2021-data.pdf
769
191
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
express support for the Universal Guidelines for AI, Germany’s policies
reflect elements found in the UGAI.
192
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Hong Kong
National AI Strategy
Hong Kong is making significant advances in AI development and
policy implementation through the issuance of guidelines, policies and AI
technology application in different sectors. Although this special
administrative region largely controlled by China does not have a national
strategy for the regulation of AI, there exists a guideline on AI to guide
organizations in adopting accountable and ethical processes. This guideline
issued by the Office of the Privacy Commissioner for Personal Data (PCPD)
contains statements on the ethical development and use of AI which is
aimed at guiding the adherence to personal data and privacy requirements
by organizations within the region.772 The PCPD guidelines are made up of
seven principles that set out to get ethical AI policy implementation. These
principles consist of themes involving accountability, human oversight,
transparency and interpretability, data privacy, fairness, beneficial AI,
reliability, robustness, and security. The PCPD guidelines also provide
practical steps to help organizations in managing their AI systems, covered
under four major areas namely:
● Establishing AI strategy and governance;
● Conducting risk assessment and human oversight;
● Executing development of AI models and managing overall AI
systems; and
● Fostering communication and engagement with stakeholders.
Additionally, in November 2020, the PCPD sponsored a resolution and
played a key role in encouraging greater accountability in the development
and use of AI to the Global Privacy Assembly (GPA). This sponsored
resolution was in response to the GPA’s adopted Declaration on Ethics and
Data Protection in Artificial Intelligence two years earlier.773 The resolution
by the PCPD called for greater accountability as it relates to the measures
below774:
772
Office of the Privacy Commissioner for Personal Data, Hong Kong. PCPD Publishes
“Guidance on Ethical Development and Use of AI” and Inspection Report on Customers’
Personal Data Systems of Two Public Utility Companies (August 2021)
https://www.pcpd.org.hk/english/news_events/media_statements/press_20210818.html
773
Bryan Cave Leighton Paisner. Hong Kong issues guidance on the use of AI
(November 2021). https://www.bclplaw.com/en-US/insights/hong-kong-issues-guidanceon-the-use-of-ai.html
774
Office of the Privacy Commissioner for Personal Data, Hong Kong. Global Privacy
Assembly Adopted a Resolution to Encourage Accountability in the Development and
Use of AI" -- Privacy Commissioner's article contribution at Hong Kong Lawyer
193
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
● assessing the potential impact to human rights (including privacy
rights) before the development and/or use of AI;
● testing the robustness, reliability, accuracy and data security of AI
systems before putting them into use; and
● disclosing the results of the privacy and human rights impact
assessment of AI, and the use of AI, the data being used and the
logic involved to enhance transparency.
In 2019, the Hong Kong Monetary Authority (HKMA) published a
12 principle guideline detailing adherence items for banks that engage in
designing and implementing AI and big data analytics applications.775 The
issuance of these principles was aimed at ensuring some form of safeguards
for banks as they deal with the increased need for AI technology adoption
in Hong Kong’s vibrant financial sector.
Hong Kong’s technology development sector, specifically AI, is
seeing increased integration with mainland China. In Carrie Lam’s, the
Chief Executive of Hong Kong, 2021 policy address she states “...the
developments of Hong Kong and our country are closely related. Only by
leveraging the Central Government’s policies in support of Hong Kong can
we give full play to our unique strengths, which will, in turn, bring
continuous impetus to our economy.776” This, coupled with the passage of
the Hong Kong National Security Law in 2020 by China’s top legislature,777
cast some doubts on the future of the “One Country, Two Systems” model
for governance of Hong Kong778.
Fundamental Rights and OECD AI Principles
Hong Kong is not one of the adopters of the Universal Declaration
of Human Rights. However, there has been the International Covenant on
Economic, Social, and Cultural Rights (ICESCR) and International
(November 2020).
https://www.pcpd.org.hk/english/news_events/newspaper/newspaper_202011.html
775
DLA Piper. Hong Kong banks must follow new AI framework (November 2019).
https://www.lexology.com/library/detail.aspx?g=aff8347f-447c-4155-801b8174a5d4668e
776
The Chief Executive’s 2021 Address.
https://www.policyaddress.gov.hk/2021/eng/p38.html
777
Hong Kong Free Press. “In full: Official English translation of the Hong Kong
national security law” (July 1 2020). https://hongkongfp.com/2020/07/01/in-full-englishtranslation-of-the-hong-kong-national-security-law/
778
Overholt, William. Hong Kong: The Rise and Fall of “One Country, Two Systems”
(December 2019).
https://ash.harvard.edu/files/ash/files/overholt_hong_kong_paper_final.pdf
194
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Covenant on Civil and Political Rights (ICCPR) in effect within the region.
These human rights conventions are also complemented by other articles
that protect the rights of citizens. One of these key articles is Article 30 that
protects citizens’ right to privacy covered by the personal data (privacy)
ordinance that clearly points out principles to be adhered to in the use of
personal data.779
The Freedom House 2021 Report, ranks Hong Kong as partly free
with a total score of 52/100.780 Political rights and civil liberties are scored
very low on the report as there is still a large prevalence of pro-Beijing
interests in the country’s political system and the freedom and autonomy of
citizens have to a large extent been controlled by political interventions
from mainland China.
Hong Kong is not a member of the OECD and has not adopted the
OECD AI Principles.781
Public Participation
Hong Kong does not have a structured process for public
participation in the development of AI policy, although, some AI and digital
technology projects have sought the engagement of citizens as part of their
roll-out plans, for instance, the Hong Kong government embarked on a twomonth public engagement drive to gather and understand the views of
citizens on the digital identity project they embarked on782.
Data Protection and Algorithmic Transparency
Hong Kong passed the Personal Data (Privacy) Ordinance (PDPO)
in 1995, which instilled a principles-based approach to data privacy and
established the Office of the Privacy Commissioner for Personal Data
(PCPD) as an independent data privacy regulator. The PCPD has been an
active participant in international discussions on data protection,
algorithmic transparency, and many other key issues in the use of AI,
especially in the General Privacy Assembly. The PDPO saw amendments
779
Hong Kong e-legislation. Cap. 486 Personal Data (Privacy) Ordinance
https://www.elegislation.gov.hk/hk/cap486
780
Freedom House. Freedom in the World (2021)
https://freedomhouse.org/country/hong-kong/freedom-world/2021
781
OECD Legal Instruments. Recommendation of the Council of Artificial Intelligence
(2021) https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449
782
Medha Basu. Exclusive: Hong Kong’s Vision for Artificial Intelligence
https://govinsider.asia/smart-gov/exclusive-hong-kongs-vision-for-artificial-intelligence/
195
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
in 2012783 and 2021784 to address direct marketing and criminalize doxxing,
respectively. Allowing the PCPD to conduct investigations without a
warrant, press charges independently, force content to be taken down, and
charge non-compliant internet platforms, there are concerns that the
doxxing amendment will be used to restrict dissenting opinions785. The
PDPO applies to both private and public data usage; however, it allows for
specific exemptions for criminal investigations, the performance of judicial
functions, security and defense, and emergency situations.786 In the context
of the recent Hong Kong National Security Law and associated protests, the
broader implications of these exemptions on human rights are less clear. For
example, police can request content be taken down or have online platforms
provide information about users, although it’s unclear if this information
can or will be shared with mainland China.787
The AI Guidance presented by the PCPD makes several
recommendations to increase transparency around the use of AI, including
“putting in controls that allow human oversight and intervention of the
operations of the relevant AI system.”788 Similarly, the guidelines for banks
using AI from the Hong Kong Monetary Authority push banks to hold
leadership accountable for AI decision-making, to ensure results from AI
systems are explainable and auditable, and to provide transparency to
783
Office of the Privacy Commissioner for Personal Data, Hong Kong. “Amendments
2012”.
https://www.pcpd.org.hk/english/data_privacy_law/amendments_2012/amendment_2012
.html
784
Office of the Privacy Commissioner for Personal Data, Hong Kong, “The Personal
Data (Privacy) (Amendment) Ordinance 2021 Takes Effect Today to Criminalise
Doxxing Acts” (Oct 8 2021),
https://www.pcpd.org.hk/english/news_events/media_statements/press_20211008.html
785
Law Society of Hong Kong, “Proposed Doxxing Offence - Personal Data (Privacy)
(Amendment) Bill 2021” (Aug 18, 2021), https://www.hklawsoc.org.hk//media/HKLS/pub_e/news/submissions/20210818.pdf
786
Hong Kong e-legislation, Cap. 486 Personal Data (Privacy) Ordinance, Part 8,
Exemptions, https://www.elegislation.gov.hk/hk/cap486
787
The Government of the Hong Kong Special Administrative Region, “Implementation
Rules for Article 43 of the Law of the People's Republic of China on Safeguarding
National Security in the Hong Kong Special Administrative Region gazetted” (Jul 6
2020), https://www.info.gov.hk/gia/general/202007/06/P2020070600784.htm
788
Office of the Privacy Commissioner for Personal Data, Hong Kong, Guidance on the
Ethical Development and Use of Artificial Intelligence (Aug 2021),
https://www.pcpd.org.hk/english/resources_centre/publications/files/guidance_ethical_e.
pdf
196
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
consumers on the use of AI789. These recommendations align with
established principles such as the OECD AI Principles, and similar
recommendations in the proposed EU AI Act; however, this is just a
guidance to businesses, and non-binding.
Hong Kong was a signatory to the 2018 GPA Declaration on Ethics
and Data Protection in Artificial Intelligence and a primary sponsor of the
2020 GPA Resolution on Accountability in the Development and Use of
Artificial Intelligence.790
AI and Surveillance
Hong Kong has long-standing concerns about AI surveillance,
especially as it relates to its relations with mainland China and the
preservation of democracy in Hong Kong. Hong Kong Chief Executive
Carrie Lam invoked the Emergency Regulations Ordinance in order to ban
the use of face masks during anti-government protests in 2019.791 This
ordinance gives the chief executive the power to “make any regulations
whatsoever which he may consider desirable in the public interest.”792 In
December 2020, Hong Kong’s Court of Final Appeal largely upheld the
application of the Emergency Regulations Ordinance for the facemask
ban793. The use of face masks in the context of the 2019 protests was a
deliberate effort by protestors to shield their identity from the government
and the subsequent mask ban casts doubt over the future of the right to freely
protest in Hong Kong.
Hong Kong has also been a strong proponent of the use of
technology in public spaces, laying out their future plans through the Hong
789
Hong Kong Monetary Authority, “High Level Principles on Artificial Intelligence” (1
Nov 2019), https://www.hkma.gov.hk/media/eng/doc/key-information/guidelines-andcircular/2019/20191101e1.pdf
790
International Conference on Data Protection and Privacy Commissioners, Declaration
on Ethics and Data Protection in Artificial Intelligence (Oct. 23, 2018),
https://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf; Global Privacy Assembly, Resolution on
Accountability in the Development and Use of Artificial Intelligence (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/FINAL-GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN-1.pdf
791
BBC News, “Hong Kong: Face Mask Ban Prompts Thousands to Protest” (Oct 4
2019), https://www.bbc.com/news/world-asia-china-49939173
792
Hong Kong E-Legislation, Cap. 241 Emergency Relations Ordinance,
https://www.elegislation.gov.hk/hk/cap241
793
Washington Post, “Hong Kong’s highest court upholds ban on masks at protests”
(Dec 21, 2020), https://www.washingtonpost.com/world/hongkong-mask-banruling/2020/12/20/f2722af0-4340-11eb-a277-49a6d1f9dff1_story.html
197
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Kong Smart Cities Blueprint 2.0, published in December 2020.794 These
plans include the use of the “StayHomeSafe” mobile app and companion
tracking bracelet, who’s use was required for all new entrants into Hong
Kong.795 Additionally as an outcome of the Smart City Blueprint, Hong
Kong has pushed to adopt concrete policy standards for the use of
autonomous vehicles (AVs).796 The Legislative Council Panel on Transport
released a proposal for legislation on this issue, with the goal of increased
testing of AVs and a clear process for licensing new test cars.797 The
blueprint and associated initiatives also outline many plans such as the
increased rollout of public wifi and contact tracing for COVID, as well as
funding for robotic patrols of airport terminals and a new LawTech Fund.
Despite the clear potential upside, there is no mention of potential concerns,
such as data privacy, algorithmic bias, or the potential violation of human
rights in the documents rolling out these new initiatives. Furthermore, there
was no significant public engagement, despite concerns about negative
impacts798. The rollout of new technology in public places, the Hong Kong
National Security Law and the development and use of facial recognition
technology by local companies799 all threaten the increased use of AI for
surveillance purposes in Hong Kong.
Evaluation
Hong Kong is a relatively new player in AI policy adoption and
implementation, although it could be safe to link this special administrative
794
Hong Kong Innovation and Technology Bureau, Hong Kong Smart Cities Blueprint
2.0 (Dec 2020),
https://www.smartcity.gov.hk/modules/custom/custom_global_js_css/assets/files/HKSma
rtCityBlueprint(ENG)v2.pdf
795
Government of Hong Kong Special Administrative Region, “StayHomeSafe User
Guide”, https://www.coronavirus.gov.hk/eng/stay-home-safe.html
796
Legislative Council Secretariat, Information Services Division, “Policy on testing and
deployment of autonomous vehicles” (Jan 12 2021), https://www.legco.gov.hk/researchpublications/english/essentials-2021ise13-policy-on-testing-and-deployment-ofautonomous-vehicles.htm
797
Legislative Council Panel on Transport, “Proposed Regulatory Framework for
Autonomous Vehicles” (May 21 2021), https://www.legco.gov.hk/yr2021/english/panels/tp/papers/tp20210521cb4-987-3-e.pdf
798
Lai, Neville, Chan, Justin, “People have to be at the heart of Hong Kong’s smart city
plans” (27 Nov 2021), https://www.scmp.com/comment/opinion/hongkong/article/3157263/people-have-be-heart-hong-kongs-smart-city-plans
799
Markay, Lachlan, “Scoop: Chinese tech firm sidesteps sanctions” (29 Sept 2021),
https://www.axios.com/chinese-tech-firm-sidesteps-sanctions-de43feaf-7df5-46ad-85bd8a37ab468e2e.html
198
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
unit with some of China’s AI efforts and achievements due to the
administrative relationship between both. Hong Kong has not been a
signatory to the OECD AI Principles or the Universal Declaration for
Human Rights; however, it has seen some effort to propose and implement
similar principles and guidelines. The leading role played by Hong Kong
through the PCPD-sponsored resolution on greater accountability in the
development and use of AI to the Global Privacy Assembly (GPA) shows
that the region has the potential to play a prominent role in key global AI
policy development and implementation. However, despite being one of the
first places to have an independent commissioner for data privacy, Hong
Kong has not been quite as proactive in the adoption of policy with regards
to the safe use of artificial intelligence. Irrespective of its complicated
relationship with mainland China especially as it relates to surveillance and
data protection issues, there is some effort by the government in regulating
and promoting ethical AI use within the country. It’s unclear, however, how
much of this extends to new government initiatives, such as the Smart Cities
Blueprint, or issues of national security. More effort is needed by the
government in the adoption of a comprehensive national AI strategy that
promotes democratic values and human rights, as well as alignment with
international commitments to AI principles.
199
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
India
National AI Strategies
Recognizing the potential of AI to transform and advance its
economy, the government of India has initiated and implemented multiple
strategies to address research, development, innovation, governance,
standards setting, and accountability of AI in India. The Task Force on
Artificial Intelligence for India’s Economic Transformation,800 which has
produced a benchmarking report (2018),801 has played a role in setting forth
India's vision regarding AI.
In 2015, India's Ministry of Commerce and Industry established
NITI Aayog, which a policy think tank of the Government of India
providing directional and policy inputs.802 The NITI Aayog replaced the
erstwhile Planning Commission, formed shortly after India’s independence
which analyzed several factors and developed goals and strategies for India
through 5 year plans. In June 2018, NITI Aayog released a discussion
paper803 the National Strategy on Artificial Intelligence where the role of
the Government has been clearly delineated to develop the research
ecosystem, promote adoption and address development of skills while
undertaking exploratory proof-of-concept AI projects in various areas,
crafting a national strategy for building a vibrant AI ecosystem in India as
well as collaborating with various experts and stakeholders. 804
NITI focused on five strategic focus areas for AI development:
healthcare, agriculture, education, smart cities and transportation.805 The
commission also identified five barriers that need to be addressed in order
to realize the full potential of AI:
1) Lack of broad-based expertise in research and application of
AI;
2) Absence of enabling data ecosystems – access to intelligent
data;
3) High resource cost and low awareness for adoption of AI;
Artificial Intelligence Task Force, Ministry of Commerce and Industry, Government
of India, https://www.aitf.org.in.
801
India, Department of Promotion of Industry and Internal Trade, Report of Task Force
on Artificial Intelligence (Mar. 2018), https://dipp.gov.in/whats-new/report-task-forceartificial-intelligence
802
NITI Aayog, (Jun. 2019), http://164.100.94.191/niti/content/overview
803
https://niti.gov.in/sites/default/files/2019-01/NationalStrategy-for-AI-DiscussionPaper.pdf
804
https://niti.gov.in/national-strategy-artificial-intelligence
805
OECD, AI in Society (2020).
800
200
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
4) Privacy and security, including a lack of formal regulations
around anonymization of data; and
5) Absence of collaborative approach to adoption and application
of AI.
The discussion paper, while highlighting the ethical factors of AI
such as fairness, transparency, privacy, and security, recommended creation
of a multi-stakeholder marketplace, facilitating creation of a large
foundational annotated data set, setting up partnerships and collaborations,
spreading awareness on the advantages of AI offers and supporting startups.In in the interim budget of 2019, the Ministry of Finance cleared
funding of approximately USD 950 million towards the NITI Aayog’s
proposal for formation of a task force to identify projects and initiatives in
which to implement Ai technology.806
Finally, the 2018 strategy discusses important issues in ethics and
AI—including fairness and bias, transparency and explainability, privacy,
and security—and advances visions for responsible AI development in its
government.
The #AIForAll strategy proposes a two-tiered framework to AI
research and development: the creation of Centres of Research Excellence
in AI (COREs), which will be academic research hubs; and the creation of
International Centres for Transformational Artificial Intelligence, which
will be industry-led.
In November 2020, NITI Aayog published an additional draft
outlining its AI Strategy, Enforcement Mechanisms for Responsible AI for
All.807 In this draft, which allowed for public participation and comments,
NITI Aayog proposed an oversight body and articulated its role and
proposed duties. These include:
• Manage and update Principles for Responsible AI in India,
• Research technical, legal, policy, and societal issues of AI,
• Provide clarity on responsible behavior through design
structures, standards, guidelines,
• Enable access to Responsible AI tools and techniques,
• Education and awareness on Responsible AI,
• Coordinate with various sectoral AI regulators, identify gaps,
and harmonize policies across sectors,
Finance Ministry Clears Niti Aayog's Artificial Intelligence Proposal, Sep. 17, 2019,
https://www.business-standard.com/article/economy-policy/finance-ministry-clears-nitiaayog-s-artificial-intelligence-proposal-119090901345_1.html
807
NITI Aayog, Working Document: Enforcement Mechanisms for Responsible #AIforAll
(Nov. 2020), https://niti.gov.in/sites/default/files/2020-11/Towards-Responsible-AIEnforcement-of-Principles.pdf
806
201
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Represent India and other emerging economies in International
AI dialogue on Responsible AI
The 2020 draft is an important addition that adds oversight and
accountability to the initial 2018 strategy, which made key
recommendations to promote research, education, and protection of civil
liberties in AI development, including the following:
1) Create two-tiered research institutes to nurture both academic
and industry research;
2) Establish learning platforms for the workforce skill/reskill
development;
3) Create targeted data sets and incubation hubs for start-ups to
facilitate cooperation; and
4) Establish a regulatory framework for data protection and cyber
security.
In 2021, NITI Aayog published two documents in the context of the
‘AI for All’ strategy. The first one proposed a series of principles for the
responsible management of AI systems that may be leveraged by relevant
stakeholders in India. These principles are: 1) Principle of Safety and
Reliability; 2) Principle of Equality; 3) Principle of Inclusivity and Nondiscrimination; 4) Principle of Privacy and security; 5) Principle of
Transparency; 6) Principle of Accountability; and 7) Principle of protection
and reinforcement of positive human values. 808 The second document
identified the various mechanisms needed for operationalizing these seven
principles by detailing a series of actions for the government, the private
sector and research institutions that must be adopted to drive responsible
AI.809
In September 2021, NITI Aayog launched a ‘New Experience
Studio’ in collaboration with Amazon Web Services and Intel, which will
help showcase the potential of technologies such as AI, machine learning,
Internet of Things (IoT), augmented reality and virtual reality, blockchain,
and robotics to accelerate their application in public sector use cases. 810
•
NITI Aayog, Responsible AI #AIForAll Part 1 – Principles for Responsible AI (Feb.
2021) https://www.niti.gov.in/sites/default/files/2021-02/Responsible-AI-22022021.pdf
809
NITI Aayog, Responsible AI #AIForAll Part 2 - Operationalizing Principles for
Responsible AI
https://www.niti.gov.in/sites/default/files/2021-08/Part2-Responsible-AI-12082021.pdf
810
India News Network, NITI Aayog, Amazon and Intel come together to accelerate digital
innovation
in
India
(Sept,
30,
2021)
https://www.indiavsdisinformation.com/20210930/niti-aayog-amazon-and-intel-cometogether-to-accelerate-digital-innovation-in-india
808
202
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Quad Group
In 2020, at a strategic meeting the foreign ministers of India, United
States, Australia, and Japan discussed a collective vision for national
security (Quadrilateral Security Dialogue), and India also expressed support
of AI policies that favor democratic nations.811 In an address at a ministerial
meeting of the Quad Group in Tokyo, external affairs Minister also spoke
about upholding “rules-based international order.” In addition to this, the
NITI Aayog has specifically recommended that India seek out ways to
harmonize its approach to AI with other emerging economies and OECD
countries. In September 2021, the Quad Group released principles on
Technology Design, Development, Governance and Use affirmed that the
ways in which technology is designed, developed, governed, and used
should be shaped by shared democratic values and respect for universal
human rights. It referred to the following principles: 1) Support of universal
values, such as freedom of expression, privacy, autonomy, agency, and
dignity of individuals; 2) Building trust, integrity and resilience; and 3)
Fostering healthy competition and international collaboration to advance
the frontier of science and technology.812
In March 2021, the U.S. - India Artificial Intelligence (USIAI)
Initiative was launched to serve as a platform to discuss opportunities for
bilateral AI R&D collaboration, share ideas for developing an AI
workforce, and recommend modes and mechanisms for catalyzing the
partnerships.813
AI Policy Development and Oversight
As discussed previously, India's Ministry of Commerce and Industry
AI commission (NITI Aayog), is charged with developing a National
Program on AI to support innovative AI projects, craft a national strategy
for building an AI ecosystem in India, and facilitate collaboration with
experts and stakeholders in key sectors. The NITI Aayog published a draft
811
The Indian Express, Jaishanker at Quad Meet: India committed to respecting
territoria integrity (Oct. 6, 2020), https://indianexpress.com/article/india/quadjaishankar-india-us-china-6705339/
812
The White House, Quad Principles on Technology Design, Development,
Governance, and Use (Sept. 24, 2021) https://www.whitehouse.gov/briefingroom/statements-releases/2021/09/24/quad-principles-on-technology-designdevelopment-governance-and-use/
813
U.S. - India Artificial Intelligence (USIAI) Initiative https://usiai.iusstf.org/
203
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
report that sets out the goals, functions, and operations of India's AI
Oversight Body.814
In addition to the NITI Aayog, the Ministry of Electronics and
Information Technology (MeitY) is also pursuing work on AI policy. The
MeitY four AI Committees charged with promoting AI initiatives and
developing policy frameworks to address 1) platforms and data on AI; 2)
leveraging AI for identifying national missions in key sectors; 4) mapping
technological capabilities, key policy enablers required across sectors,
skilling, reskill; and 4) cybersecurity, safety, legal and ethical issues.815
Recommendations have emerged from each of these committees,
including:
• Development of an Open National AI Resource Platform (NAIRP)
to become the central hub for knowledge integration and
dissemination in AI and ML;
• Stakeholders need to deliberate on whether AI systems should be
recognized as a legal person in the event of a civil liability claim;
• Sharing of best practices by the government around security,
privacy, and other issues;
• Constitute a stakeholder committee to review existing laws to
understand needed modifications for AI applications;
• AI framework should provide broad principles, and organizations
should design their internal compliance programs to maximize
flexibility with changing technologies;
• Standards should be set to address the AI development cycle. The
Bureau of Indian Standards (BIS) has established a new committee
for standardization in AI;
• Develop rigorous government safety parameters and thresholds so
that AI applications are designed to minimize harm to people and
property.
India’s AI Stack and Aadhaar
In 2009, India created UIDAI (Unique Identity Authority of India)
and embarked on the creation of an ambitious digital biometric identity
ecosystem, the Aadhaar identity system. This ecosystem utilizes AI and
machine learning techniques throughout. To facilitate a fully digital,
NITI Aayog, Working Document: Enforcement Mechanisms for Responsible #AIforAll
(Nov. 2020), https://niti.gov.in/sites/default/files/2020-11/Towards-Responsible-AIEnforcement-of-Principles.pdf
815
Government of India, Ministry of Electronics and Information Technology, Artificial
Intelligence Committee Reports, https://www.meity.gov.in/artificial-intelligencecommittees-reports
814
204
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
cashless society and economy, a large number of open application
programming interfaces or APIs were now associated with the Aadhaar
ecosystem. When these APIs are linked to services or data, including those
utilizing AI, it is called "The India Stack."816 The India Stack is the largest
system of its kind in the world, and by extension, among the largest
AI/Machine Learning based systems in the world.
The Aadhaar system, originally a voluntary pilot program, became
mandatory overtime and was tied to many services, which eventually
created substantive human rights and privacy issues.817 However, a series of
landmark rulings from the Supreme Court of India, culminating in the
landmark Aadhaar Privacy Decision of 2018,818 attempted to curtail the
negative uses of the Aadhaar system but a significant national mandate for
technological, procedural, and policy improvements remained.819 Because
of this, throughout India there is a strong impetus toward implementing
mitigations for privacy and autonomy concerns in the Aadhaar system, the
India Stack, and in general, AI systems. This mandate includes public
participation in, and understanding of, AI systems. Since the Aadhaar
ruling, many improvements regarding have been made regarding AI-based
identity systems and services, the "India Stack." However, the use and
implementation of Aadhaar with AI tends to fall with the grey-zone of the
Aadhaar Privacy Decision of 2018 as the said judgment prohibits the use of
Aadhaar by private entities and limits the use of Aadhaar by the
Government towards providing subsidies, benefits and services. In this
regard, the Attorney General for India later issued a guiding opinion820 in
2019, the Aadhaar Act was further amended821 make the statute comply
with the Aadhaar judgment. Another concern regarding constitutional
validity of the Aadhaar statute has also arisen as it was passed as a “money
The India Stak, https://www.indiastack.org/about/
Pam Dixon, A Failure to “Do No Harm” – India’s Aadhaar biometric ID program and
its inability to proect privacy in relation to measures in Europe and the U.S., Health
Technology (May 4, 2017), https://link.springer.com/epdf/10.1007/s12553-017-0202-6
818
Software Freedom Law Center, Full text of decision (Sept. 26, 2018),
https://sflc.in/updates-aadhaar-final-hearing/aadhaar-judgement
819
The Hindu, Reactions to the Aadhaar verdict: Original Aadhaar petitioner Justice
Puttaswamy welcomes parts of the judgment (Sept. 26, 2018),
https://www.thehindu.com/news/national/reactions-to-aadhaarverdict/article25046282.ece
820
Opinion of the Attorney General for India, Oct. 15, 2018,
https://uidai.gov.in/images/Circular%20Use%20of%20Aadhaar%20for%20opening%20bank%20accounts%20and%20withdrawa
l%20of%20money%20through%20AePS-reg.pdf
821
The Aadhaar And Other Laws (Amendment) Act, 2019,
https://www.uidai.gov.in/images/news/Amendment_Act_2019.pdf
816
817
205
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
bill.”822 However, privacy concerns remain and this area needs to be
observed in the future. In addition, the implementation of GDPR in Europe
renewed the focus on implementation of privacy and data protection
features within India as well.
In 2019, to further address concerns of standardization of AI
development, the Department of Telecommunications formed an AI
standardization committee to develop interface standards and design India’s
AI Stack, a framework designed to provide standards for all sectors
addressing: data privacy, protection, federation, and minimization; defined
data structures; interfaces and protocols; ethical standards; digital rights;
and trustworthiness.823 The committee released a report in 2020, and invited
public comments on the design of India’s AI Stack.
In addition, in January 2020 the NITI Aayog released
recommendations that an AI-explicit computer framework (AIRAWAT) be
established to serve the needs of innovation hubs, AI research, and students,
as well as a new discussion paper regarding the issue of Responsible AI.
In January 2022, the Ministry of Electronics and Information
Technology (MeitY) has proposed a new model of “Federated Digital
Identities” under which a citizen’s multiple digital IDs — from PAN
and Aadhaar to driving licence and passport numbers could be interlinked,
stored, and accessed via one unique ID envisaged under India Digital
Ecosystem Architecture 2.0.824
Public participation
The government of India has conducted several public consultations
on AI policy. Most recently, in July 2020, the AI policy commission of India
(NITI Aayog) requested public comments on its working document
“Towards Responsible #AIforAll.825 In addition, the Department of
Telecommunications invited public comments on the AI standardization
committee’s design of India’s AI Stack, a framework designed to provide
standards for all sectors addressing: data privacy, protection, federation, and
minimization; defined data structures; interfaces and protocols; ethical
standards; digital rights; and trustworthiness (AI Standardization
committee, 2020). The Ministry of Electronics and Information Technology
Decoding the Tribunal Judgment, Bar and Bench, Nov. 15, 2019,
https://www.barandbench.com/columns/column-decoding-the-tribunal-judgment
823
AI Standardisation committee, 2020.
824
InDEA 2.0, India Digital Ecosystem Architecture, January 2022,
https://www.meity.gov.in/writereaddata/files/InDEA%202_0%20Report%20Draft%20V
6%2024%20Jan%2022_Rev.pdf
825
AI Standardisation committee, 2020.
822
206
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
(MeitY) publishes reports from each of its four AI Committees, each
charged with promoting AI initiatives and developing policy frameworks.826
In November 2020, the Software Freedom Law Center India (SFLC)
wrote to the Chairperson of the Joint Parliamentary Committee regarding
the Personal Data Protection bill.827 The SFLC noted “core deficiencies in
the draft bill “including the lackof surveillance reforms, wide exemptions
and the problems with the Data Protection Authority.” The NGO asked the
Committee to invite civil society organizations that “defend the rights of
citizens in the digital space for consultation on the draft Data Protection
Bill.”
Further, the NITI Aayog “proposed setting up of an oversight body
to set up standards, guidelines and benchmarks for use of artificial
intelligence across sectors, which will be mandatory for public sector
procurement. The body is expected to have field experts from computer
science, AI, legal experts, sector specialists and representatives from civil
societies, humanities and social science.”828 The overarching body would
also be responsible for educating and creating awareness on responsible AI,
coordinate with various sectoral AI regulators as well as identify gaps and
harmonize policies across sectors. “Further, it would represent India (and
other emerging economies) in International AI dialogue on responsible AI.
OECD/G20 AI Principles
As a G20 member, India endorsed the G20 AI Principles at the 2019
G20 Leader’s Summit in Japan. According to the OECD, most but not all,
of the OECD AI principles are addressed in the national AI strategy.829
According to OECD AI Observatory, following the AIforAll initiatives,
India is now addressing accountability.830
Data Protection
The Supreme Court of India's Aadhaar privacy decision (2018)
created meaningful opportunities for public participation in AI policy, and
Ministry of Electronics and Information Technology, Artificial Intelligence
Committees Reports, https://www.meity.gov.in/artificial-intelligence-committees-reports
827
SFLC, Letter to Joint Parliamentary Committee on inviting civil societies for
consultation on draft Data Protection Bill, (Nov. 18, 2020), https://sflc.in/updatesaadhaar-final-hearing/aadhaar-judgement
828
Yogima Seth Sharma, NITI Aayog wants dedicated oversight body for use of artificial
intelligence, The Economic Times,
https://economictimes.indiatimes.com/news/economy/policy/niti-aayog-wants-dedicatedoversight-body-for-use-of-artificial-intelligence/articleshow/
829
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
830
OECD.AI, AI in India, https://oecd.ai/en/dashboards/countries/India
826
207
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
a series of assertive architectural, procedural, legislative, and other
improvements have been undertaken. The Aadhaar privacy decision, which
reaffirmed the centrality of privacy in one of the world's largest AI-based
identity systems, has provided a strong legal foundation for data protection
and respect of the individual. In December 2021, the Joint Parliamentary
Committee submitted its report to the Indian Parliament after two years of
deliberations on the draft bill with seven dissent.831 The latest version of the
bill also refers to algorithmic transparency (Clause 23), adding that
companies should be transparent about the fairness of algorithms used for
processing of personal data.832 The bill has also been criticised because it is
conflating issues and bringing in social media and non-personal data into its
ambit and at the same time granting complete exemption to the government
from the scope of the statute. In fact, Justice Srikrishna who had led the
drafting of the original bill criticised the revision by calling it
"Orwellian."833
Algorithmic Transparency
Prime Minister Modi addressed directly the issue of algorithmic
transparency in October 2020.834 Speaking the Responsible AI for Social
Empowerment (RAISE) summit, he said “It remains our collective
responsibility to ensure trust in how AI is used. Algorithm Transparency is
key to establishing this Trust. Equally important is accountability. We must
protect the world against weaponistion of AI by Non-State Actors.”
Joint Parliament Committee, Report of the Joint Committee on the Personal Data
Protection Bill, 2019 (Dec. 2021) https://www.ahlawatassociates.com/wpcontent/uploads/2021/12/17-Joint-Committee-on-the-Personal-Data-Protection-Bill2019.pdf
832
For an appraisal IAPP, A look at proposed changes to India's (Personal) Data
Protection Bill (Jan. 5, 2022) https://iapp.org/news/a/a-look-at-proposed-changes-toindias-personal-data-protection-bill/; Atlantic Council, Experts react: India’s Personal
Data Protection Bill tabled in Parliament
(Jan. 4, 2022) https://www.atlanticcouncil.org/blogs/southasiasource/experts-react-indiaspersonal-data-protection-bill-tabled-in-parliament/
833
Data protection bill is Orwellian, loaded in favour of the government: Justice BN
Srikrishna, Nov. 26, 2021, https://www.moneycontrol.com/news/business/dataprotection-bill-is-orwellian-loaded-in-favour-of-the-government-justice-bn-srikrishna7763331.html
834
PM Narendra Modi, We want India to become a global hub for Artificial Intelligence
(Oct. 5, 2020), https://www.narendramodi.in/text-of-pm-s-address-at-the-inauguration-ofresponsible-ai-for-social-empowerment-2020-summit-551754
831
208
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Human Rights
India was among the original 48 countries that voted in favor of the
Universal Declaration of Human Rights in 1948. According to Freedom
House, India receives generally high marks for political rights and civil
liberties.835 Freedom House reports “India maintains a robust electoral
democracy with a competitive multiparty system at the federal and state
levels, though politics are marred by corruption. The constitution
guarantees civil liberties including freedom of expression and freedom of
religion, but harassment of journalists and other government critics has
increased.” In 2021, India’s status declined, from 71/100 to 67/100, due to
a multiyear pattern in which the Hindu nationalist government have
presided over rising violence and discriminatory policies affecting the
Muslim population and pursued a crackdown on expressions of dissent by
the media, academics, civil society groups, and protesters.836
Evaluation
India has endorsed the G20 AI Guidelines. India has set out a
national strategy that addresses key concerns about the use of AI, has a
Constitutional guarantee for data protection, and has created meaningful
opportunities for public participation in AI policy. But there are still
significant gaps in the national AI policy as well as concerns about the
expanded use of the Aadhaar database database as well as implications of
strategically linking multiple disparate databases.
Freedom House, Freedom in the World 2020 – India (2020),
https://freedomhouse.org/country/india/freedom-world/2020
836
Freedom House, Freedom in the World 2021 – India (2021),
https://freedomhouse.org/country/india/freedom-world/2021
835
209
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Indonesia
National AI Strategy
Indonesia published the National Strategy for Artificial Intelligence
(Stranas KA) in August 2020.837 The National Strategy is aimed at
advancing the Indonesian economy through leadership in AI, through the
year 2045.838 The AI Strategy for Indonesia is intended to implement Visi
Indoesia 2024, the country’s economic, social, governance and technology
development strategy. The National Strategy for AI also follows Making
Indonesia 4.0, a government sponsored program, announced in 2018, to
promote the automation of the Indonesian society. Through investments in
AI, robotics and technology-based Indonesian firms as well as encourage
investment from leading Japanese, Chinese and Korean tech firms.839
The Indonesia National AI Strategy identified four key focus areas:
(1) Ethics and Policy, (2) Talent Development, (3) Infrastructure and Data,
and (4) Industrial Research and Innovation. In the focus area of Ethics and
Policy, the goals include implementing data sharing ethics, establishing a
Data Ethics Board, strengthening laws to crack down on the abuse of
technology and the misuse of data privacy.
Indonesia has already made progress in AI.840 A 2018 International
Data Corporation survey found that Indonesian companies had the highest
rates of AI adoption in Southeast Asia; a number of state projects employ
AI, to anticipate state fires for example; and some government agencies are
promoting AI development and technology-based tools at schools and other
learning institutions.841 However, the guidelines cite data misuse as a hurdle
and note that the country has neither the provisions to regulate AI, nor an
official agency to oversee AI development. They recommend establishing
a data ethics board that would set national standards for AI innovation.
837
KA Menuju Visi Indonesia 2045: Pusat Inovasi Kecerdasan Artifisial Indonesia,
https://ai-innovation.id
838
Made Anthony Iswara, Indonesia Sets Sights on Artificial Intelligence Strategy, The
Jakarta Post (Aug. 14, 2020), www.thejakartapost.com/news/2020/08/13/indonesia-setssights-on-artificial-intelligence-in-new-national-strategy.html
839
Ministry of Investment/BKPM, Making Indonesia 4.0: Indonesia’s Strategy to Enter
the 4th Generation of Industry Revolution, https://www2.investindonesia.go.id/en/whyinvest/indonesia-economic-update/making-indonesia-4.0-indonesias-strategy-to-enterthe-4th-generation-of-ind
840
Global Government Forum, Indonesia publishes AI strategy (Aug. 20, 2020),
https://www.globalgovernmentforum.com/indonesia-publishes-ai-strategy/
841
People Matters, The Journey of AI Adoption in ASEAN Countries, People Matters
(Oct. 23, 2018), https://www.peoplemattersglobal.com/article/technology/the-journey-ofai-adoption-in-asean-countries-19636
210
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI Initiatives
According to the OECD, Indonesia considers the availability of an
integrated trustworthy health data system a key challenge for trustworthy
AI in the health sector. 842 The Indonesian government is also facilitating the
development of public cloud services that will provide AI services for the
wider public. The services will also provide shared infrastructures and
platforms through which digital companies can distribute metadata, data
examples, computing and learning services that are free to use by AI
developers. Indonesia is also fostering a quadruple helix collaboration in AI
research and innovation initiatives.
According to the United Nations E-Government Development Index
(EGDI), the country is lagging in the implementation of digital services.843
The country’s president, Joko Widodo made the promise to create a
“citizen-centric digitised service government (Pemerintahan Digital
Melayani) in the next five years.” After winning his second term in April
2019, President Widodo announced that government agencies have been
ordered to replace top civil servants with AI during 2020. This would
consolidate the current top four tiers into two tiers.844 Bureaucratic reform
was also revisited in the National AI strategy, in which it is one of the five
priority areas.845
Another priority area is smart cities and mobility. There are
currently 98 smart cities and 416 smart districts planned under Indonesia’s
100 Smart Cities Plan. In 2019, President Widodo announced a new capital
on the island of Borneo, to replace Jakarta. It is planned to be a smart city
that will “rely heavily on sustainable smart city systems, cleantech and
infrastructure run by emerging technologies such as 5G, AI and IoT
(Internet of Things).”846
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
843
United Nations: Department of Economic and Social Affairs, 2020 United Nations EGovernment Survey (July 2020),
https://www.un.org/development/desa/publications/publication/2020-united-nations-egovernment-survey
844
Reuters, Indonesia aims to replace some top civil service jobs with AI in 2020 (Nov.
28, 2019), https://uk.reuters.com/article/us-indonesia-economy/indonesia-aims-toreplace-some-top-civil-service-jobs-with-ai-in-2020-idUKKBN1Y20AE
845
Kecerdasan Artifisial Indonesia, AI towards Indonesia Vision 2045, https://aiinnovation.id/stranas-ka
846
Forbes, As Jakarta sinks a new futuristic capital city will be built on Borneo, (Jan. 20,
2020), https://www.forbes.com/sites/jimdobson/2020/01/20/as-jakarta-sinks-a-newfuturistic-capital-city-will-be-built-on-borneo/
842
211
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Jakarta Smart City Initiative
In a bid to solve Jakarta’s traffic gridlocks, flooding, and waste
management, the city has turned to AI. The Indonesian government
launched the Jakarta Smart City (JSC) initiative.847 Built on six pillars, the
program uses AI to tackle the city’s governance, people, living, mobility,
economy, and environmental issues. The Smart City initiative encourages
public comment to promote transparency of the local government’s work
and better public services. At the same time, data misuse remains a hurdle
given that country has neither the provisions to regulate AI, nor an official
agency to oversee AI development.
AI Summit 2020
In November 2020, the Indonesian government held the Artificial
Intelligence Summit 2020.848 Speakers from several different countries
representing the industry, science, academia and government took part in
discussions on AI. One of the national keynote speakers spoke on “Ethics
of using health data for training data on the use of artificial intelligence.”
Another spoke on “The trustworthy, policy and talent development for
Indonesia artificial intelligence technology.”
Pancasila Values
The National Strategy states that Indonesian AI policy should be
based on Pancasila values. Pancasila is the philosophical theory that is the
foundation of Indonesian government and policy. It is comprised of five
principles: (1) Belief in The One True God, (2) A fair-minded and civilized
humanity, (3) Unity of Indonesia, (4) Democracy (from the people) led by
Wisdom of consultation (of the) representatives (of the people), and (5)
Social justice for every person in Indonesia.849 The AI Strategy sets out the
importance of establishing public trust through transparency, social and
ecological welfare, robustness and technical safety, diversity, justice and
non-discrimination, amongst others. The Strategy emphasize the
importance of AI being reliable, safe, open and accountable. Synergy
between stakeholders is also mentioned to ensure that policy is relevant and
helpful.
Techwire Asia, AI to be a US$366b industry in Indonesia by 2030 (Oct. 20, 2020),
https://techwireasia.com/2020/10/ai-to-be-a-us366b-industry-in-indonesia-by-2030/
848
Kecerdasan Artifisial Indonesia, Speakers, https://ai-innovation.id/jadwal-ais2020
849
Wikipedia, Pancasila (politics), https://en.wikipedia.org/wiki/Pancasila_(politics)
847
212
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI Oversight
According to reports, there is currently no national data protection
authority. However, certain sectors have their own authorities to ensure
compliance with the regulatory regime. For example, the Indonesian
Financial Services Authority (FSA) has the authority to act as the regulator
of data privacy in the capital markets sector and with regard to banks'
customer data privacy issues.850 At the moment, the Minister of
Communication and Informatics (MoCI) is responsible for overseeing
compliance with the data protection regime.851
The National Human Rights Commission of Indonesia, Komnas
HAM, is an independent institution that carries out studies, research,
counseling, monitoring and meditation of human rights.852 Komnas HAM
was established in 1993 by Presidential Decree and in 1999 the Law
Number 39 established its “existence, purpose, function, membership,
principles, completeness, duties and authority.” Komnas HAM also has the
authority to conduct investigation into human rights violations and
supervise of regional and central governmental policies. The goal of
Komnas HAM is to “improve the protection and enforcement of human
rights in order to develop the whole Indonesian human person and the
ability to participate in various fields of life.”
Public Participation
According to the Jakarta Post, AI providers and experts have lauded
the move to establish a foundation for AI development while urging the
government and other stakeholders to improve on the strategy, fix current
flaws and anticipate risks.853 University of Indonesia AI and robotics
professor Wisnu Jatmiko described AI as an “extraordinary challenge.” He
told The Jakarta Post that the country needs to nurture high-quality talent in
the field of AI and to bolster infrastructure, including fixing internet
connection issues and developing its own cloud computing system to
prevent the leak of confidential information. Big Data and AI Association
chairman Rudi Rusdiah and Institute for Policy Research and Advocacy
DLA Piper, Data Protection Laws of the World: Indonesia (Nov. 2020),
https://www.dlapiperdataprotection.com/?t=law&c=ID
851
Lexology, Q&A: the data protection legal framework in Indonesia (Aug. 2020),
https://www.lexology.com/library/detail.aspx?g=430e1444-ba8d-43d0-82dc86ed44d416bc
852
Komnas Ham, Legal Foundation,
https://www.komnasham.go.id/index.php/about/1/tentang-komnas-ham.html
853
Made Anthony Iswara, Indonesia Sets Sights on Artificial Intelligence Strategy, The
Jakarta Post (Aug. 14, 2020), www.thejakartapost.com/news/2020/08/13/indonesia-setssights-on-artificial-intelligence-in-new-national-strategy.html
850
213
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
researcher Alia Yofira Karunian said the national strategy should uphold
principals of fairness, accountability and transparency as pillars of AI
implementation. Karunian called on government to detect and iron out
biases in automated decision-making through human intervention, and to
ensure people have the right not to have AI make decisions about them. “We
must learn from the mistakes of other countries,” she said.
The Agency for the Assessment and Application of Technology,
coordinated the development of the National AI Strategy. The development
was carried out with help of a “wide variety of public and private sector
organization” who “contributed to the plan including government
ministries, universities, industry associations and national telecom
providers.”854
A website dedicated to the National AI Strategy illustrates the
strategy and provides material from the AI Summit 2020. The website also
provides an Artificial Intelligence Map that describes the research institutes,
universities, industry, and communities who develop and utilize innovative
Artificial Intelligence in Indonesia.
The Jakarta Smart City initiative also encouraged community
participation and government responsiveness through social media, public
figures and a public reporting system. Further, the “management of
community complaints was conducted with effective coordination between
the Jakarta Smart City team and various government departments.”855
Data Privacy Law
The Indonesian data protection regime comprises of several laws,
however, there is no general law on data protection.856 The primary law is
the law regarding Electronic Information and Transactions of 2008 (the EIT
Law). There is a 2016 amendment and implementing regulations in 2019.
In January 2020, Indonesia’s government submitted a bill to
parliament aimed at protecting consumer data. The bill includes a penalty
of up to seven years in jail for distribution of personal data without
consent.857 Data protection law is important, relevant in the global life as the
Carrington Malin, Indonesia National AI Strategy published this month (Aug. 16, 2020),
https://www.carringtonmalin.com/2020/08/16/indonesia-national-ai-strategy-set-in-motionthis-month/
855
1 World Connected, Jakarta Smart City (Sept. 1, 2020),
https://1worldconnected.org/project/asia_egov_jakartasmartcityindonesia/
856
DLA Piper, Data Protection Laws of the World: Indonesia (Nov. 2020),
https://www.dlapiperdataprotection.com/?t=law&c=ID
857
Jessica Damiana, Indonesia to step up data protection with new bill amid booming
digital economy, Reuters (Jan. 28, 2020), https://www.reuters.com/article/us-indonesiadata-idUSKBN1ZR1NL
854
214
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
economy has transformed lives in the digital era,” Communications
Minister Johnny G. Plate told a news conference. Indonesia's Personal Data
Protection Bill was initially planned to be issued in October 2020. Its
issuance and enactment were, however, delayed.858
The Indonesian government has not signed the Council of Europe’s
Convention for the Protection of Individuals with regard to Automatic
Processing of Personal Data.859 Indonesia did however participate as an
observer on the Council of Europe Convention 108 Consultative
Committee.
OECD/G20 AI Principles
Indonesia is a member of the G20 and endorsed the G20 AI
Principles in 2019. According to the OECD, the Indonesia National AI
Strategy (Strategi Nasional Kecerdasan Artifisial) addresses several OECD
AI principles.860
Indonesia will host the G20 for 2022, following the Italian
presidency in 2021, and preceding the Indian presidency in 2023. The 17th
G20 Heads of State and Government Summit will take place at the end of
the year 2022 in Bali. The Summit will be the pinnacle of the G20 process
and intense work carried out within the Ministerial Meetings, Working
Groups, and Engagement Groups throughout the year. Indonesia has
prioritized three priority issues for 2022: Global Health Architecture,
Sustainable Energy Transition, and Digital Transformation.861 The 17th G20
Heads of State and Government Summit will take place at the end of the
year 2022 in Bali. Through the Sherpa track, 11 working groups, 1 initiative
group, and 10 engagement groups meets to discuss and provides
recommendation on the G20 agenda and priorities.862 The Digital Economy
Task Force, established in 2016, supports the work of the Ministers with
competence on issues related to the digital economy and highlights the
Freddy Karyadi and Novario Asca Hutagalung, Personal Data Protection Bill To
Address Privacy Issues In Indonesia, Lexology (Nov. 20, 2020),
https://www.lexology.com/library/detail.aspx?g=b2417bcf-5548-4ba8-95921a0a299e7115
859
Council of Europe, Chart of signatures and ratifications of Treaty 108 (Status as of
Dec. 4, 2020), https://www.coe.int/en/web/conventions/full-list//conventions/treaty/108/signatures
860
OECD AI Policy Observatory, Indonesias National AI Strategy, (Dec. 9, 2021),
https://oecd.ai/en/dashboards/policy-initiatives/http:%2F%2Faipo.oecd.org%2F2021data-policyInitiatives-26968
861
G20 Presidency of Indonesia, https://g20.org/g20-presidency-of-indonesia/
862
G20 Sherpa Track, https://g20.org/sherpa-track/
858
215
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
central role of digital transformation in the broader context of economic and
social growth.
Human Rights
Indonesia has ratified the Universal Declaration of Human Rights.
According to Freedom House, Indonesia is “partly free,” with wellestablished safeguards for elections and political rights but lower marks for
civil liberties.863 Freedom House reports that “Indonesia has made
impressive democratic gains since the fall of an authoritarian regime in
1998, establishing significant pluralism in politics and the media and
undergoing multiple, peaceful transfers of power between parties. However,
the country continues to struggle with challenges including systemic
corruption, discrimination and violence against minority groups.”
Evaluation
Indonesia has endorsed the G20 AI Principles and is in the early
stages of AI policy development. While there is substantial AI investment
and several significant government undertakings, including the Jakarta
Smart City Initiative, the government has not yet developed the regulations
or created the agencies necessary for trustworthy AI. However, the
presidency of the G20 provides the opportunity for Indonesia to advance
work on AI policy.
Freedom House, Freedom in the World 2020 – Indonesia,
https://freedomhouse.org/country/indonesia/freedom-world/2020
863
216
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Iran
National AI Strategy
“Digital Iran,”864 a national road map for the advancement of
technology, outlines Iran’s digital transformation agenda. This framework
has three layers: Enabler, Application, and Impact. The enabler layer
consists of six pillars, regulation, security, infrastructure, identity, literacy,
and open data. The application layer includes digital society, digital
government, and digital business and the impact layer covers the social,
economic, and environmental dimensions. The vision of the framework is
to build “a healthy cooperative society coupled with the smart economy,
sustainable development and effective governance with transparency as a
result of utilizing digital technologies.” The framework is implemented
through 13 strategies, 30 policies and 42 confirmed projects.
In 2020, there was talk of launching a “National Centre for the
Development of AI Innovation” by the Information Telecommunication and
Technology (I.T.C) Minister,865 however, no further details have emerged
about this move. There is some evidence that AI-driven technologies have
been imported or are being used in Iran, both by the government and by the
private sector.866
The ministry of I.T.C867, responsible for implementation of modern
technologies in the IT sector, supports AI developments by hosting
international technology868 conferences, cultivation and creation of new
industries with the use of digital technologies, and investment in academic
research.869
864
Iran Digital Transformation Project, Digital Iran: National Roadmap Executive
Summary 2020-2025, https:///irandigitaltransformation.ir/wpcontent/uploads/2020/06/Digital-Iran-Roadmap-Executive-Summary.pdf
865
Twitter account of Ministry of I.C.T, https://digiato.com/article/2020/02/12/-ﺑﮫ-ﺟﮭﺮﻣﯽ
ﻣﺼﻦ-ھﻮش-ﻧﻮآوری-ﺗﻮﺳﻌﮫ-ﻣﺮﮐﺰ-زودی
866
Iran News Agency (Sept. 2021), https://en.irna.ir/news/84025992/Iran-Russia-tocooperate-on-artificial-intelligence-research
867
Information Telecommunication and Technology (I.C.T),
https://www.ict.gov.ir/en/topmainmenu/aboutus
868
20th International Exhibition of Telecommunications, Information Technology &
Innovative CIT Solutions, http://www.irantelecomfair.com/en/ and
https://calendar.iranfair.com/en/companies/index/252/The-21st-International-Exhibitionof-Telecommunications-Information-Technology-Innovative-CIT-Solution
869
AmirKabir University of Technology, The “Simorgh” Supercomputer, (June 2021)
https://aut.ac.ir/content/7995/The-“Simorgh”-Supercomputer-was-Launched-at-AUT
217
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Freedom of Internet and Digital Rights
In February 2020, the Supreme Council for Cyberspace (SCC),
Iran’s top internet policymaking body, initiated meetings to set five-year
targets for the expansion of the National Information Network (NIN), the
country’s localized internet architecture. The plan was approved by the SCC
in September 2020.870 The NIN established several targets to improve
internet access.871
The Telecommunication Company of Iran (TCI), which is run by
the I.C.T Ministry, controls internet traffic flowing in and out of the
country.872 The Internet dominance creates opportunity for TCI to monitor
online activities, where the majority of TCI’s shareholder is the Islamic
Revolutionary Guard Corps (IRGC), a powerful branch of the security
forces that also controls large portions of the economy.873
Data Protection
There are currently no comprehensive data protection laws in place
in Iran.874 In February 2018, the I.C.T Minister unveiled the first drafts of
five newly proposed bills for internet and cyberspace regulation in Iran. The
legislation address e-government, electronic identification, and the
responsibilities of service providers, electronic financial transactions, and
data protection, respectively.875 A draft bill on data protection and privacy
was presented to the cabinet in July 2018, and it is awaiting review from the
Islamic Parliament of Iran without any clarifications on the expected
870
Filter Watch, Policy Monitor (Sept. 2020)
https://filter.watch/en/2020/10/26/policy-monitor-september-2020/
871
Filter watch, Policy Monitor (Feb. 2020)
https://medium.com/filterwatch/filterwatch-policy-monitor-february-2020-41db0293f2e0
872
Small Media Research, Iranian Internet Infrastructure and Policy Report (2015),
https://smallmedia.org.uk/media/articles/files/IIIP_Jul15.pdf#page=9
873
Gholam Khiabany and Annabelle Sreberny, Blogistan: The Internet and Politics in
Iran, London:IB Tauris, 2010),
p.5.https://www.bloomsburycollections.com/book/blogistan-the-internet-and-politics-iniran/?clearSearch
874
Filter Watch, Data Insecurity On Iran’s Localised Internet,2020,
https://filter.watch/en/2020/06/19/data-insecurity-on-irans-localised-internet/
875
Iran I.C.T Ministry, Protection of Personal Data - Draft Bill ( Chrome translation
services),
https://www.ict.gov.ir/fa/newsagency/21691/%D9%84%D8%A7%DB%8C%D8%AD%
D9%87-%D8%B5%DB%[…]DB%8C%D8%B1%D9%88%D9%86%D9%85%D8%A7%DB%8C%DB%8C%D8%B4%D8%AF
218
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
timeframe.876 The draft law provides for the establishment of the
Supervisory Board of Personal Data which would be tasked with receiving
and processing stakeholder complaints in order to protect personal data. In
the absence of an overarching data privacy law, the legal framework for
privacy derives from a combination of other laws and regulations dealing
with data protection alongside additional issues. Such legislation includes
the Law on Publication and Access877 to Data 2009, the Electronic
Commerce Law 2004,878 and the Cybercrime Law 2009.879
Digital ID
Iran is currently working on implementing digital IDs as part of
Iran’s e-government program.880 Iran’s Executive Council of Information
Technology is building new digital platforms as part of its ongoing egovernment initiatives. In addition to two pilots with the Ministry of
Agriculture and the Ministry of Culture and Islamic Guidance, the council
also announced the addition of the Ministry of Cooperatives Labor and
Social Welfare, the Ministry of Economic Affairs and Finance, the Iranian
Health Insurance Organization, the Ministry of Welfare and Social Security,
and the Central Insurance of Iran for the next phase of the digital
government.881 Iran’s Law Enforcement Forces signed an agreement with
an Iranian digital ID and biometric firm, to acquire digital ID detection
authentication platform.882 The digital ID platform obtains 5 to 15-second
876
One Trust Data Guidance Solution, Iran
Report,https://www.dataguidance.com/jurisdiction/iran
877
Ministry of Culture and Islamic Guidance, Law on Dissemination of and Free Access
to Information,(2009), (Chrome Translation Services) https://foia.farhang.gov.ir/en/law
878
Iran Ministry of Commerce, Electronic Commerce Law (2010),
https://www.wipo.int/edocs/lexdocs/laws/en/ir/ir008en.pdf
879
International Labour Organization, Database of National Labour, Social Security and
Related Human Rights Legislation, Criminal and penal law (2009), Law No. 71063 on
Computer Crimes.(Google Translation Services)
http://ilo.org/dyn/natlex/docs/ELECTRONIC/91715/106512/F1311829502/IRN91715.pdf
880
Biometric Update, Iran unveils new e-government components as digital ID
importance grows, 2021,
https://www.biometricupdate.com/202103/iran-unveils-new-e-government-componentsas-digital-id-importance-grows
881
Biometric Update, Iran unveils new e-government components as digital ID
importance grows, 2021
https://www.biometricupdate.com/202103/iran-unveils-new-e-government-componentsas-digital-id-importance-grows
882
Biometric Updates, “UID to supply biometric digital ID app to Iran’s national police,
2021
219
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
selfie videos and then runs it through an ID liveness detection algorithm to
verify user identity against a user’s original image registered with Iran’s
Civil Registration Organization.
Facial Recognition
Iran is one of many countries worldwide actively integrating facial
recognition and biometrics into its law enforcement system for border
protection.883 In 2015, the Iranian government launched a biometric
national identity card - a card “issued to all new applicants and to anyone
renewing an expired national identity card.”884 Iran’s banking sector is
adopting biometric methods and requires customers to provide their smart
identity for many banking transactions. Iranian officials have also
announced plans that would require citizens to verify their identity using the
smart identity card to access the internet.885
In January 2020, the Iranian government ceased allowing applicants
for the card to choose ‘Other’ in the religion field on the application form,
which had previously been one of the available options.886 Instead,
applicants must now choose one of the four officially recognized religions
given on the form – Islam, Christianity, Judaism or Zoroastrianism.
OECD/G20 Principles
Iran has not yet adopted OECD AI Principles, nor did it define
ethical norms and standards for AI.887
https://www.biometricupdate.com/202102/uid-to-supply-biometric-digital-id-app-toirans-national-police and https://www.ilna.news/-ﺗﻔﺎھﻢ-اﻣﻀﺎی-1036837/158-آوری-ﻓﻦ-ﺑﺨﺶ
ﯾﻮآﯾﺪی-ﺷﺮﮐﺖ-ﭘﻠﯿﺲ-ﺑﯿﻦ-ھﻤﮑﺎری-ﻧﺎﻣﮫ
883
ICAO Regional Seminar, Iranian ePassport and Border Management Technical
Report, 2016. https://www.icao.int/Meetings/icaotrip-Iran-
2016/Documents/Presentations/D2%20S4%20EBRAHIMI.pdf
Minority and Indigenous Trends, 2020, Case Studies, Middle East and North Africa:
Iran, https://minorityrights.org/trends2020/
885
Identity Review, “Iran Begins Integrating Facial Recognition for Better Border
Protection”, (secondary resource) https://identityreview.com/iran-begins-integratingfacial-recognition-for-better-border-protection/
886
Minority and Indigenous Trends, 2020, Case Studies, Middle East and North Africa:
Iran, https://minorityrights.org/trends2020/iran/
887
State of Implementation of OECD AI Principle, June 2021, https://www.oecdilibrary.org/science-and-technology/state-of-implementation-of-the-oecd-aiprinciples_1cd40c44-en?_ga=2.209594642.1644595301.16373276841452510560.1637152644
884
220
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
UNESCO AI Ethics Recommendation
Iran is one of signatories to the UNESCO recommendations on
ethical AI which aims to deliver the advantages of technology while
reducing associated human rights risks.888
Human Rights
The Freedom House Index indicates Iran as a “Not Free” country,
with a rating of 16/100, a slight drop from 2020.889 Internet freedom
remained highly restricted and State-aligned cyber operations spread
disinformation and manipulated the online space. Iran and its judicial
system have shown little inclination to curb or confront serious rights
violations perpetrated by Iranian security agencies. The country’s security
and intelligence apparatus, in partnership with Iran’s judiciary, harshly
crack down on any form of dissent, including through excessive and lethal
force against protesters.890
Even though Iran is a signatory to the UDHR in 1948, there is scant
regard for women’s/children’s rights, gender identity, and religious
freedom. “The government also discriminates against other religious
minorities, including Sunni Muslims and restricts cultural and political
activities among the country’s Azeri, Kurdish, Arab, and Baluch ethnic
minorities.”891
Lethal Autonomous Weapons
The Iranian military is very interested in the development of AI &
Autonomous Weapon systems in the pursuit of force-multiplying
asymmetric warfare capabilities. At the Human Rights Council in May
2013, Iran expressed interest in opening multilateral talks on lethal
autonomous weapons systems. Iran however has not commented on the
concerns raised by removing human control from the use of force or
supported proposals to negotiate a new international ban treaty. Iran is not
a party to the Convention on Conventional Weapons (CCW), but it attended
CCW meetings on killer robots in 2016 and 2018-2019.892
First-ever Global Agreement on the Ethics of Artificial Intelligence, November 2021,
https://news.un.org/en/story/2021/11/1106612
889
Freedom House, Freedom in the World – 2021: Iran,
https://freedomhouse.org/country/iran/freedom-world/2021
890
Human Rights Watch, Iran Report 2020, https://www.hrw.org/worldreport/2021/country-chapters/iran
891
Human Rights Watch, Iran Report 2020, https://www.hrw.org/worldreport/2021/country-chapters/iran
892
Human Rights Watch, Stopping Killer Robots: Country Positions on Banning Fully
Autonomous Weapons and Retaining Human Control, 2020,
888
221
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Evaluation
Iran has launched the “Digital Iran” framework to oversee and
implement new technologies. Designing mechanisms for protecting vital
data is one of the priorities under this framework, but no appropriate law is
yet put in place. The absence of protections for fundamental rights as the
country seeks to expand national identification and systems for facial
recognition is concerning. Iran is implementing digital IDs and AI
surveillance technologies, but this report has not found any evidence on
implementing oversight legislation and responsible use of AI ethics
standards and principle and public participation and dialogue between
different stakeholders, like civil society, government and business on the
topic of AI, privacy and ethical framework.
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#
222
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Ireland
National AI Strategy
Ireland’s national enterprise strategy, the Economic Recovery Plan
and Enterprise 2025 Renewed, set out Ireland’s ambition to be at the frontier
of disruptive technologies, including AI. In 2019, Ireland published the
Future Jobs Ireland Plan, which included a commitment to develop a
national AI strategy for Ireland and created a Top Team on Standards for
AI to focus on increasing Ireland’s AI development and assisting AI
enterprises.893
Ireland’s national AI strategy AI - Here for Good, sets out a longterm roadmap for artificial intelligence. It was published by the Department
of Enterprise, Trade, and Employment (DETE) in 2021 after participation
from a wide range of stakeholders from industry, academia, and the public
in meetings and through an online public consultation.894
The aim of AI - Here for Good is to make Ireland “an international
leader in using AI to the benefit of our population, through a people-centred,
ethical approach to AI development, adoption and use.”895 Its objectives are
grouped into eight strands: AI and Society; A Governance Ecosystem that
Promotes Trustworthy AI; Driving Adoption of AI in Irish Enterprise; AI
Serving the Public; A Strong AI Innovation Ecosystem; AI Education,
Skills and Talent; Supportive and Secure Infrastructure for AI; and
Implementation of the Strategy.
Ireland's national AI policy states that "ensuring explainability,
accountability, and fairness" and addressing discrimination are some of the
main challenges to be addressed through AI regulation. Ireland's "Strategic
Approach to AI" emphasizes the importance of AI that is "accountable and
acceptable to society." Under its objective to create an agile AI governance
and regulatory framework, Ireland recognizes the risk of unfair
discrimination and unequal treatment arising from biased training data,
design, and use, along with challenges of "explainability, accountability,
and fairness." The Top Team on Standards for AI is responsible for
developing certification schemes and codes of conduct for AI to determine
Government of Ireland, Future Jobs Ireland 2019: Preparing Now for Tomorrow’s
Economy (March 2019), https://enterprise.gov.ie/en/Publications/Publication-files/FutureJobs-Ireland-2019.pdf
894
Department of the Taoiseach Press Release, Taoiseach and Minister Troy launch
Government Roadmap for AI in Ireland (July 8, 2021), https://www.gov.ie/en/pressrelease/f4895-taoiseach-and-minister-troy-launch-government-roadmap-for-ai-in-ireland/
895
Department of Enterprise, Trade and Employment, AI - Here for Good: National
Artificial Intelligence Strategy for Ireland (last updated September 15, 2021),
https://www.gov.ie/en/publication/91f74-national-ai-strategy/
893
223
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
or demonstrate fairness, estimate bias in training data, develop auditing
mechanisms, and support GDPR rights.
Ireland’s National AI Strategy favors the Council of Europe's Ad
Hoc Committee on AI (CAHAI) for its explicit aim to ensure protection of
democracy and rule of law and its “mix of binding and non-binding legal
instruments.” Ireland seeks to apply the Council of Europe’s Ethical Charter
on the Use of AI in Judicial Systems to promote rule of law in the use of AI
in the justice sector.
Public Participation
To receive engagement and input in the development of its National
AI Strategy, Ireland held a public consultation process from October 16,
2019, to November 7, 2019, which was open to all stakeholders and
interested parties. The consultation was meant to “better understand the
views of the public on the opportunities, enablers and challenges for AI in
Ireland and to gather views on key areas and issues that should be addressed
by the strategy.”896
Strand 1 of AI - Here for Good, “AI and Society,” asserts that Ireland
“must also prioritise measures to raise awareness about AI.” To that end,
Ireland plans to appoint an “AI ambassador to promote awareness among
the public and businesses of the potential that AI offers.” The AI
ambassador is expected to engage with the public, lead a “national
conversation around the role of AI” with an emphasis on “an ethical and
compliant approach,” and champion AI as a positive force for Ireland.
Ireland published a call for Expression of Interest for the AI Ambassador
position on October 26, 2021, and applications were due November 12,
2021.897 There is no indication at this time that the AI Ambassador has been
chosen.
Ireland plans to convene a “Youth Assembly on AI” to discuss
young people’s views of AI’s “benefits, risks and impacts on different
groups in society.” University College Cork (UCC) currently hosts “The
Elements of AI,” a massive open online course (MOOC) made available to
all EU member states. Ireland plans to use “Elements of AI,” which is freely
available, “to deliver AI education to at least 1% of the population.”
896
Department of Enterprise, Trade and Employment, National AI Strategy for Ireland:
Public Consultation Report, https://enterprise.gov.ie/en/Consultations/Consultationsfiles/AI-Strategy-Public-Consultation-Report.pdf
897
Department of Enterprise, Trade and Employment, Call for Expression of Interest:
Artificial Intelligence (AI) Ambassador (Oct. 26, 2021),
https://www.gov.ie/en/publication/07668-call-for-expression-of-interest-artificialintelligence-ai-ambassador/
224
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Trust in AI
Strands 1 and 2 of AI - Here for Good are aimed at “building public
trust in AI.” Strand 1 seeks to build public trust in AI through public
awareness programs and research and grants to develop AI applications for
societal good and sustainability, including health and climate change.
Strand 2 addresses the need for a “robust governance framework to
safeguard against [AI] risks and underpin public trust in AI.” It establishes
three pillars that Ireland will use to ensure a strong AI governance
framework: 1) “an agile and appropriate framework,” 2) “active promotion
of ethics guidance and frameworks,” and 3) “a robust system of standards
and certification.”
Ireland has played an active part in EU discussions of the AI Act
and the EU’s prior work related to AI, and Ireland’s AI strategy states that
it will continue to do so. Ireland’s National AI Strategy endorses the EU AI
Act as a “‘smart mix’ of voluntary and mandatory measures [that] will help
to protect our people, facilitate innovation in AI and respect our democratic
values,” Strand 2 emphasized the AI Act’s voluntary and self-regulatory
oversight of non-high-risk AI and its integration of impact assessments,
codes of practice, and ethical guidelines.
Human Rights
According to Freedom House, Ireland is “free,” with high scores for
political rights and civil liberties (97/100). Regarding transparency and
openness, Freedom House reports: “The public has broad access to official
information under the 2014 Freedom of Information Act, though partial
exemptions remain for the police and some other agencies. A Transparency
Code requires open records on the groups and individuals that advise public
officials on policy. The government has been criticized for failing to consult
meaningfully with civil society groups and relevant stakeholders in policy
formulation, particularly regarding the Roma, Travellers, and people living
with disabilities.”898
Ireland has endorsed the Universal Declaration of Human Rights
and ratified seven of the nine core international human rights instruments,899
Freedom House, Freedom in the World 2021-Ireland (2021),
https://freedomhouse.org/country/ireland/freedom-world/2021
899
UN Office of the High Commissioner for Human Rights, Ratification Status for
Ireland,
https://tbinternet.ohchr.org/_layouts/15/TreatyBodyExternal/Treaty.aspx?CountryID=83
&Lang=EN
898
225
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
along with the European Convention on Human Rights.900 Strand 1 of
Ireland’s national AI strategy commits to “making human rights and ethical
principles a key focus” of its AI strategy, although it largely focuses on AI
R&D that can improve access and inclusion, e.g., AI tools that help people
with impaired hearing through real-time live captioning.
G20/OECD Principles
Ireland, a member of the OECD and Council of Europe, endorsed
the OECD AI Principles in May 2019. According to its National AI
Strategy, Ireland’s AI policies are “underpinned by [Ireland’s] engagement”
with AI policy processes at the EU, UN, and OECD.
Global Partnership on AI
Ireland is not a member of the Global Partnership on AI (GPAI). As
a part of its National AI Strategy, however, Ireland stated its goal of joining
GPAI as part of its effort to develop an “agile and appropriate” AI
governance and regulatory environment.
Global Privacy Assembly
Ireland’s Data Protection Commissioner is an accredited member of
the Global Privacy Assembly and has endorsed the Global Privacy
Assembly’s 2018 Declaration on Ethics and Data Protection in AI. Ireland
has not endorsed the Global Privacy Assembly’s 2020 Resolution on
Accountability in the Development and Use of AI.
Algorithmic Transparency
Transparency is an important part of Ireland's national AI strategy.
Ireland views transparency as an essential component of accountability and
trust in AI. Strand 1 of AI - Here for Good states that “AI must be developed
and used with trust, transparency and accountability” and that Ireland is
committed to ensuring that AI systems are “trustworthy, fair and inclusive.”
The GDPR’s transparency requirements apply to Ireland as an EU
member. Ireland’s strategy incorporates the transparency requirements of
the EU AI Act. Although Ireland has not created its own AI ethical
guidelines, its National AI Strategy incorporates the seven requirements of
the EU High-Level Expert Group (EU HLEG) on AI’s Ethics Guidelines
for Trustworthy AI. UCC’s Insight Centre currently hosts an online version
of EU HLEG’s Assessment List for Trustworthy AI, as an accessible
checklist for adhering to the seven requirements of the Ethics Guidelines
Irish Human Rights and Equality Commission, Understanding Rights—Human Rights,
(accessed Nov. 26, 2021), https://www.makerightsreal.ie/understanding-rights/humanrights/
900
226
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
for Trustworthy AI. The Top Team on Standards for AI is tasked with
developing codes of conduct that address explainability and transparency.
AI Oversight
Established under the Data Protection Act of 2018, the Data
Protection Commission (DPC) is Ireland's national independent supervisory
authority responsible for upholding the data protection rights of individuals
in the EU. The DPC is charged with monitoring the application of the
GDPR, including its transparency and automated decision-making
provisions, and other Irish and EU regulatory frameworks and directives.
The DPC has faced criticism from members of the European Parliament for
failing to enforce the GDPR by choosing to bring a legal proceeding in the
Schrems II case instead of making a decision on its own, issuing only one
GDPR sanction out of thousands of complaints, and failing to reach a draft
decision on any case referred to Ireland in 2018. On March 25, 2021, the
European Parliament passed a resolution expressing its concern with the
slow pace of the Irish DPA and calling for faster investigations to show EU
citizens that “data protection is an enforceable right.”901 On May 20, 2021,
the European Parliament voted in favor of a resolution calling for the
European Commission to open an infringement procedure against the
DPC.902 The failure of the DPC to exercise enforcement powers now raises
concerns about the Commission’s ability to safeguard fundamental rights
with regard to AI.
Ireland's Department of Enterprise, Trade, and Employment
(DETE) is currently spearheading Ireland's national AI policy, which lists
strategic actions to be conducted by various Irish agencies, including the
Department of Environment, Climate and Communications, the Department
of Justice, the National Standards Authority of Ireland, and the Department
of Children, Equality, Disability, Integration, and Youth. It has also
established the Top Team on Standards for AI to develop certification
schemes and codes of conduct to address AI at different stages of
development.
Strand 8 of AI - Here for Good, “Implementation of the Strategy,”
lists several entities that compose Ireland’s “whole of Government strategy”
for AI. The Enterprise Digital Advisory Board will soon be established to
represent government departments, businesses, and AI experts, oversee the
European Parliament, Resolution on the Commission evaluation report on the
implementation of GDPR two years after its application (March 25, 2021),
https://www.europarl.europa.eu/doceo/document/TA-9-2021-0111_EN.html
902
European Parliament, Resolution on the ruling of the CJEU of ‘Schrems II’ (May 20,
2021), https://www.europarl.europa.eu/doceo/document/TA-9-2021-0256_EN.pdf
901
227
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
implementation of business elements of the National AI Strategy, and
advise the Irish Government on driving enterprise adoption of AI. The AI
Ambassador, a member of the Enterprise Digital Advisory Board, and the
GovTech Delivery Board, which is responsible for AI adoption in the public
sector, are other entities that Ireland plans to incorporate into its governance
of AI.
Lethal Autonomous Weapons
Ireland is a High Contracting Party to the Convention on Certain
Conventional Weapons (CCW) and has been an active participant in CCW
discussions related to lethal autonomous weapons systems (LAWS).
According to Human Rights Watch, Ireland has expressed interest in
multilateral talks on LAWS in the UN since at least 2013 and has
participated in every CCW meeting on LAWS between 2014 and 2019.903
In 2019, Ireland joined the other High Contracting Parties to the CCW to
adopt 11 guiding principles for addressing challenges to international
humanitarian law posed by LAWS.904 It also joined eight other CCW parties
in finding that the 11 guiding principles were a “useful and valuable starting
point” and calling for the development of a “normative and operational
framework” for ensuring human control of LAWS.905 Ireland has not called
for a prohibition on or new international treaty for the regulation of LAWS.
Public Services Card Facial Recognition Controversy
In March 2013, the Department of Employment Affairs and Social
Protection (DEASP), Ireland’s national social security office, implemented
facial image matching software with the aim of minimizing fraud and error
in the use of Public Services Cards (PSC) to verify the identity of social
security applicants. The Irish Council for Civil Liberties (ICCL), an
independent non-profit in Ireland, has criticized the PSC for collecting and
storing more data than necessary for its purpose, increasing the risk of a
security breach, potentially collecting extremely sensitive biometric data,
Human Rights Watch, Stopping Killer Robots: Country Positions on Banning Fully
Autonomous Weapons and Retaining Human Control (Aug. 10, 2020),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#_ftn141
904
Meeting of the High Contracting Parties to the Convention on Certain Conventional
Weapons, Final Report (Dec. 13, 2019), https://documents-ddsny.un.org/doc/UNDOC/GEN/G19/343/64/PDF/G1934364.pdf?OpenElement
905
Austria, Belgium, Brazil, Chile, Ireland, Germany, Luxembourg, Mexico, and New
Zealand, Joint Commentary on Guiding Principles A, B, C and D (Sept. 1, 2020),
https://documents.unoda.org/wp-content/uploads/2020/09/GGE20200901-AustriaBelgium-Brazil-Chile-Ireland-Germany-Luxembourg-Mexico-and-New-Zealand.pdf
903
228
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
forcing economically vulnerable people to exchange their private data for
access to services to which they are legally entitled, and contradicting
Ireland’s position on privacy at the EU.906 DEASP has not published
information on the accuracy of the facial image matching software used for
the PSC and has continued to use the software despite being under
investigation by the Irish Data Protection Commissioner since October
2017. Human Rights Watch has documented reports of eligible applicants
being denied social security support due to documentation requirements or
objections to the PSC for privacy reasons.907
Evaluation
Ireland has endorsed the OECD AI Principles and recently
established its national AI strategy to address issues of trust and
transparency and improve Ireland’s competitiveness in AI. As an EU and
Council of Europe member, Ireland enjoys data protection and algorithmic
transparency guarantees. Ireland has consistently advocated for
comprehensive LAWS regulations. Its national AI strategy was developed
with mechanisms for public participation and discussion and adopts EU
HLEG-AI’s Ethics Guidelines for Trustworthy AI. Ireland is fairly new to
AI policy, having only established its strategy in 2021. As a result,
implementation is in its early stages and gaps in Ireland’s AI policies
remain. Ireland’s national AI strategy focuses heavily on business adoption
of AI and the importance of trust, through safeguards and public awareness,
for ensuring that Ireland will be an AI innovation hub. The continued use of
facial recognition by DEASP for determining access to social services and
the Irish DPC’s weak GDPR enforcement record remain areas of concern.
Irish Council for Civil Liberties, The Public Services Card,
https://www.iccl.ie/2019/the-public-services-card-contd/
907
Human Rights Watch, Q&A: How the EU’s Flawed Artificial Intelligence Regulation
Endangers the Social Safety Net (Nov. 10, 2021),
https://www.hrw.org/sites/default/files/media_2021/11/202111hrw_eu_ai_regulation_qa_
0.pdf
906
229
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Israel
National AI Strategy
Although Israel is described as one of the world’s top three countries
in the field of AI research, 908 at present Israel does not have a national
strategy for AI. In January 2018 former Prime Minister Benjamin
Netanyahu stated, “Artificial intelligence is changing everything,
connectivity is important and these are changing the structure of growth.”909
In July 2018, Netanyahu established an AI Steering Committee with experts
from academia, government, business and civil society, led by Professors
(Major General retired) Izhak Ben-Israel and Eviatar Matania from the TelAviv University.910 Subcommittees were established to explore such topics
as robotic and autonomous systems, IOT and Sensors, distributed intelligent
systems, quantum computing, academic research centers, cyber-Security
and AI, and ethics and regulation. The committees completed discussions
in 2019.
In November 2019, Ben-Israel and Matania released a draft report,
during AI Week at Tel Aviv University, announcing a focus on the
digitization of government services and the agriculture sector. Ben-Israel
and Matania also stated that the Steering Committee recommended a
coordination agency for AI within the Prime Minister's Office, an Israeli AI
cloud, the classification of an Israel city as a "trial city" for smart
transportation and autonomous vehicles, and the creation of research centers
in universities.
In a public statement, other members of the Steering Committee
objected to the summary of the recommendations and the manner of the
announcement. 911 They pointed that the Steering Committee was one of 15
subcommittees set up to discuss the needs and best policy concerning the
various aspects of promoting AI, such as professional training, ethical use
of the technology, computer infrastructure, and national projects. They
recalled that the purpose of the Steering Committee was to coordinate the
various work committees' conclusions for submission to government
ASGARD, The Global Artificial Intelligence Landscape, https://asgard.vc/global-ai/
Srishti Deoras, Israeli PM Benjamin Netanyahu Believes Big Data And Artificial
Intelligence Will Reshape The World, Analytics India (Jan. 21, 2018).
https://analyticsindiamag.com/israeli-pm-benjamin-netanyahu-believes-big-dataartificial-intelligence-will-reshape-world/
910
Uri Berkovitz, Israel's national AI plan unveiled, Globes (Nov. 20, 2019),
https://en.globes.co.il/en/article-israels-national-ai-plan-unveiled-1001307979
911 Uri Berkovitz, Critics slam draft national AI plan, Globes (Nov. 24, 2019),
https://en.globes.co.il/en/article-critics-slam-draft-national-ai-plan-1001308287.
908
909
230
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
approval. They specifically objected to the emphasis on agriculture and the
creation of an organizing agency. Ben-Israel said in response, "The report
is a draft, and discussion of it has not ended. We will finish the work we
started."
In October 2020, the Institute for National Security Studies (INSS)
recommended setting up an agency like the National Cyber Directorate to
take charge of integrating artificial intelligence into the defense
establishment and maintain Israeli leadership.912 The INSS argued that
developing a national strategy for artificial intelligence, including its ethical
aspects, is critical for Israel’s future security.
National AI Program
In December 2020, Israel launched a five-year national AI program
with a budget of approximately $1.55 billion).913 The national program was
created for TELEM (The National Infrastructure Forum for Research and
Development), which is a voluntary organization that aims to promote R&D
programs and projects in scientific and technological fields through
establishment of national R&D infrastructures and inter-organizational,
inter-departmental and international collaborations. The program was
written by a committee of experts in AI, which recommended several urgent
projects, including the establishing of an HPC (High Performance
Computing) supercomputer, promoting generic R&D projects with a focus
on NLP (Neuro-linguistic programming) capabilities and the development
of training of manpower and the purchasing of advanced equipment for
academia. This is because, according to the committee, despite the bustling
startup industry in Israel, there are gaps in academia and research, in
sophisticated computer infrastructure (mainly the lack of supercomputing)
and in regulation that would enable AI applications to be used in both the
public and the private sectors.
Ethical and Legal Aspect of AI
The ethical dimensions of AI have received attention from the Israeli
Parliament and academics. In June 2018, the Knesset Science and
Technology Committee called upon the government to examine various
912
Sagi Cohen, AI is the next national security frontier, but Israel may be losing its edge,
Haaretz (Oct. 12, 2020), https://www.inss.org.il/wp-content/uploads/2020/10/AI-is-thenext-national-security-frontier-but-Israel-may-be-losing-its-edge-Liran-Antebi-Haaretzfor-site.pdf
913
Calcalist, Israel launches national AI program, but lack of budget threatens its
implementation (Dec. 22, 2020) https://www.calcalistech.com/ctech/articles/0,7340,L3883355,00.html
231
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
regulatory aspects of AI, including privacy and legal responsibility.914 A few
months later, the CEO of the Israel Innovation Authority (IIA) also urged
Israel to “close the gap with other countries who already make enormous
investments in artificial intelligence infrastructures. In order for Israel to
continue to lead in the global technological race, it is necessary to allocate
resources and a national artificial intelligence strategy shared by the
government, academia, and the industry.”915
The 2019 IIA Report pointed identified four challenges to Israel’s
continued leadership AI: (1) a shortage of human capital skilled in the field
of AI; (2) limited access to public and government databases for use by new
companies; (3) inadequate supercomputing infrastructure for the
development of advanced technologies; and (4) slow-changing regulation
and a lack of ethical guidelines.
The IIA Report also drew attention to privacy and ethics concerns
for AI: “Implementation of smart systems raises ethical concerns that call
for clear regulatory definitions. Using large databases poses privacy
challenges that mandate information classification by sensitivity levels.
Moreover, at times, it is not clear how AI systems make decisions. The
responsibility of the manufacturer or the user for the machine’s
‘independent’ activity, in the case of critical error, remains unclear (for
example, in an autonomous car accident).”916
The Ethics and Regulation subcommittee, chaired by Prof. Karine
Nahon, released its report in November 2019.917 The committee was
“commissioned to suggest guiding principles in the Israeli context that
would be taken into account as part of the national plan to turn Israel into
an AI leader.” The Committee recommended the following ethical
principles for AI:
1) Fairness
Science and Technology Committee, First discussion on the government's readiness
for the field of artificial intelligence (June 4, 2018) [GT],
https://m.knesset.gov.il/news/pressreleases/pages/press04.06.18ec.aspx
915
Israel Innovation Authority, 2018-19 Report (Jan. 14, 2019),
https://innovationisrael.org.il/en/news/israel-innovation-authority-2018-19-report
916
Israel Innovation Authority, Bolstering Artificial Intelligence: What Can Be Done for
Israel to Maintain its Leading Position in the Field of AI? (2019) (section: Changes
Needed to Privacy and Ethics Policy in AI),
https://innovationisrael.org.il/en/reportchapter/bolstering-artificial-intelligence0#footnote3_fzh0scp
917
Ethics and Regulations Team, Subcommittee of the Israeli National Intelligent Systems
Project on Artificial Intelligence, Ethics and Regulation (Nov. 19, 2020),
https://ekarine.org/wp-admin/pub/AIEthicsRegulationReport-English.pdf
914
232
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
2) Accountability (including transparency, explainability, ethical
and legal responsibility)
3) Protecting human rights (including bodily integrity, privacy
autonomy, civil and political rights)
4) Cyber and information security
5) Safety (including internal safety and external safety)
6) Maintaining a competitive market
The Committee found that “Privacy protection regimes are currently
facing a significant gap between the principled importance of consent to
collect and use information and a reality where this agreement is based on
standard forms that often do not serve the purpose of agreement. This
complexity also affects the AI areas, as it is based on the processing of
personal information.” The Committee also recommended that the Privacy
Protection Authority would be responsible for AI applications and decision
involving personal data. The Committee recommended new authorities and
resources for the agency. The Committee noted that “The ability to
anonymize personal data, at a reasonable confidence level, is fundamental
to the development and promotion of AI.”
International Cooperation
In June 2021, a bill was proposed in requiring the Department of
State to establish the United States–Israel Artificial Intelligence Center.918
The purpose of the Center will be to “leverage the experience, knowledge,
and expertise of institutions of higher education and private sector entities
in the United States and Israel to develop more robust research and
development cooperation in specified areas”, such as machine learning,
image classification, object detection, speech recognition, natural language
processing, data labeling, computer vision and model explainability and
interpretability.
In November 2021, Israel joined the Global Partnership on Artificial
Intelligence in recognition of its advanced artificial intelligence
technologies and its adherence to the values of equality and democracy
promoted by the OECD.919
S.2120 - United States–Israel Artificial Intelligence Center Act (Jun. 17, 2021),
https://www.congress.gov/bill/117th-congress/senate-bill/2120/text?r=49&s=1
919
Al-Monitor, Israel joins international artificial intelligence group: Israel was added
today to the Global Partnership on Artificial Intelligence, joining 19 an alliance of
technologically advanced democratic countries (Nov. 11, 2021), https://www.almonitor.com/originals/2021/11/israel-joins-international-artificial-intelligencegroup#ixzz7K2VqCY5k
918
233
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation
According to the press, the Steering Committee on AI, formed by
the Prime Minister in 2018, included 15 subcommittees composed of 300
senior people from the government, the Israel Defense Forces, institutions
of higher education, civil society, and the technology industry. The
recommendations of the AI Steering Committee were slated for submission
to the government in January 2020 but that has not occurred.920
AI Week
Tel Aviv University's first international AI Week took place in
November 2019 and explored the role of artificial intelligence in medicine,
computer vision, startups, transportation, human capital development and
more. 921 AI Week for 2020 was postponed. Tel Aviv University has
announced AI Week for 2021, February 22-24, as a virtual event.922 The
2022 edition of AI Week took place virtually in February 7-9.923
Privacy and Data Protection
Complementary laws govern data protection in Israel: legal texts
and guidelines: (1) the 1992 Basic Law: Human Dignity and Liberty
according to which the right for privacy is a constitutional right, 924 (2) the
1981 Privacy Protection Law (PPL)925 and subsequent regulations,926 such
as Israel’s 2017 Data Security Regulation,927 and (3) the guidelines of the
Israeli Privacy Protection Authority.928 Chapter 1 of the PPL covers privacy
generally, while Chapter 2 concerns data in storage and sets out various
Uri Berkovitz, Israel's national AI plan unveiled, Globes (Nov. 20, 2019),
https://en.globes.co.il/en/article-israels-national-ai-plan-unveiled-1001307979
921
AAAS EurekAlert!, First AI Week kicks off at Tel Aviv University (Nov. 19, 2019),
https://www.eurekalert.org/pub_releases/2019-11/afot-faw111919.php
922
Tel Aviv University, AI Week: Feb. 22-24, 2021, https://ai-week.com
923
Tel Aviv University, AI Week: Feb. 7-9, 2022, https://ai-week.com/
924 The Knesset, Basic Law: Human Dignity and Liberty (Mar. 17, 1992)
http://www.knesset.gov.il/laws/speciaL/eng/basic3_eng.htm
925
The Knesset, Protection of Privacy Law 1981 (unofficial English translation),
https://www.gov.il/BlobFolder/legalinfo/legislation/en/ProtectionofPrivacyLaw57411981
unofficialtranslatio.pdf
926
IAPP, Protection of Privacy Regulations (Data Security) 2017 (Unofficial translation),
https://iapp.org/media/pdf/resource_center/IS-PROTECTION-OF-PRIVACYREGULATIONS.pdf
927
Assaf Harel, 5 takeaways from the Israeli Privacy Protection Regulations, IAPP (Aug.
5, 2019), https://iapp.org/news/a/five-takeaways-on-the-first-anniversary-of-the-israeliprivacy-protection-regulations/
928
Yoram Shiv and Shira Nager, Israel - Data Protection Overview, OneTrust (Oct.
2020), https://www.dataguidance.com/notes/israel-data-protection-overview
920
234
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
registration, purpose-limitation, transparency and security requirements, as
well as individual rights of access and rectification. 929 Other Chapters
address procedural and enforcement matters as well as the disclosure or
sharing of information by public bodies and liabilities for the publication of
privacy-infringing material in newspapers.
In 2011, the European Commission determined that Israel satisfied
the “adequacy requirement” according to the European Directive 95/46, but
this status is under examination currently due the changes in the new
European Data Protection Regulation (GDPR). The Israeli Ministry of
Justice proposed updates of the PPL, due to the major gap between GDPR
and the current Israeli Law. In 2021, the Israeli Ministry of Justice
announced approval of updates to the Privacy Protection Law. The
amendments include significant new administrative enforcement powers for
the Privacy Protection Authority, including the authority to impose financial
sanctions, updating technological and social definitions, and reducing the
bureaucratic burden on organizations' obligation to register databases.930 In
January 2022, the Israeli government introduced a substantial amendment
to the Protection of Privacy Law (Bill No. 14) aimed to align the PPL with
the EU GDPR at least in part.931
Israel’s Privacy Protection Authority (PPA) is the primary regulator
for matters relating to privacy and data security. 932 The PPA sits within the
Israeli Ministry of Justice and is headed by the Registrar of Databases. The
PPA regulates and enforces data protection across all sectors, private and
public, according to the provisions of the Privacy Protection Law.
Algorithmic Transparency
There are no provisions relating to automated decision-making in
the Privacy Protection Law, but the inspection powers granted to inspectors
can be applied to disclose the usage of personal information by the database
owner. Similar powers are granted to the credit services regulator at the
Bank of Israel, according to the Credit Data Law of 2016.933
The Privacy Protection Authority, Legislation (Oct. 3, 2017) (unofficial translation),
https://www.gov.il/en/Departments/legalInfo/legislation
930
IAPP, Israel pushes forward privacy law amendments (Nov. 8, 2021),
https://iapp.org/news/a/israel-pushes-forward-privacy-law-amendments/
931
For criticism see IAPP, A turning point for privacy laws in Israel (Jan. 26, 2022)
https://iapp.org/news/a/a-turning-point-for-privacy-laws-in-israel/
932
The Privacy Protection Authority,
https://www.gov.il/en/departments/the_privacy_protection_authority
933
The Knesset, Credit Data Law, 5776-2016 (Mar. 29, 2016) (unofficial translation),
https://www.boi.org.il/en/CreditRegister/Documents/Credit%20Data%20Law,%2057762016.pdf
929
235
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Medical Data
In 2018, Israel’s Prime Minister announced the establishment of
a $300 million initiative to make Israel’s large pool of de-identified clinical
data available to researchers, entrepreneurs, and medical institutions to
develop new treatments and personalized medicine among other goals.
“With all records in a common format, AI systems – using machine learning
algorithms – will be able to parse the data, seeking correlations in conditions
and treatments to discern which treatments are likely to be most effective”
said the Director of the Israel Ministry of Health.934
The initiative will encompass a number of projects, including the
establishment of the “Mosaic” health project, which will create a national
information infrastructure for health research in the field of genetics and
medical information. Regulators will work together to make sure
information can be accessed anonymously, maintaining privacy and
securing information and access permissions. Participation in all of the
projects will be exclusively on a voluntary basis.935 Israel’s plans to combine
health maintenance organizations’ digital health records of most Israelis
into a single system for AI and data analytics to tap were confirmed by the
press in September 2019.936
Covid-19 Tracking Controversy
Beginning with emergency measures taken in March 2020, the
Israeli police used mobile-phone location data and AI techniques to attempt
to determine whether those in quarantine were indeed staying in
quarantine. The police arrested 203 people based on this phone location
tracking. A month after the tracking was authorized, the parliamentary
committee in charge of overseeing the practice halted the mobile phone
tracking. The Committee argued that the harm done to privacy outweighed
the benefits of the tracking.937
Moshe Bar Siman Tov, How Israel Turned Decades Of Medical Data Into Digital
Health Gold, Forbes (Mar 26, 2019),
https://www.forbes.com/sites/startupnationcentral/2019/03/26/how-israel-turned-decadesof-medical-data-into-digital-health-gold/?sh=1b576d873ee4
935
https://www.timesofisrael.com/despite-privacy-concerns-israel-to-put-nationsmedical-database-online/
936
Dov Lieber, Israel Prepares to Unleash AI on Health Care, Wall Street Journal (Sept.
15, 2019), https://www.wsj.com/articles/israel-prepares-to-unleash-ai-on-health-care11568599261
937
Knesset News, Foreign Affairs and Defense Committee suspends bill allowing police
to surveil civilian phones to enforce quarantine orders (Apr. 23, 2020),
https://main.knesset.gov.il/EN/News/PressReleases/Pages/press23420b.aspx
934
236
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Israel then turned to Shin Bet, the Israeli Security Agency, to
monitor the general population and track potential Covid patients and their
contacts.938 This action was criticized by human rights activists, as well as
medical associations.
On April 26, 2020, Israel's Supreme Court banned the intelligence
agency from tracing the phone location of those who may be infected with
Covid-19, until new laws are passed.939 "The state's choice to use its
preventative security service for monitoring those who wish it no harm,
without their consent, raises great difficulties and a suitable alternative...
must be found," the court said.940 The Association for Civil Rights in Israel,
one of the groups which brought the court challenge, welcomed the
decision, saying: "Israel must not be the only democracy operating its secret
security service to monitor its citizens, even in the fight against the
coronavirus."
However, the decision simply ended the program under the
emergency powers authorized in March. In July, Israel’s parliament voted
to allow the country’s Internal Security agency to track the contact relations
of Israeli cellphone users for the rest of the year amid a resurgence in new
cases.941 Human rights organizations renewed their objections.942
In a September 2020 opinion, the national Privacy Protection
Authority also objected to the use of the Israeli Internal Security Service
location tracking tool.943 The PPA said that the measure cannot be justified,
and that use would adversely impact the public’s trust in public authorities.
The PPA also questioned the effectiveness of the location tracking tool.
Jonathan Lis, Israel Extends Security Service Tracking of Coronavirus Cases for
Three More Weeks (May 27, 2020), https://www.haaretz.com/israel-news/.premiumisrael-extends-security-service-tracking-of-coronavirus-cases-for-three-more-weeks1.8875700
939
BBC News, Coronavirus: Israeli court bans lawless contact tracing (Apr. 27, 2020),
https://www.bbc.com/news/technology-52439145
940
Cardoza Law School, Versa, Ben Meir v. Prime Minister, HCJ 2109/20, (Apr. 26,
2020), https://versa.cardozo.yu.edu/opinions/ben-meir-v-prime-minister-0
941
Reuters, Israel approves cellphone tracking of COVID-19 carriers for rest of year
(July 20, 2020), https://www.reuters.com/article/us-health-coronavirus-israelsurveillanc/israel-approves-cellphone-tracking-of-covid-19-carriers-for-rest-of-yearidUSKCN24L2PJ
942
Privacy International, Israel's coronavirus surveillance is an example for others - of
what not to do (updated July 21, 2020), https://privacyinternational.org/longread/3747/israels-coronavirus-surveillance-example-others-what-not-do
943
Pearl Cohen, Israel: Privacy Protection Authority Objects to Shabak-Run Location
Tracking for Coronavirus Epidemiological Investigations (Sept. 1, 2020) (includes link to
opinion in Hebrew), https://www.pearlcohen.com/israel-privacy-protection-authorityobjects-to-shabak-run-location-tracking-for-coronavirus-epidemiological-investigations/
938
237
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Social Ranking
Following a proposal to use scoring technologies for Israelis who
may be infected with COVID-19, in April 2020, the PPA also published a
review on the use of technologies for the social ranking of citizens to
achieve social and governmental goals, and the impact of this on the right
to privacy. 944 The April 2020 Review indicated that social ranking systems
have increased in an era of Big Data and are present, in Israel, in forms
such as the credit data rating system or a potential future AI-based system
aimed to rate an individual’s likelihood of contracting COVID-19.
Specifically, the Review outlined that such a system, which would process
location, medical, and personal data, would constitute a serious violation of
the privacy of citizens and should be avoided as far as possible and, where
it cannot be avoided, it must be compliant with data protection law.
Facial recognition
Facial Recognition in Israel is implemented in border control and
Israel has a biometric database of face photos and fingerprints of citizens
and residents, as well as foreigners accessing Israel. A biometric database
was enacted in law in 2009.945 The law provides the basis for the Israeli
national ID-Card. The database includes biometric face-photos, and
voluntary supplied fingerprints. According to the biometric database law,
the information can be used for severe crime enforcement, and for state
security tasks. In May 2020 the Israeli State Comptroller reported that the
data of about 4.5 million Israeli drivers’ licenses, including facial pictures,
are not sufficiently protected from misuse or outside hacking.946
Still, Israel's military has invested tens of millions of dollars to
upgrade West Bank checkpoints with AnyVision facial recognition
technology to verify Palestinian workers’ identities and ease their entry into
Israel. The new system, which began rolling out late 2018, drew criticism
about the role the controversial technology plays in Israel's military control
over Palestinians.947
OneTrust, Israel: PPA publishes background review on the use of social ranking
systems (Apr. 24, 2020) (includes links to report and summary, in Hebrew),
https://www.gov.il/he/departments/publications/reports/social_ranking
945
Wikipedia, Biometric Database Law,
https://en.wikipedia.org/wiki/Biometric_Database_Law
946
The Jerusalem Post, 4.5 million citizens’ details insufficiently protected, comptroller
says (May 4, 2020), https://www.jpost.com/Israel-News/45-million-citizens-detailsinsufficiently-protected-comptroller-says-626847
947
Daniel Estrin, Face Recognition Lets Palestinians Cross Israeli Checkposts Fast, But
Raises Concerns (Aug. 22, 2019), https://www.npr.org/2019/08/22/752765606/facerecognition-lets-palestinians-cross-israeli-checkposts-fast-but-raises-conc
944
238
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Microsoft, which was part of a group that had invested $74 million
in AnyVision, hired a team of lawyers to audit the Israeli firm and determine
whether AnyVision’s technology applications complied with Microsoft’s
ethical principles against using facial recognition for mass surveillance. In
March 2020, Microsoft said it was pulling investments from AnyVision
although the outcomes of the audit did not substantiate claims that the
startup’s technology was used unethically. 948
In 2021, the Washington Post released further information on the
program.949 The initiative involves a smartphone technology called Blue
Wolf that captures photos of Palestinians’ faces and matches them to an
image database. The phone app flashes in different colors to alert soldiers if
a person is to be detained, arrested or left alone. To create the database,
soldiers competed in 2020 in photographing Palestinians, including children
and the elderly. The total number of people photographed is unclear. The
Israeli military has also installed face-scanning cameras in the divided city
of Hebron to assist soldiers at checkpoints identify Palestinians before they
present their ID cards. A wider network of closed-circuit television cameras,
provides real-time monitoring of the population and can sometimes see into
private homes.
A new security tool is also under development: body cameras with
facial recognition technology to enable police to scan crowds and detect
suspects in real time, even if their faces are obscured.950
OECD AI Principles
Israel has endorsed the OECD AI Principles. Israel has not joined
the Global Partnership on AI. Also, the OECD did not identify any
examples of implementation of the AI Principles in the summary 2020
report. In the 2021 report, the OECD noted that Israel is investing in
language technologies. The AI R&D Framework and Activities of the
Matt O’Brien, Microsoft divests from Israeli facial-recognition startup (March 28,
2020), https://www.timesofisrael.com/microsoft-divests-from-israeli-facial-recognitionstartup/
949
Washington Post, Israel escalates surveillance of Palestinians with facial recognition
program in West Bank (Nov. 27, 2021),
https://www.washingtonpost.com/world/middle_east/israel-palestinians-surveillancefacial-recognition/2021/11/05/3787bf42-26b2-11ec-8739-5cb6aba30a30_story.html
950
France24, Israeli firm develops body cams with facial recognition (Jan. 23, 2022)
https://www.france24.com/en/live-news/20220123-israeli-firm-develops-body-camswith-facial-recognition
948
239
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Innovation Authority is supporting the development of Hebrew-language
Natural Language Processing tools.951
Human Rights
Israel is a signatory to many international human rights treaties and
conventions and is considered a free country, receiving a score of 76/100
for political rights and civil liberties.952 Freedom House reports that “Israel
is a multiparty democracy with strong and independent institutions that
guarantee political rights and civil liberties for most of the population.
Although the judiciary is active in protecting minority rights, the political
leadership and many in society have discriminated against Arab and other
minorities, resulting in systemic disparities in areas including political
representation, criminal justice, education, and economic opportunity.”
Autonomous Weapons
Israel is developing lethal autonomous weapons, including both the
Iron Dome defensive system953 and the Harop suicide drone.954 The Israeli
mission to the GGE on LAWS of the Convention on Certain Conventional
Weapons clarified Israel’s position in August 2019.955 In August 2020,
Israel expressed further views on the Eleven Guiding Principles Adopted by
the Group of Government Expert concerning lethal autonomous weapons
system. Israel’s view is that “the law of armed conflict, or international
humanitarian law (IHL), applies to the potential development and use of
emerging technologies in the area of LAWS; that human judgment will
always be an integral part of any process regarding emerging technologies
in the area of LAWS, and will be applied during their life-cycle; and that
OECD, State of Implementation of the OECD AI Principles – Insights from National
Ai Policies (2021) https://www.oecd-ilibrary.org/docserver/1cd40c44-en.pdf
952
Freedom House, Freedom in the World 2021 – Israel (2021),
https://freedomhouse.org/country/israel/freedom-world/2021
953
Yaniv Kubovich, Israel Deploys Iron Dome Amid Islamic Jihad Leader's
Assassination Anniversary, Haaretz (Nov. 11, 2020), https://www.haaretz.com/israelnews/israel-iron-dome-gaza-islamic-jihad-leader-s-assassination-hamas-1.9303330
954
The Week India, Why Indian Army is eyeing a mini ‘suicide drone’ from Israel (July
14, 2020), https://www.theweek.in/news/india/2020/07/14/why-indian-army-is-eyeing-amini-suicide-drone-from-israel.html
955
Group of Governmental Experts of the High Contracting Parties to the Convention on
Prohibitions or Restrictions on the Use of Certain Conventional Weapons Which May Be
Deemed to Be Excessively Injurious or to Have Indiscriminate Effects, Draft Report of
the 2019 session of the Group of Governmental Experts on Emerging Technologies in the
Area of Lethal Autonomous Weapons Systems (Aug. 21, 2019) (Annex IV)
https://www.unog.ch/80256EDD006B8954/(httpAssets)/5497DF9B01E5D9CFC125845
E00308E44/$file/CCW_GGE.1_2019_CRP.1_Rev2.pdf
951
240
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
humans will always be responsible for the use of LAWS.” Moreover, in
Israel’s view, “besides the potential risks that may be associated with
LAWS, there are also operational advantages to the use of LAWS as well
as clear advantages from the humanitarian perspective.”956
According to New York Times, Israel used an AI-equipped,
remotely operated gun to kill an Iranian nuclear scientist.957 The Israeli
system used AI to correct for more than a second and a half of input delay,
allowing the system’s operator to fire the gun at a moving target while
stationed more than 1,000 miles away. As assassins able to kill targets from
afar, such attacks may become much easier to carry out and, therefore, more
frequent.
Evaluation
The current circumstances of Israel’s AI policies and practices are
confusing and complex. Although Israel is a leader in AI research and
development, efforts to develop a coherent national AI strategy are still
lagging behind other countries. There is good work underway on AI ethics
and a well-established legal system for data protection, but the general
population tracking for sensitive medical condition by the internal security
agency with AI technique is of concern. Also troubling is the use of facial
recognition technology without clear legal basis, the reluctance to support
limits on lethal autonomous weapons, and the deployment of new
techniques for AI-assisted assassination. Israel has endorsed the OECD AI
principles, and works in cooperation with other countries on AI policy, but
has not yet expressed support for the Universal Guidelines for AI,
particularly with the US and the countries participating in the Global
Partnership for AI.
Permanent Mission of Israel to the UN, Israel Considerations on the
Operationalization of the Eleven Guiding Principles Adopted by the Group of
Governmental Experts (Aug. 31, 2020), https://documents.unoda.org/wpcontent/uploads/2020/09/20200831-Israel.pdf
957
The New York Times, The Scientist and the A.I.-Assisted, Remote-Control Killing
Machine (Sep. 23, 2021), https://www.nytimes.com/2021/09/18/world/middleeast/irannuclear-fakhrizadeh-assassination-israel.html
956
241
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Italy
National AI Strategy
In July 2020, the Italian Ministry of Economic Development issued
the National Strategy for Artificial Intelligence.958 The document is the
result of the public consultation closed in September 2019 on the draft
version,959 and a background paper providing initial guiding principles and
policy recommendations as a basis for Italy’s AI strategy.960 The Italian
strategy fits within the lines of the White Paper on Artificial Intelligence of
the European Commission. The National Strategy is part of the European
Coordinated Plan for Artificial Intelligence and must be placed in the
context of a synergy between Member States and European institutions. It
therefore arises from the awareness that only with joint and coordinated
actions Europe will be able to compete with the most advanced countries.
Besides, the strategy is the result of the debate and negotiation at the
international level like the OECD and cooperation within the G7 and G20,
precisely the Global Partnership on AI, in which Italy participates together
with 13 other states and the European Union.
It is worth mentioning the White Paper on Artificial Intelligence,
presented by the Agency for Digital Italy (AgID) in 2018, underlining the
opportunities offered by AI for improvement of public services and the
relationship between public administration and citizens.961 The Italian
Ministry of Economic Development will monitor and evaluate the progress
of the national AI strategy on a continuous basis and update its
implementation where needed.
Italian Ministry of Economic Development, Proposte per una strategia nazionale per
l’intelligenza artificiale (2019),
959
Italian Ministry of Economic Development, Strategia Nazionale per l’Intelligenza
Artificiale. Ministero dello Sviluppo Economico (2019),
https://www.mise.gov.it/images/stories/documenti/Strategia-Nazionale-IntelligenzaArtificiale-Bozza-Consultazione.pdf
960
Italian Ministry of Economic Development, Proposte per una strategia italiana per
l‘intelligenza artificiale. Gruppo di Esperti MISE sull’intelligenza artificiale (2019),
https://www.mise.gov.it/images/stories/documenti/Proposte-per-una-strategia-italiana2019.pdf
961
Agency for Digital Italy, White Paper on Artificial Intelligence (2018),
https://ia.italia.it/assets/librobianco.pdf.
958
242
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The document provided 82 recommendations which will constitute
the Italian strategy within the European Coordinated Plan on Artificial
Intelligence.962 The document is structured in three parts:
1) analysis of the global, European and national market of artificial
intelligence;
2) description of the national strategy on artificial intelligence;
3) policy recommendations monitoring of the national strategy.
The strategy on AI aims to achieve not only industrial
competitiveness in the aforementioned sectors but also the well-being of
humanity and the planet, the so called “RenAIssance.” The strategy calls
for anthropocentric approach to AI based on three pillars driving the
development of technologies and policies:
• AI for human beings: The first level concerns the individual and the
relationship with “the machine.” AI technologies must be at the
service of people, guaranteeing human supervision, preventing
social and territorial imbalances deriving from unaware and
inappropriate uses. It is about defining and implementing initiatives
related to safety, public administration, health and medicine,
education, new skills, policies for work and digital humanities,
media and the cultural and creative industry.
• AI for a reliable, productive and sustainable digital ecosystem: The
second level includes industrial policies for the manufacturing
sector (Industry 4.0). AI must be designed and implemented in a
reliable and transparent way, so that it can be adopted in any area
productive. This concerns the promotion of robotics and
autonomous systems, software, data processing, IoT, finance,
pharmaceuticals and biotech.
• AI for sustainable development: The third level focuses on
sustainability. AI technologies must generate opportunities of
growth and well-being for all individuals, in line with the principles
contained in Article 3 of Italian Constitution and the United Nations
Sustainable Development Goals. This goal includes actions related
to environmental protection and sustainable infrastructures such as
smart cities, transport, agriculture, space.
The budget provides a starting point based on a mixed public and
private investment which amounts to 888 million euros in 5 years. Besides,
the strategy underlines the need of 605 million (121 million per year) of
private contributions. There are six areas of investment: 1) IoT,
Communication from the Commission to the European Parliament, the European
Council, the Council, the European Economic and Social Committee and the Committee
of the Regions, Coordinated Plan on Artificial Intelligence COM(2018) 795 final.
962
243
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
manufacturing and robotics; 2) services, health and finance; 3) transports,
agriculture and energy; 4) aerospace and defense; 5) public administration;
6) culture, creativity and digital humanities.
Strategic Programme on AI 2022-2024
Building on the earlier National Strategy for Artificial Intelligence,
in November 2021, the Italian government announced its Strategic
Programme on AI for 2022-2024.963 The Programme recognised that the
Italian AI ecosystem has vast potential, yet not fully exploited. The
ecosystem is characterized by vibrant research communities but these often
lack scale, struggle to attract foreign talent and could also benefit from
improvements in the gender representation and patenting performance.
With regard to AI solution providers, the Italian industry is growing rapidly
but its economic contribution still remains below its potential, especially
compared to peer countries in Europe. The current Italian context and
international position thus call for a radical upgrade and update of Italy’s
national AI strategy characterised by the need to build on the positive
elements of its ecosystem while focusing on reforms and investments on the
specific areas of weakness. To that end, the Strategic Programme indicated
6 objectives of the Italian strategy in line with an EU-centred approach to
AI. These are to:
1) Advance frontier research in AI;
2) Reduce AI research fragmentation;
3) Develop and adopt human-centred and trustworthy AI;
4) Increase AI-based innovation and the development of AI
technology;
5) Develop AI-driven policies and services in the public sector; and
6) Create, retain and attract AI talent in Italy.
Furthermore, the Strategic Programme identified 11 priority areas:
industry and manufacturing; education system; agri-food; culture and
tourism; health and well-being; environment, infrastructures and networks;
banking, finance and insurance; public administration; smart cities, areas
and communities; national security; and information technologies. The
Strategic Programmes also stated three areas of intervention: strengthening
and attracting the talents and competences that will enable the AI-driven
economy; expanding funding of advanced research in AI; and favouring the
adoption of AI and its applications both in the public administration and in
the Italian economy at large.
OECD, Strategic Programme on Artificial Intelligence 2022-2024 (2021)
https://wp.oecd.ai/app/uploads/2021/12/Italy_Artificial_Intelligence_Strategic_Program
me_2022-2024.pdf
963
244
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
National AI Ecosystem
Several centers of excellence characterize the Italian AI research
ecosystem, precisely, the Artificial Intelligence and Intelligent Systems
Laboratory (AIIS) of the Italian Interuniversity Consortium for Informatics
(CINI), the Italian Institute of Technology (IIT) and the Institute for
Calculation and Networks for High Services (ICAR) of the National
Research Council (CNR). The Italian government will reinforce public
funding and encourage public-private venture capital support in the field of
artificial intelligence, blockchain and Internet of Things. For instance,
Smart&Start Italia is government-funded scheme for new businesses in the
digital economy.964 The National Innovation Fund established in 2019 is
another source of resources up to €1 billion.965 The government is also
setting up advisory services through the appointment of innovation
managers that will help SMEs during the technological and digital
transformation process. Concerning the public sector, the Agency for
Digital Italy recently released a white paper on artificial intelligence at the
service of citizens (see below).
In terms of networking, 8 Competence Centers, established by the
Ministry of Economic Development, and 12 European Technology
Clusters, set up by the Ministry of Education, will form the basis for a
national network for knowledge exchange and collaboration. These
integrate the Digital Europe Programme for the period 2021-2027,966
together with the establishment of Digital Innovation Hubs. The Italian
strategy mentions its proactive support to European initiatives like the
Confederation of Artificial Intelligence Laboratories in Europe (CLAIRE)
and the public-private partnerships for electronic components and systems
(ECSEL).
The strategy also aims to encourage the development of the data
economy by supporting the creation of a Common European Data Space.967
This is based, for instance, on improving the interoperability and
Sostegno alle startup innovative (Smart & Start Italia),
https://www.mise.gov.it/index.php/it/incentivi/impresa/smart-start.
965
Fondo Nazionale Innovazione,
https://www.mise.gov.it/index.php/it/incentivi/impresa/fondo-nazionale-innovazione.
966
Proposal for a regulation of the European Parliament and of the Council establishing
the Digital Europe programme for the period 2021-2027, COM/2018/434 final (June 6,
2018), https://eur-lex.europa.eu/legalcontent/EN/TXT/?uri=COM%3A2018%3A434%3AFIN
967
Communication from the Commission to the European Parliament, the Council, the
European Economic and Social Committee and the Committee of the Regions, Towards a
common European data space (COM 2018), https://eur-lex.europa.eu/legalcontent/en/ALL/?uri=CELEX%3A52018DC0232
964
245
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
accessibility of public administration data through API interfaces. To
facilitate data exchanges, it is proposed to focus on Data Sharing
Agreements, in particular in strategic sectors, and Data Trust models to
ensure data sharing in a fair, safe and equitable way.
Concerning the development of digital infrastructures, the Italian
government is participating in the Joint Undertaking to develop a
competitive European computing ecosystem (EuroHPC).968 Italy is further
expanding its ultra-broadband optical fibre network and 5G network. The
plan also considers high-performance computing (HPC). The worldwide
excellences such as Eni’s Green Data Center in Ferrera Erbognone and
Cineca’s Leonardo supercomputer are two examples showing how Italy
weighs 1.2% in the global HPC panorama (around 50 petaflops). The plan
proposes to double this capacity by investing € 70 million in 5 years.
The OECD notes that Italy has an increasing number of healthcare
applications and AI technologies, leveraging data in the research sector,
hospital medical records, reports and laboratory tests. Italy’s Ministry for
University and Research has launched a National AI Doctoral Program
which aims at recruiting around 200 doctoral candidates all over the
country. There is now a Memorandum of Understanding between the
Minister of Technological Innovation and Digitization and Fondazione
Leonardo to shape the framework and boundaries for AI adoption in Public
Administration. There is also exploration of a specific platform to improve
the level of citizen education on AI matters, with a view to fostering idea
generation for future adoption and ensuring a better understanding of
trustworthiness on use cases where AI is used.969
Human capital
The development and implementation of AI technologies firmly
depends on skills and competences. The Government has already shown its
intention to strengthen the provision of AI competences at all education
levels. At the primary and secondary education level, the government has
launched the National Plan for the Digital School to update school curricula
and promote new skills in digital education and AI-related courses.970 At
higher education levels, the government is encouraging the integration of
courses with AI-related themes in bachelors, masters and doctoral
968
The European High Performance Computing Joint Undertaking,
https://ec.europa.eu/digital-single-market/en/eurohpc-joint-undertaking
969
OECD, Examples of National AI National Policies 61-62 (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
970
Scuola Digitale, https://www.istruzione.it/scuola_digitale/allegati/Materiali/pnsdlayout-30.10-WEB.pdf
246
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
programs. The planned budget also aims to support projects among PhDs,
researchers and professors.
Besides, literacy campaigns will be fostered via broadcasting and
multimedia. Special attention will be devoted to informing about fake news
and issues of cyber security. At the same time, the strategy underlines the
need of new plans to support small and mid-size business in the AI
deployment and update the skills of the workforce. To increase
the international attractiveness of Italy in the field of AI, Italy will focus on
attracting foreign talents through instruments such as the EU Blue
card,971 and the Italian Startup Visa.972
The Italian Institute for Artificial Intelligence (I3A)
The strategy also includes the creation of the Italian Institute for
Artificial Intelligence as a single point of contact at the international level
which can collect different interests and perspectives on AI technologies.
The Institute aims to become one of the leading research institutes in
Europe. It will consist of a hub with central laboratories and 7 centers
specialized in the priority sectors identified by the Strategy who will work
in connection with universities or other institutes already active.
The Institute will work according to a multi-year strategic plan with
periodically updated objectives and an autonomous governance but
synchronized with the strategic lines of national governance and with
universities and other centers of excellence already active also to be able to
seize opportunities for development in connection with other technological
trends (e.g., 5G, Industry 4.0, cybersecurity).
White Paper on Public Administration
In 2018, the AgID launched the White Paper on Artificial
Intelligence. The objective is to give an important impulse to innovation in
the public sector. The White Paper defines a plan to facilitate the adoption
of AI technologies in the Italian Public Administration and improve the
quality of public services. Artificial intelligence technologies can indeed be
implemented in healthcare, education, security, urban management. The
White Paper includes a set of recommendations defining the challenges for
developing and implementing AI technologies in the public sector. The
White paper defines nine challenges:
• The ethical challenge: the anthropocentric vision on artificial
intelligence technologies leads to look at AI technologies as at the
EU Blue Card Network, Italy, https://www.apply.eu/BlueCard/Italy/
The Italian Government policies to attract and retain innovative entrepreneurs from
all over the world, http://italiastartupvisa.mise.gov.it/#homepage
971
972
247
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
•
•
•
•
service of humans. In this case, it is important to ensure that these
technologies meet universal needs. The characteristics of AI
technologies leads to raising questions concerning the quality of
data, transparency and accountability, as well as protection of rights
and freedoms. This step is critical in the public sector to ensure
transparency and the respect of individuals’ rights and freedoms.
The technological challenge: AI technologies cannot still replicate
the functioning of the human mind. There is the interest in
improving and implementing these technologies to make the work
of the Public Administration more effective.
The skills challenge: citizens increasingly deal with digital
technologies. Therefore, it is critical they understand how the Public
Administration implements and uses artificial intelligence
technologies to take decisions or provide public services. Civil
servants need to constantly improve their skills to ensure they can
effectively be aware of the opportunities and challenges of the
implementation of AI technologies in the public sector.
The data challenge: data quality is one of the primary issues when
implementing artificial intelligence technologies. Open data of
public bodies can provide important information that would be very
useful to generate applications of artificial intelligence at the service
of the citizens. Therefore, it is critical to ensure equal and nondiscriminatory access to public data.
The legal challenge: in the field of AI technologies, is necessary to
reconcile the principle of transparency of administrative acts and
procedures with the protection of privacy and personal data. A
second issue of transparency concerns intellectual property rights
over algorithms. Moreover, when the public administration
implements decision-making process, it is necessary to deal with
accountability.
The implementation challenge: training public employees,
particularly officials and managers, on the functioning, benefits, as
well as ethical and technical implications on the use of AI
technologies is critical to ensure the development of the public
sector.
The inequalities challenge: AI solutions can reduce social
inequalities in the field of education and training, health and
disability, knowledge and human rights. However, AI technologies
can also increase inequalities like in the case of biased outputs.
Therefore, the Public Administration should focus on implementing
248
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
•
•
these technologies ensuring inclusiveness, accessibility,
transparency, non-discrimination.
The measurement challenge: The implementation of new
technologies impact on citizens and institution. The Public
Administration has not always the instruments to measure these
effects. However, the introduction of AI technologies in the public
sector can provide more information while requiring an impact
assessment.
The human being challenge: citizens and institutions should be
aware of the effects of automated systems. Artificial intelligence
systems are not only a matter of technology but also social
innovation.
The Rome Call for Ethics
This initiative is aimed at increasing awareness of the role of ethics
in AI.973 The document was signed in February 2020 by the Pontificia
Accademia per la Vita, Microsoft, IBM, FAO and the Italian Government
and proposes a more human-centric approach to AI. The Declaration sets
out a program of “Algorithm Ethics” according to the “fundamental
principles of good innovation,” including Transparency, Responsibility,
Impartiality, Reliability, Security and privacy. The Call is based on three
principles:
• Ethics: All human beings are born free and equal in dignity and
rights.
• Education: Transforming the world through the innovation of AI
means undertaking to build a future for and with younger
generations.
• Rights: The development of AI in the service of humankind and the
planet must be reflected in regulations and principles that protect
people – particularly the weak and the underprivileged – and natural
environments.
Public Participation and Access to Documents
The national AI strategy followed a 2018 consultation. The Italian
Ministry of Economic Development formed a 30-member group of experts
to draft a national strategy on AI.974 The group was comprised of ten
Pontificia Accademia per la Vita, Rome Call for AI Ethics
(Feb. 28, 2020), http://www.academyforlife.va/content/pav/it/events/workshopintelligenza-artificiale.html
974
Governo Italiano, Ministry of Economic Development, Artificial intelligence (AI): call
for experts (Sept. 14, 2018), https://www.mise.gov.it/index.php/en/news/2038605artificial-intelligence-ai-call-for-experts
973
249
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
representatives of enterprises operating in the field of AI, ten
representatives of research centres / think tanks or academia, and ten
representatives of the labour market, professions, consumers and civil
society.975 The group was tasked with developing recommendations on:
• improving, coordinating and strengthening the research in the AI
field;
• promoting public and private investments in AI, also benefitting
from the dedicated EU funds;
• attracting talent and developing business in the field of AI;
• encouraging the development of the data-economy, paying
particular attention to the spreading and valorisation of nonpersonal data, adopting the better standards of interoperability
and cybersecurity;
• the legal framework with specific regard to safety and
responsibility related to AI-based products and services;
• the socio-economic impact of development and widespread
adoption of AI-based systems, along with proposals for tools to
mitigate the encountered issues.
A 2020 survey of Italian consumers by BEUC, the European
Consumer organization, found substantial public concern about the
deployment of AI.976 More than half of respondents disagreed or strongly
disagreed that current regulation is adequate to efficiently regulate AI. Over
70% of respondents in Italy “strongly agreed that users should be able to
say ‘no’ to automated decision-making.” More than half “(strongly) agreed
that companies use AI to manipulate consumer decisions.”
In 2021, Italy hosted the G20 summit. Research institutions in Italy
and around the world participated actively in the preparations for the
Summit. Recognizing the “benefits stemming from the responsible use and
development of trustworthy human-centered Artificial Intelligence (AI),”
the G20 Leaders said in Rome they would encourage competition and
innovation, “as well as diversity and inclusion.”977 Artificial intelligence
figured prominently in the G20 Declaration of the Digital Ministers who
At 14-15. https://www.mcit.gov.sa/sites/default/files/examples-of-ai-nationalpolicies.pdf
976
BEUC, Artificial Intelligence: what consumers say – Finding and policy
recommendations of a multi-country survey on AI (2020),
https://www.beuc.eu/publications/beuc-x-2020078_artificial_intelligence_what_consumers_say_report.pdf
977
G20 Rome Leaders Advance AI Policy, Elevate Privacy, Gender Equality, CAIDP
Update 2.40 (Oct. 31, 2021), https://www.caidp.org/app/download/8352831663/CAIDPUpdate-2.40.pdf
975
250
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
met in Triese, Italy. They stated, “We reaffirm our willingness to implement
trustworthy Artificial Intelligence (AI) and to commit to a human- centred
approach, as decided in 2019 under the Japanese G20 Presidency, guided
by the G20 AI Principles, drawn from the OECD Recommendations on
AI.”978 The Ministers further said, “In the design of our policies, we consider
the specific needs of MSMEs and start-ups, for the implementation of
trustworthy AI that is human-centred, fair, transparent, robust, accountable,
responsible, safe and secure and protects privacy, so as to encourage
competition, innovation, diversity and inclusion.”
Facial Recognition
According to Privacy International, the municipality of Como, Italy,
purchased a facial recognition system “with little transparency and despite
the lack of a clear legal framework.”979 Privacy International reported that
Como “embraced a narrative of technological innovation pushed by
Huawei” within the broader concept of smart city and innovation tech, but
was forced, after the intervention of the Italian Data Protection Authority,
to suspend the system. The Garanti determined that there was no legal basis
to collect facial images. Subsequent reporting by Wired indicated that the
municipality had changed vendors and also that the system installed most
recently failed to work as proposed.980 In September 2020, AlgorithmWatch
also reported that Italy is exploring the use of facial recognition in football
stadiums.981
In January 2021, EDRi reported how the Italian Police are deploying
dehumanizing biometric systems against people at Italy’s borders.982 These
systems use Automatic Image Recognition System (SARI), initially
G20 Information Centre, Declaration of G20 Digital Ministers: Leveraging
Digitalisation for a Resilient, Strong, Sustainable and Inclusive Recovery (Aug. 5, 2021),
http://www.g20.utoronto.ca/2021/210805-digital.html
979
Privacy International, How facial recognition is spreading in Italy: the case of Como
(Sept. 17, 2020), https://privacyinternational.org/case-study/4166/how-facial-recognitionspreading-italy-case-como
980
Laura Carrer, The Municipality of Como has discovered that his facial recognition
system is not what he had bought: The testing of the video surveillance system with facial
recognition revealed inconsistencies and discrepancies with the tender specifications
(Sept. 28, 2020), https://www.wired.it/attualita/tech/2020/09/28/como-riconoscimentofacciale-collaudo/
981
AlgorithmWatch, In Italy, an appetite for face recognition in football stadiums (Sept.
16, 2020), https://algorithmwatch.org/en/story/italy-stadium-face-recognition/
982
European Digital Rights, Chilling use of face recognition at Italian borders shows why
metric mass surveillance (Feb. 10, 2021) https://edri.org/our-work/face-recognitionitalian-borders-ban-biometric-mass-surveillance/
978
251
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
acquired by the Italian police in 2017, and employed to monitor arrivals of
migrants and asylum seekers on the Italian coasts and related activities. In
doing so, according to EDRi, the Italian Ministry of Interior has ignored the
questions of the national Data Protection Authority (DPA) that is
investigating the facial recognition system that the police wants to use.
In April 2021, the Italian DPA blocked the deployment of the Sari
Real Time system, a facial recognition system that would rely on several
cameras installed in a particular geographical area and capable of scanning
individuals’ faces in real time and then compared against a governmental
biometric database available to law enforcement agencies.983 The DPA
acknowledged that the system would immediately delete images of
individuals who are not deemed as suspects. However, the data protection
agency added that the fact that everyone’s faces would be scanned
indiscriminately in the first place may lead to an evolution of the very nature
of surveillance, marking a shift from targeted surveillance of certain
individuals to the possibility of universal surveillance.
In December 2021, the Italian parliament introduced a moratorium
on video surveillance systems that use facial recognition technologies. This
law introduces for the first time in an EU Member State, a temporary ban
for private entities to use these systems in public places or places accessible
to the public.984 The moratorium will be in force until December 31, 2023
at the latest, unless a new law is introduced before that date. As reported by
EDRi, this is an important development, yet the moratorium contains major
exceptions: it only covers video surveillance systems with facial recognition
and the moratorium allows the police to use such systems subject to a caseby-case approval by the Italian DPA and exempts judicial authorities and
public prosecutors from any control.
OECD/G20 AI Principles
Italy endorsed the OECD and the G20 AI Principles and is a
founding member of the Global Partnership for AI. Italy will host the G20
Ministers in 2021. Progress on the implementation of the AI Principles will
be considered.
DigWatch, Italian data protection authority: Sari facial recognition system proposed
by Ministry of Interior could lead to mass surveillance (Apr. 16, 2021)
https://dig.watch/updates/italian-data-protection-authority-sari-facial-recognition-systemproposed-ministry-interior/
984
European Digital Rights, Italy introduces a moratorium on video surveillance systems
that use facial recognition (Dec. 15, 2021) https://edri.org/our-work/italy-introduces-amoratorium-on-video-surveillance-systems-that-use-facial-recognition/
983
252
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection
In January 2021, an Italian court determined that an algorithm to
evaluate employee performance violates labor laws.985 The case concerned
the ranking algorithm of the food delivery service Deliveroo. The
judge ruled that the algorithm unfairly assessed absent workers noting that
it failed to take account of permissible reasons for absence. The court
ordered the company to pay a fine and legal costs and to post the judgment.
In May 2021, the Supreme Court released its judgment in Garante
per la Protezione dei Dati Personali v. Associazione Mevaluate
Onlu.986 The case concerned the Italian DPA’s 2016 order to Mevaluate
Italia s.r.l., which was originally quashed by the Court of Rome, to suspend
the implementation of its online Artificial Intelligence system capable of
analysing documents voluntarily uploaded by users to provide
reputational ratings. The Court quashed the ruling of the Court of
Rome, which had opined in favour of the lawfulness of the system as data
subjects had provided their consent, and found that the lack of transparency
regarding Mevaluate's algorithms invalidates such consent, thus
violating Article 8 of the EU Charter of Fundamental Rights, a series of
articles of Legislative Decree No. 196 of 2003, and the GDPR. It was
further found that consent can only be valid if the data subject is
appropriately informed about the purposes of processing and freely and
specifically expresses their consent to the same. Consent cannot be
considered as informed, if the logic involved in the algorithm remains
unknown to the data subjects, as was the case in Mevaluate reputational
ranking system.
In July 2021, the Italian DPA issued a 2.6 million Euros penalty to
the on-demand delivery company Foodinho, which was ordered to make a
number of changes to how it operates in the market and amend how its
algorithms function.987 One of the issues of concern was the risk of
discrimination arising from a rider rating system and of relevance has been
the decision by the Supreme Court, a discussed above.
Forbes, Deliveroo Rating Algorithm Was Unfair To Riders, Italian Court Rules (Jan.
5, 2021), https://www.forbes.com/sites/jonathankeane/2021/01/05/italian-court-findsdeliveroo-rating-algorithm-was-unfair-to-riders/?sh=34eb0a9e22a1
986
Data Guidance, Italy: Court of Cassation rules that algorithm must be transparent for
consent to be valid (May 25, 2021), https://www.dataguidance.com/news/italy-courtcassation-rules-algorithm-must-be For the judgment (in Italian) see
http://www.italgiure.giustizia.it/xway/application/nif/clean/hc.dll?verbo=attach&db=snci
v&id=./20210525/snciv@s10@a2021@[email protected]
987
IAPP, Italian DPA fines food delivery app 2.6M euros for GDPR violations (Jul. 6,
2021), https://iapp.org/news/a/italian-dpa-fines-food-delivery-app-3m-euros-for-gdprviolations/
985
253
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Italy was a primary sponsor of the 2018 GPA Declaration on Ethics
and Data Protection in Artificial Intelligence and a signatory to the 2020
GPA Resolution on Accountability in the Development and Use of
Artificial Intelligence.988
Algorithmic Transparency
Italy is a member of the European Union and has ratified Council of
Europe Convention 108+. Italians have a general right to obtain access to
information about automated decision-making and to the factors and logic
of an algorithm. There is a data protection agency in Italy with independent
authority. In 2021, the Italian DPA ruled that opaque automated decisionmaking violated the GDPR.989
Human Rights
Italy is a signatory to the major international human rights
instruments, and generally ranks highly for the defense of human rights.
Freedom House rated Italy 90/100 in 2021, for political rights and civil
liberties, a slight increase from 2020.990
Evaluation
Italy has emerged as a leader in the field of AI policy. Italy has
endorsed the OECD/G20 AI Principles. The national strategy incorporates
a strong commitment to fundamental rights and reflects the active
participation of many public and private constituencies. Italy is subject to
the GDPR and has ratified the modernized Council of Europe Convention
108, providing a high level of protection for personal data and specific right
of algorithmic transparency. Moreover, the Rome Call for AI Ethics,
undertaken by Pope Francis with the support of the Italian government and
private companies, sets out a powerful vision for AI that is human-centric
and that diminishes social inequality.
International Conference on Data Protection and Privacy Commissioners, Declaration
on Ethics and Data Protection in Artificial Intelligence (Oct. 23, 2018),
https://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf; Global Privacy Assembly, Resolution on
Accountability in the Development and Use of Artificial Intelligence (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/FINAL-GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN-1.pdf
989
Id.
990
Freedom House, Freedom in the World 2021 – Italy (2021),
https://freedomhouse.org/country/italy/freedom-world/2021
988
254
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Japan
National AI Strategy
Under the direction of former Prime Minister Shinzo Abe, Japan has
emerged as a global leader for both AI policy and data governance. Abe
declared in 2019 that “Artificial Intelligence (AI) must be used solely for
humans and humans must be held responsible for its outcome. We will take
the lead in establishing human-centered ethical principles for AI.”991
Earlier, in 2016, Prime Minister Abe called for the Japanese
government to establish an “Artificial Intelligence Technology Strategy
Council.”992 The Council set out an Artificial Intelligence Technology
Strategy and Industrialization Roadmap.993 The Roadmap focuses on
public-private collaboration along the AI “full pipeline from Rresearch and
devlopment to social implementation.” Priority areas include productivity;
health, medical care, and long-term care; mobility; and information
security. The roadmap includes three phases: (1) the development and
application of AI within various domains, (2) the public use of data and AI
across those domains, and (3) the creation of ecosystems that integrate
domains together. In August 2018, an action plan specified the objectives
and timetable for accomplishment for each initiative under the Strategy.
The government established in parallel separate opportunities for
examination of ethical aspects of AI technology, intellectual property rights,
personal information protection, and promotion of open data, as crosssectional items.994
Japan’s updated the AI Strategy in 2019995 and again 2021.996 “AI
for Everyone: People, Industries, Regions and Governments” focuses on the
measures that the Japanese government should immediately take in a
Prime Minister of Japan, Speeches and Statements by the Prime Minister, Policy
Speech by Prime Minister Shinzo Abe to the 198th Session of the Diet (Jan. 28, 2019),
https://japan.kantei.go.jp/98_abe/statement/201801/_00003.html
992
Prime Minister of Japan, Council for Science, Technology and Innovation (Sept. 15,
2016), https://japan.kantei.go.jp/97_abe/actions/201609/15article2.html
993
Strategic Council for AI Technology, Artificial Intelligence Technology Strategy
(Mar. 31, 2017), https://www.nedo.go.jp/content/100865202.pdf; MIC, AI Strategy and
Related Activities in Japan (Oct. 25, 2017), http://events.sciencejapon.org/dlai17/doc/MIC%20-%20France-Japan%20Symposium%2020171025.pdf
994
Strategic Council for AI Technology, Artificial Intelligence Technology Strategy
(Mar. 31, 2017), https://www.nedo.go.jp/content/100865202.pdf (top page 6)
995
Prime Minister’s Office, Japan, AI Strategy 2019: AI for Everyone: People, Industries,
Regions and Governments (June 11, 2019),
https://www.kantei.go.jp/jp/singi/ai_senryaku/pdf/aistratagy2019en.pdf
996
Cabinet Office, Science and Technology / Innovation, AI Strategy 2021,
https://www8.cao.go.jp/cstp/ai/index.html
991
255
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
concerted manner. It “establishes an integrated policy package for AI that
encompasses educational reform, research and development (R&D) and
social implementation in order to contribute to the world, overcome
challenges, and ultimately improve Japan's industrial competitiveness.”
The 2019 strategic objectives are: (1) to align human resources with
the needs of the AI era; (2) to strengthen industrial competitiveness; (3) to
achieve a sustainable society that incorporates diversity; (4) to build
international research, education, and social infrastructure networks in the
AI field, and (5) to accelerate AI-related R&D. The Strategy notes that it
respects the basic principles set out by the government in the "Social
Principles of Human-Centric AI."
AI R&D Guidelines
The Japanese AI R&D Guidelines influenced global AI policies.
The Japanese government proposed international discussion on AI policy at
the G-7 ICT Ministers’ meeting in 2016 and presented the Guidelines to the
G-7 Leaders meeting in Turin, Italy in 2017.997 The Japanese AI R&D
Guidelines also contributed significantly to the development of the OECD
AI Principles, the first global framework for AI Policy. The OECD AI
Principles were adopted by 42 countries in May 2019, and then by G-20
Nations at the Leaders’ Summit hosted at Osaka, in June 2019. OECD
Secretary General thanked Prime Minister Abe and said that the OECD AI
Principles, endorsed by the G-20 nations, are “affirming that the AI we want
is centered on people, respects ethical and democratic values, is transparent,
safe and accountable.”
Social Principles of Human-Centric AI
Japan’s 2019 "Social Principles of Human-Centric AI"998 were
developed by the “Council for Social Principles of Human-centric AI"
chaired by Professor Osamu Sudoh. The Social Principles specify the form
of society that Japan should aim for, discuss impacts on society, present a
set of AI social principles and identify issues to consider in AI R&D and
social implementation. They call for all relevant stakeholders to cooperate
and interact closely.
The philosophy that underpins the Social Principles of HumanCentric AI consists of three basic principles: (1) Dignity - a society in which
human dignity is respected; (2) Diversity and Inclusion - a society in which
Conference toward AI Network Society, Draft AI R&D Guidelines (July 28, 2017)
https://www.soumu.go.jp/main_content/000507517.pdf
998
Cabinet Secretariat, Government of Japan, Social Principles of Human-Centric AI
(Feb. 15, 2019), https://www.cas.go.jp/jp/seisaku/jinkouchinou/pdf/humancentricai.pdf
997
256
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
people with diverse backgrounds can pursue their own well-being; and (3)
Sustainability - a sustainable society.
The social principles themselves are meant to be implemented
across the Japanese society, including national and local governments, as
well as in multilateral frameworks. They include seven principles for AI:
(1) Human-Centric - the utilization of AI must not infringe upon the
fundamental human rights guaranteed by the Constitution and international
standards and AI should be developed, utilized, and implemented in society
to expand the abilities of people and allow diverse people to pursue their
own well-being; (2) Education/Literacy – all stakeholders must have an
accurate understanding of AI, knowledge and ethics permitting appropriate
use of AI in society; (3) Privacy Protection – AI should not infringe on a
person's individual freedom, dignity or equality, AI using personal data
should have mechanisms to ensure accuracy and legitimacy, and to allow
individuals to be substantially involved in managing the privacy of their
personal data, personal data must be protected appropriately according to
its degree of importance and sensitivity; (4) Ensuring Security – a risk
management approach is necessary; (5) Fair Competition; (6) Fairness,
Accountability, and Transparency - it is necessary to ensure fairness and
transparency in decision-making, appropriate accountability for the results,
and trust in the technology, so that people who use AI are not subject to
undue discrimination with regard to personal background, or to unfair
treatment in terms of human dignity; and (7) Innovation.
AI R&D Guidelines and AI Utilization Guidelines
The original AI R&D Guidelines are directed at developers. 999
They include 9 principles related to: (1) collaboration; (2) transparency; (3)
controllability; (4) safety; (5) security; (6) privacy; (7) ethics (respect
human dignity and individual autonomy); (8) user assistance; and (9)
accountability.
The more recent (2019) AI Utilization Guidelines1000 provide
practical guidance on matters to be considered by various stakeholders,
including developers, end users, and data providers. Aimed to promote the
benefits of AI and mitigate risk, the Guidelines aim to help AI service
providers and business users to establish their own AI development and
utilization guidelines, based on the Social Principles for Human-centric AI.
The Conference toward AI Network Society, Draft AI R&D GUIDELINES for
International Discussions (July 28, 2017),
https://www.soumu.go.jp/main_content/000507517.pdf
1000
The Conference toward AI Network Society, AI Utilization Guidelines Practical
Reference for AI utilization (Aug. 9, 2019),
https://www.soumu.go.jp/main_content/000658284.pdf
999
257
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Guidelines set out ten principles to be considered, in full or in part,
according to the purpose and social context of AI utilization: (1) proper
utilization; (2) data quality; (3) collaboration; (4) safety; (5) security; (6)
privacy; (7) human dignity and individual autonomy; (8) fairness; (9)
transparency; and (10) accountability.
Data Free Flows with Trust
Prime Minister Abe also put forward the concept of Data Free Flows
with Trust (DFFT) in a speech at the World Economic Forum in January
2019.1001 Abe said, “We must, on one hand, be able to put our personal data
and data embodying intellectual property, national security intelligence, and
so on, under careful protection, while on the other hand, we must enable the
free flow of medical, industrial, traffic and other most useful, non-personal,
anonymous data to see no borders, repeat, no borders.” Abe underscored the
importance of privacy protection, explaining that the DFFT regime should
be built on “non-personal data.” Abe further emphasized that the
appropriate framework for protection and governance on data according to
their sensitivity would allow higher freedom of data flow across borders.
At the 2019 G20 Summit in Osaka, OECD Secretary
Gurria described Abe’s vision for Data Free Flows with Trust as “ambitious
and timely.”1002 The G20 Leaders adopted the concept at the 2019
Summit,1003 and reaffirmed the goal at the 2020 Summit in Riyadh.1004 The
phrase “Data Free Flows with Trust” also appears, with emphasis, in the
December 2020 Joint Communication from the European Communication,
proposing a New US Agenda for Global Change.1005
Ministry of Foreign Affairs, Japan, Speech by Prime Minister Abe at the World
Economic Forum Annual Meeting: Toward a New Era of "Hope-Driven Economy" (Jan.
23, 2019), https://www.mofa.go.jp/ecm/ec/page4e_000973.html
1002
OECD, 2019 G20 Leaders’ Summit - Digital (AI, data governance, digital trade,
taxation), Remarks by Angel Gurría (June 28, 2019),
https://www.oecd.org/g20/summits/osaka/2019-g20-leaders-summit-digital-osaka-june2019.htm
1003
The Japan Times, Full text of the G20 Osaka leaders' declaration (June 29, 2019),
https://www.japantimes.co.jp/news/2019/06/29/national/full-text-g20-osaka-leadersdeclaration/
1004
G20 Riyadh Summit, Leaders' Declaration (Nov. 21-22, 2020),
https://g20.org/en/media/Documents/G20 Riyadh Summit Leaders Declaration_EN.pdf
1005
European Commission and High Representative of the Union for Foreign Affairs and
Security Policy, Joint Communication to the European Parliament, the European
Council and the Council: A New EU-US Agenda for Global Changes, (Dec. 2, 2020)
(emphasis in the original), https://ec.europa.eu/info/sites/info/files/joint-communicationeu-us-agenda_en.pdf
1001
258
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation
Japan organized a conference with public participation in advance
of the 2016 G-7 Ministerial.1006 The conclusions of the conference informed
the 2016 Takamatsu Declaration.1007 The G7 ICT Ministers agreed to
promote ICT technology R&D for Artificial Intelligence. In October 2016,
Japan1008 also launched a new public conference on the theme “Toward AI
Network Society” with the participation of experts from industry, academia,
and citizens to examine the social, economic, ethical, and legal implications
of AI.1009 The AI Network Society conference, chaired by Dr. Osamu
Sudoh,1010 formulated the AI R&D Guidelines. The Japanese government
presented AI R&D Guidelines to the G-7 meeting in Turin, Italy in 2017.1011
A subsequent meeting of the Toward AI Network Society conference
produced the AI Utilization Guidelines, “a commentary on the principles
expected to be taken into consideration in the utilization of AI.”1012
Japan’s AI R&D Guidelines and the AI Utilization Guidelines
influenced the development of AI policy frameworks at the OECD and
elsewhere. The Guidelines promoted the development of AI and addressed
public concerns, with the goal of building trust in the technology. The
Conference is continuously studying the safe, secure, and trustworthy
implementation of AI in the society.1013
In 2021, the Ministry of Economy, Trade, and Industry issued a call
for Public Comments on Call for Public Comments on "AI Governance
The event was organized by the Institute for Information and Communications Policy
(IICP) of the Ministry of Internal Affairs and Communications (MIC).
1007
G7 Information Center, Joint Declaration by G7 ICT Ministers (Action Plan on
Implementing the Charter) (Apr. 30, 2016), http://www.g8.utoronto.ca/ict/2016-ictdeclaration.html
1008
The event was organized by the Institute for Information and Communications Policy
(IICP) of the Ministry of Internal Affairs and Communications (MIC).
1009
Ministry of Internal Affairs and Communications, Japan, The Conference toward AI
Network Society―Release of 2020 Report (July 21, 2020),
https://www.soumu.go.jp/main_sosiki/joho_tsusin/eng/pressrelease/2020/7/21_1.html
1010
Professor at the Faculty of Global Informatics, Chuo University and Project Professor
at the Graduate School of Interdisciplinary Information Studies, University of Tokyo.
1011
The Conference toward AI Network Society, Draft AI R&D GUIDELINES for
International Discussions (July 28, 2017),
https://www.soumu.go.jp/main_content/000507517.pdf
1012
The Conference toward AI Network Society, AI Utilization Guidelines Practical
Reference for AI utilization (Aug. 9, 2019),
https://www.soumu.go.jp/main_content/000658284.pdf
1013
Ministry of Internal Affairs and Communications, Institute for Information and
Communications Policy, The Conference toward AI Network Society―Release of 2020
Report (July 21, 2020),
https://www.soumu.go.jp/main_sosiki/joho_tsusin/eng/pressrelease/2020/7/21_1.html
1006
259
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Guidelines for Implementation of AI Principles Ver. 1.”1014 The call was
undertaken to gather a wide range of opinions and to clarify the trends in
Japan and overseas regarding AI principles and rule-making and to present
an overall picture of how Japan should approach AI governance, including
creating interim guidelines for implementing the Social Principles of
Human-Centric AI.
On February 1, 2022, the Japanese Government published the AI
Governance Guidelines. The guideline clarifies and presents the whole
picture of AI governance that Japan should be, including the creation of
intermediate guidelines for the implementation of human-centered AI social
principles.1015 The Governance Guidelines reflect the advice of companies,
academics, legal experts and auditors. Further consultations will be
conduced with experts in standards and consumer protection.
Data Protection
The Act on the Protection of Personal Information (APPI) governs
data processing in the private sector. The 2020 amendments to the APPI
bring the law closer to the EU’s General Data Protection Regulation
(GDPR).1016 The amendments upgrade individuals’ rights, introduce the
concept of pseudonymization, reinforce data breach reporting and increase
penalties for offenders. The updated APPI also broadens the definition of
personal data to capture facial recognition/biometric data. In January 2019,
the European Commission adopted an adequacy decision on Japan’s APPI,
allowing personal data to flow freely between the two economies.1017
Two laws regulate data processing by government.1018 The APPI
also requires national and local governments to be “responsible for
METI, Call for Public Comments on "AI Governance Guidelines for Implementation
of AI Principles Ver. 1.0" Opens (July 9, 2021) [GT],
https://www.meti.go.jp/english/press/2021/0709_004.html
1015
METI, Governance Guidelines for the Practice of AI Principles Ver. 1.1" has been
compiled (Jan. 28, 2022) [GT],
https://www.meti.go.jp/press/2021/01/20220125001/20220124003.html; see also IAPP,
Japan publishes AI governance guidelines, https://iapp.org/news/a/japan-publishes-aigovernance-guidelines/;
1016
https://www.ppc.go.jp/files/pdf/overview_amended_act.pdf - The 2020 Amendments
will come into force on a date specified by a cabinet order, within two years after
promulgation (June 12, 2020).
1017
European Commission, European Commission adopts adequacy decision on Japan,
creating the world's largest area of safe data flows (Jan. 23, 2019),
https://ec.europa.eu/commission/presscorner/detail/en/IP_19_421.
1018
Act on the Protection of Personal Information Held by Incorporated Administrative
Agencies, etc., No 59 (May 30, 2003),
http://www.japaneselawtranslation.go.jp/law/detail_main?re=&vm=2&id=3397
1014
260
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
comprehensively formulating and implementing the necessary measures to
ensure the proper handling of personal information in conformity with the
purport of this Act.”
The Personal Information Protection Commission (PPC),
established in 2016, supervises the implementation of the APPI. The
members of the PPC exercise their official authority independently. The
PPC also supervises the implementation of the My Number Act, which
regulates the use numeric identifiers for social security and taxation.1019
OECD/G20 AI Principles
Japan endorsed the OECD and the G20 Principles and is a member
of the Global Partnership on AI (GPAI). Japan was also a catalyst for the
adoption of the OECD AI Principles by the G20 Ministerial meeting in
Tsukuba and the G20 Leader’s Summit in Osaka, Japan, in 2019.1020
Algorithmic Transparency
Japanese law does not contain a general right of algorithmic
transparency. However, there are specific provisions for certain sectors. For
example, for financial services, the "Comprehensive Guidelines for
Supervision over Major Banks" require that the concerned individual be
provided with specific explanations on the reasons for the rejection of a
request to conclude a loan agreement.1021
Use of AI for policy decisions
The government is considering a data analysis system developed by
Palantir for public agency decision-making, according to Japan Times.1022
AI systems are also under consideration for defense, national security, trade
management, and public health. The move complements the plans by the
Personal Information Protection Commission, Act on the Use of Numbers to Identify
a Specific Individual in the Administrative Procedure,
https://www.ppc.go.jp/files/pdf/en3.pdf
1020
CAIDP Update 1.7, Prime Minister Abe’s AI and Data Governance Legacy (Aug. 30,
2020), https://dukakis.org/center-for-ai-and-digital-policy/caidp-update-prime-ministerabes-ai-and-data-governance-legacy/
1021
Official Journal of the European Union, Commission Implementing Decision (EU)
2019/419 of 23 January 2019 pursuant to Regulation (EU) 2016/679 of the European
Parliament and of the Council on the adequate protection of personal data by Japan
under the Act on the Protection of Personal Information (March 19, 2019) (par. 93),
https://eur-lex.europa.eu/legalcontent/EN/TXT/HTML/?uri=OJ:L:2019:076:FULL&from=DE
1022
The Japan Times, Japan considers using AI for speedy policy decisions (Nov. 2,
2020), https://www.japantimes.co.jp/news/2020/11/02/national/japan-ai-policygovernment/.
1019
261
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
administration of Prime Minister Yoshihide Suga to accelerate
digitalization.1023
Facial Recognition
Japan has deployed facial recognition in several sectors, including
transportation, banking (ATMs), police and immigration. According to
Japan Times, Japan planned to use facial recognition technology, originally
intended for security purposes, to prevent the spread of the novel
coronavirus when it hosted the Tokyo Olympics and Paralympics in
2021.1024 1025 Osaka Metro Co. has developed automated ticket gates
with facial recognition with a view to equip all metro stations in Osaka by
2024, ahead of the 2025 World Expo.1026 Likewise, the Japanese Ministry
of Economy, Trade and Industry is testing facial recognition ticketing on
driverless buses in several cities across the country.1027
In September 2020, Japan Times reported that Japanese Police
Forces have been using facial recognition technology across the nation since
March 2020 to locate criminal suspects. Critics warned that the system
could transform the country into a surveillance society unless it run under
strict regulations, a senior National Police Agency (NPA) official said “we
are using the system only for criminal investigations and within the scope
of the law. We discard facial images that are found to be unrelated to
cases.”1028 The NPA manages and utilizes facial images under rules set by
the National Public Safety Commission,1029 as it does with fingerprints and
1023
Analytics India Magazine, Use Of Algorithmic Decision Making & AI In Public
Organisations (Nov 11, 2020), https://analyticsindiamag.com/use-of-algorithmicdecision-making-ai-in-public-organisations/ - 13/11/2020
1024
The Japan Times, Facial Recognition, https://www.japantimes.co.jp/tag/facialrecognition.
1025
Find Biometrics, Japan to Pair Face Recognition with Mask and Temperature
Detection During Tokyo Olympic (Oct. 22, 2020), s https://findbiometrics.com/japanpair-face-recognition-mask-temperature-detection-during-tokyo-olympics-102209/
1026
The Japan Times, Osaka Metro unveils ticket gate with facial recognition tech (Dec.
10, 2019), https://www.japantimes.co.jp/news/2019/12/10/business/corporatebusiness/osaka-metro-facialrecognition/#:~:text=on%20Tuesday%20started%20testing%20a,around%201%2C200%
20Osaka%20Metro%20employees.
1027
NFCW, Japanese passengers test facial recognition ticketing on driverless buses
(Sept. 10, 2020) https://www.nfcw.com/2020/09/10/367826/japanese-passengers-testfacial-recognition-ticketing-on-driverless-buses/
1028
Biometric Update, Police in Japan reveal use of facial biometrics in criminal probes
(Sept. 16, 2020), https://www.biometricupdate.com/202009/police-in-japan-reveal-useof-facial-biometrics-in-criminal-probes
1029
The National Public Safety Commission is a Japanese Cabinet Office commission
which guarantees the neutrality of the police system by insulating the force from political
262
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
DNA. The agency’s database currently holds 10 million facial images of
criminal suspects.
Japan does not have specific legislation for facial recognition in the
government sector. As of September 2020, the Japanese APPI covers the
use of facial biometric data gathered from security cameras. 1030 Law
enforcement is however exempt from this type of privacy regulations. The
APPI also allows the use of anonymized facial recognition data beyond the
intended purposes if such data is sufficiently protected from being restored
to its original form.
Human Rights
Japan is signatory to many international human rights treaties.
According to Freedom House, Japan rates among the top countries in the
world for political rights and civil liberties.1031
Evaluation
Japan is a pioneer in the field of AI policy and has endorsed the
OECD/G20 AI Principles. The Conference toward AI Network Society,
established in 2016, is broadly influential. The Japanese R&D Guidelines
provided the basis for the OECD AI Principles. Japan also hosted the G20
Leaders’ meeting in Osaka in 2019 at which time the G20 nations endorsed
the OECD AI Principles. And former Prime Minister Shinzo Abe promoted
the concept of Data Free Flow with Trust (DFFT), a core concept for
human-centric AI, that carries forward in the policy recommendations of
the OECD, the G20, and the European Commission. However, concerns
about the unregulated use of facial recognition remain. While there has been
no express support for the Universal Guidelines for AI, Japan’s policies
reflect elements found in the UGAI.
pressure and ensuring the maintenance of democratic methods in police administration. It
administers the National Police Agency, and has the authority to appoint or dismiss
senior police officers.
https://en.wikipedia.org/wiki/National_Public_Safety_Commission_(Japan)
1030
Biometric Update, Police in Japan reveal use of facial biometrics in criminal probes
(Sept. 16, 2020), https://www.biometricupdate.com/202009/police-in-japan-reveal-useof-facial-biometrics-in-criminal-probes
1031
Freedom House, Freedom in the World 2021 – Japan (2021),
https://freedomhouse.org/country/japan/freedom-world/2021
263
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Kazakhstan
Overview and National AI Strategy
In recent years, Kazakhstan has taken an active interest in AI and
new technologies to reform the country’s economy and extend internal
security and policing.1032 Kazakhstan has also expressed interest in Russia’s
AI development.1033 Although the precise AI strategy has not yet been
defined, the government has set out an AI and other smart technologies
agenda. This includes the cultivation and creation of new industries with the
use of digital technologies, and productivity growth through the widespread
introduction of automation, robotics, AI, and the exchange of big data. 1034
The state program “Digital Kazakhstan” describes the implementation of
this agenda through the realization of the following projects in the AI
sector:1035
• Creation of an international technopark of IT start-ups (Astana
Hub)
• Creation of model factories based on Industry 4.0 technologies
• Development of open platforms (Open API), Big Data, and AI
• Development of telecommunications infrastructure, including
broadband internet access
• Development of innovative financial technologies
• Implementation of Smart City components
The country has established several IT and research centers that are
planned to be the flagships for the development of AI in Kazakhstan:
Nazarbayev University, Astana International Financial Centre, Astana
International Technology Park of IT Startups Despite these aspirations,
Kazakhstan has only scored 46.55 out of 100 in the Government AI
Trend News Agency, Work is underway in Kazakhstan to introduce the concept of
"Data-Driven Government" (Dec. 4, 2020),
https://www.trend.az/casia/kazakhstan/3345220.html
1033
Tass, Kazakhstan interested in Russia’s experience in AI development (Dec. 4, 2020)
(“President Kassym-Jomart Tokayev pointed out that digitalization of the Eurasian
Economic Union (EAEU) should become a top priority for the Eurasian Economic
Commission”), https://tass.com/world/1231509
1034 The President of Kazakhstan Nursultan Nazarbayev’s Address to the Nation of
Kazakhstan. (Jan. 31, 2017), Third Modernization of Kazakhstan: Global
Competitiveness http://www.akorda.kz/en/addresses/addresses_of_president/thepresident-of-kazakhstan-nursultan-nazarbayevs-address-to-the-nation-of-kazakhstanjanuary-31-2017
1035 On approval of the State Program "Digital Kazakhstan"( Об утверждении
Государственной программы "Цифровой Казахстан")
http://adilet.zan.kz/rus/docs/P1700000827
1032
264
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Readiness Index, with the lowest score in the technology sector.1036 A dearth
of qualified IT specialists1037 and low R&D spending (2.70 out of 100)1038
are two significant barriers to a dynamic and innovative technology sector.
According to the prime minister Askar Mamin, Kazakhstan simply does not
have financial resources for the development of its own AI technologies.1039
The country is trying to address this problem with the help of foreign
investors1040 and international partners.1041
It was announced in 2020 that World Bank will work with
Nazarbayev University to create a National Cluster of Artificial Intelligence
with its own laboratory, a data processing research center and a science park
for the development of artificial intelligence.1042 Among other plans was
establishing active cooperation in developing common standards, rules, and
policies in the field of data exchange and integration. In April of the same
year, the Kazakh Ministry of Education and Science, along with the World
Bank, launched the Fostering Productive Innovation Project (FPIP)1043 to
support and develop high-quality scientific research on, and the
commercialization of new technologies.
AI Core Values
One of Kazakhstan’s primary purposes of embracing AI is to spark
foreign investment in the country to diversify the economy and reduce its
The Government AI Readiness Index 2020, Oxford Insights
https://www.oxfordinsights.com/government-ai-readiness-index-2020
1037
В будущем «цифровом Казахстане» не хватает IT-специалистов The future,
"digital Kazakhstan" lacks IT specialists, Radio Free Europe,
https://rus.azattyq.org/a/programma-cifrovoi-kazakhstan-deficit-itspecialistov/28625463.html
1038
Research and development expenditure (% of GDP) – Kazakhstan, The World Bank
https://data.worldbank.org/indicator/GB.XPD.RSDV.GD.ZS?locations=KZ&view=map
1039
Готов ли Казахстан к технологической гонке? Is Kazakhstan ready for a
technology race? https://forbes.kz//process/intellektualnaya_zadacha_1588745463/?
1040 Caspian Policy Center, Kazakhstan — The Buckle in the Belt and Road Initiative
Seeks Investment and Growth (Jan. 31, 2020), https://www.caspianpolicy.org/kazakhstanthe-buckle-in-the-belt-and-road-initiative-seeks-investment-and-growth/
1041
Kazakhstan seeks high-tech, agricultural cooperation with China, says Tokayev
during Beijing Business Council meeting, https://www.euractiv.com/section/centralasia/news/kazakhstan-seeks-high-tech-agricultural-cooperation-with-china-says-tokayevduring-beijing-business-council-meeting/
1042 EAEU prime ministers participate in Digital Almaty Forum
https://primeminister.kz/en/news/premer-ministry-eaes-prinyali-uchastie-v-forumedigital-almaty1
1043 Kazakhstan: Fostering Productive Innovation Project
https://projects.worldbank.org/en/projects-operations/project-detail/P150402
1036
265
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
economic dependence on natural resources.1044 Social governance and
welfare is another purpose of developing AI. 2017’s AI agenda mentions
the threat of terrorism growth and the prevention of religious extremism
propaganda on the Internet and social networks.1045
Facial Recognition and Smart Cities
Facial recognition surveillance technology is becoming increasingly
widespread in Kazakhstan. In October 2019, facial recognition technologies
were first installed on buses.1046 Notably, President Tokayev had even paid
a visit and discussed future cooperation with Hikvision,1047 a Chinese stateowned surveillance company under U.S. sanctions1048 that provided the
hardware for Kazakhstan’s newly established surveillance system. In the
same year, the small city of Akkol was proclaimed the first complete “Smart
City” in Kazakhstan. Akkol is digitally monitored by an AI-based facial
recognition surveillance system, the functions of which include thermal
imaging, searching for a car by number plates, recognizing missing persons,
detecting the presence of weapons in schools, hospitals and other public
places.1049 Similarly, over 4,000 cameras blanket Nur-Sultan, the
capital.1050 In 2020, the authorities announced that Kazakhstan would be
Kazakhstan's Ai Aspirations https://www.rebellionresearch.com/blog/kazakhstan-sai-aspirations
1045
The President of Kazakhstan Nursultan Nazarbayev’s Address to the Nation of
KazakhstanThird Modernization of Kazakhstan: Global Competitiveness (Jan. 31, 2017),
http://www.akorda.kz/en/addresses/addresses_of_president/the-president-of-kazakhstannursultan-nazarbayevs-address-to-the-nation-of-kazakhstan-january-31-2017
1046
The Four Big Issues Central Asia Faced In 2019 (And They're Not Going Away),
Radio Free Europe, https://www.rferl.org/a/central-asia-2019-challenges-security-chinafacial-recognition/30356077.html
1047
«Распознает даже людей в масках». Нужны ли Казахстану камеры Hikvision? "It
even recognizes people in masks." Does Kazakhstan need Hikvision cameras? Radio Free
Europe, https://rus.azattyq.org/a/kazakhstan-china-survelliance-camera/30210035.html
1048
Bloomberg, U.S. Blacklists Eight Chinese Tech Companies on Rights Violations:
Move comes as U.S.-China high-level trade talks set to resume Action targets Chinese
surveillance companies, public entities (Oct. 7, 2019),
https://www.bloomberg.com/news/articles/2019-10-07/u-s-blacklists-eight-chinesecompanies-including-hikvision-k1gvpq77
1049 Видеонаблюдение, безопасность и комфорт. Как живет самый умный город
Казахстана - Smart Aqkol Video surveillance, security and comfort. How the smartest
city of Kazakhstan lives - Smart Aqkol, Tengrinews,
https://tengrinews.kz/article/videonablyudenie-bezopasnost-komfort-jivet-samyiyumnyiy-1353/
1050
Как работает проект "Сергек". Репортаж Informburo.kz How the Sergek project
works. Informburo.kz report https://informburo.kz/stati/kak-rabotaet-proekt-sergekreportazh-informburokz.html
1044
266
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
spending $23 million to install facial recognition software in its largest city,
Almaty.1051
While the government insists that the main goal is to maintain
public safety, many activists are worried that this will ultimately create a
totalitarian surveillance state,1052 especially since the companies that are
behind the surveillance system in Kazakhstan are under U.S. sanctions for
unethical use of AI technology.
Medical AI
Since mid-March 2020, Kazakhstan's government has been fighting
the novel coronavirus. The Kazakhstani Ministries of Health and Internal
Affairs ultimately turned to AI technological solutions to confront the
coronavirus outbreak. The range of technologies being reoriented to enforce
quarantine and curfews include traffic cameras, facial recognition
technologies, and smartphone apps.1053 The ministry has required the 8,000
or so Kazakhstani citizens currently under quarantine to use the
SmartAstana tracking app allowing officials to guarantee these individuals
remain in isolation and monitored citizens through facial recognition video
surveillance technology to find violators of the quarantine regime in
Almaty. By the end of the country’s two-month state of emergency on May
11, 2,424 people had been charged with violating quarantine in Almaty and
3,347 in Nur-Sultan.1054According to experts, the pandemic exacerbated the
existing arbitrary and uneven policing practices as surveillance is
augmented by national and municipal authorities without public
oversight.1055
AI Ethics
Kazakhstan has not yet adopted OECD AI Principles, nor did it
define ethical norms and standards for AI. Nevertheless, the country’s major
AI research center, Institute of Smart Systems and Artificial Intelligence at
1051
На камеры с распознаванием лиц в Алма-Ате выделили $23 млн (Alma-Ata
allocated $ 23 million for cameras with face recognition), (Feb. 8, 2019),
http://fergana.agency/news/105020/
1052
Kazakhstan embraces facial recognition, civil society recoils, Eurasianet,
https://eurasianet.org/kazakhstan-embraces-facial-recognition-civil-society-recoils
1053
Technology and Policing a Pandemic in Central Asia
https://thediplomat.com/2020/05/technology-and-policing-a-pandemic-in-central-asia/
1054
World Politics Review, Police States Expand Under the Cover of COVID-19 (July
14, 2020), https://www.worldpoliticsreview.com/articles/28910/across-central-asiapolice-states-expand-under-the-cover-of-covid-19
1055
The Diplomat, Technology and Policing a Pandemic in Central Asia (May 13, 2020),
https://thediplomat.com/2020/05/technology-and-policing-a-pandemic-in-central-asia/
267
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Nazarbayev University,1056 states that it operates in accordance with the
following ethical principles:
• Societal Well-being
• Human Centered Values
• Transparency
• Technical Resilience and Robustness
• Accountability
Data Protection
In 2019, the country experienced a wave of major data breaches
from the databases of the CEC and the Prosecutor General's Office.1057 Soon
after that, the personal information of 11 million people were published
online and could be accessed by anyone through a published database.1058
These incidents led to the amendment of the existing data protection
law, which was revised to mostly align with the GDPR.1059 Amendments to
the regulation of digital technologies and to the Personal Data Law entered
into force on July 7, 2020.1060 The new regulations establish a data
protection agency, create rules for the collection and processing of personal
data and introduce the concept of “personal data safety protection service.”
The Personal Data Law includes a requirement that the content and amount
of personal data collected strictly correspond to the specific, previously
declared and legal purposes of their processing. Nevertheless, the GDPR
requires “the appropriate data protection training to personnel having
permanent or regular access to personal data” whereas Kazakhstan’s
amendments do not require data protection training. Training is important
because human error is one of the major causes of data breaches across the
world.
Nazarbayev University, Institute of Smart Systems and Artificial Intelligence,
https://issai.nu.edu.kz/about/
1057
Catalin Cimpanu, Extensive Hacking Operation Discovered in Kazakhstan, ZDNet,
(Nov. 23, 2019), https://www.zdnet.com/article/extensive-hacking-operation-discoveredin-kazakhstan/.
1058
Злоумышленники выложили в сеть данных миллионов казахстанцев Attackers
have posted data of millions of Kazakhstanis to the network
https://kursiv.kz/news/obschestvo/2019-07/zloumyshlenniki-vylozhili-v-set-dannyemillionov-kazakhstancev
1059
The Law of the Republic of Kazakhstan No. 94-V dated May 21, 2013 “On Personal
Data and Their Protection” (hereinafter, the “Personal Data Law”).
1060
Dentons, Amendments on Personal Data Protection Issues in Kazakhstan (July 14,
2020), https://www.dentons.com/en/insights/articles/2020/july/14/amendments-onpersonal-data-protection-issues-in-kazakhstan
1056
268
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Algorithmic Transparency
Although Kazakhstan is not directly subject to the GDPR, the
Personal Data Law reflects modern concepts of transparency and data
subject access. Kazakhstan is also eligible to ratify the Council of Europe
Modernized Convention on Privacy.1061
OECD AI Principles
Kazakhstan has not endorsed the OECD AI Principles. According
to the OECD AI Observatory, the national initiatives also do not address
any of the OECD AI principles.1062
Human Rights
According to Freedom House, Kazakhstan rates poorly (23/200) for
political rights and civil liberties.1063 Freedom House reports
“Parliamentary and presidential elections are neither free nor fair, and major
parties exhibit continued political loyalty to the government. The authorities
have consistently marginalized or imprisoned genuine opposition figures.
The dominant media outlets are either in state hands or owned by
government-friendly businessmen. Freedoms of speech and assembly
remain restricted, and corruption is endemic.” On transparency, Freedom
House states “The government and legislature offer little transparency on
their decision-making processes, budgetary matters, and other operations.
The media and civil society do not have a meaningful opportunity to provide
independent commentary and input on pending laws and policies. A law on
public access to government information was adopted in 2015, but it is
poorly implemented in practice.”
Kazakhstan is eligible for admission to the Council of Europe. In
recent years, Kazakhstan has increased cooperation with the Council of
Europe. A previous agreement was limited to criminal justice. The
Neighbourhood Co-operation Priorities for Kazakhstan 2019-2022
introduces new areas of co-operation, including the fight against economic
crime, promoting a common legal space and human rights standards, and
assistance in the electoral field. The document was adopted by the
Committee of Ministers in April 2019. Kazakhstan participates in the
Council of Europe, Chart of signatures and ratifications of Treaty 108 (Status as of
Nov. 11, 2019), https://www.coe.int/en/web/conventions/full-list//conventions/treaty/108/signatures?p_auth=UMypWMxn
1062
OECD AI Policy Observatory, “Digital Kazakhstan Government Programme (Oct. 6,
2021), https://oecd.ai/en/dashboards/policyinitiatives/http:%2F%2Faipo.oecd.org%2F2021-data-policyInitiatives-25280
1063
Freedom House, Freedom in the World 2021 – Kazakhstan (2021),
https://freedomhouse.org/country/kazakhstan/freedom-world/2021
1061
269
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Central Asia Rule of Law Programme, launched in 2020, which has the goal
of “Improving the lives of citizens by reinforcing human rights, democracy
and rule of law.”1064
Evaluation
Kazakhstan has set an ambitious goal of embracing new
technologies and boosting productivity. As Kazakhstan rushed into the
digital future by quickly importing and implementing AI surveillance
technologies, it failed to implement oversight legislation, responsible use of
AI ethics standards and principles and allow broad public discussion of what
constitutes public safety and privacy. While AI can provide security and
prosperity, advanced surveillance technologies and deep troves of
identifying data can pose a threat to citizens if oversight mechanisms and
ethical standards are not properly established.
Council of Europe / European Union, Central Asia Rule of Law Programme,
https://pjp-eu.coe.int/en/web/central-asia
1064
270
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Kenya
National AI Strategy
Kenya does not have legislation, national policy or strategy on
Artificial Intelligence (AI).1065 However, Kenya is one of the top five
African countries in the 2021 global Government Artificial Intelligence
Readiness Index, ranking 71st.1066 It is estimated that over the last decade,
Kenya's total value of investment in AI is Sh13 billion (US$120 million).1067
In February 2018, the government empanelled a blockchain and AI
taskforce. 1068 The task force's mandate was to provide the government with
recommendations on harnessing the emerging technologies of blockchain
and AI in Kenya. In addition, the taskforce was tasked to explore the use of
AI in public service delivery, financial inclusion, cyber-security, and
election processes.1069
In July 2019, the taskforce published its report which identified
three domains of AI development and application. First, it identified the
need to leverage blockchain and AI in the fight against corruption.1070
Secondly, the report identified the critical role of AI in the financial
sector.1071 Lastly, the report explores the application of AI in elections and
states that AI could “bolster election fairness through fast tallying and
providing real-time polling results, and by extension, strengthen
democracy”.1072 The report also avers that effective regulation of the
technologies will potentially balance citizen protection and private sector
innovation.1073 The report recommended:
1065
J Kabubu, Official Intelligence in Kenya (Jan. 26, 2021),
https://mman.co.ke/content/artificial-intelligence-ai-kenya.
1066
A Gwagwa & Others ‘Artificial intelligence (AI) deployments in Africa: Benefits,
challenges and policy dimensions’ (2020) 26 The African Journal of Information and
Communication (AJIC) at 1-28.
1067
F Ngila ‘Kenya, Africa hurdles in artificial intelligence race’ Business Daily 7
January 2021 https://www.businessdailyafrica.com/bd/corporate/technology/kenyaafrica-hurdles-in-artificial-intelligence-race-3249180 .
Kabubu
Ibid
1070
Ibid
1071
K Abuya, Kenya Blockchain Taskforce Findings Rally for Use Cases in Poll
Transparency, Teckweez 26 July 2019 https://techweez.com/2019/07/26/blockchaintaskforce-report/.
1072
Abuya (n 2).
1073
D Mpala ‘Kenyan taskforce calls for state to regulate AI and blockchain’
https://ventureburn.com/2019/08/kenya-report-blockchain-ai/.
1068
1069
271
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
1) Drafting a digital asset framework that will enable Kenyan citizens
to raise funds through initial coin offerings (ICOs)1074;
2) Create a digital locker (complementing a digital ID) for every
citizen to securely store official documents like credit reports and
birth certificates;
3) Use blockchain to track agricultural produce from end-to-end
(from seeds to marketplace) and to use AI and analytics to detect
fraud, trace unsafe products;
4) Distribute farming subsidies through a blockchain controlled agritoken;
5) Introduce blockchain technology to enable customers to trace the
supply chain of medication;
6) Develop a health token incentive to reward citizens who maintain a
healthy lifestyle; and
7) Use blockchain supply-chain networks to flag and report
counterfeit goods.1075
Although the taskforce presented the report to the government, the
proposed recommendations have not been translated into any public
policy, legislation or strategy on AI.1076
In addition, in November 2019, Kenya enacted the Data Protection
Act, establishing some protections for personal data.1077 AI systems use
private data to perform better, and this has considerable privacy and social
risks. As Das observes, this is particularly associated with “how some
organizations are collecting and processing a vast amount of user data in
their AI-based system without their knowledge or consent, which can lead
to concerning social consequences”.1078 However, the Act does not include
AI. According to Ngila: “Kenya's Data Protection Act of 2019 only takes
care of data privacy, totally leaving out AI, with the two becoming
increasingly inseparable”.1079
Initial Coin Offerings (ICOs) are a popular fundraising method used primarily by
startups wishing to offer products and services, usually related to the cryptocurrency and
blockchain space. See https://www.investopedia.com/terms/i/initial-coin-offering-ico.asp
1075
D Mpala ‘Kenyan taskforce calls for state to regulate AI and blockchain’
https://ventureburn.com/2019/08/kenya-report-blockchain-ai/.
1076
Kabubu
1077
Freedom House, Freedom in the World 2020,
https://freedomhouse.org/country/kenya/freedom-world/2020
1078
S Das ‘The Social Impact of Artificial Intelligence and Data Privacy Issues’
https://www.red-gate.com/simple-talk/development/data-science-development/the-socialimpact-of-artificial-intelligence-and-data-privacy-issues/
1079
Ngila (n 3 above).
1074
272
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation
The development of Kenya’s AI report by the Blockchain and
Artificial Intelligence taskforce followed a public consultation process. The
taskforce reported that it received 150 presentations and consulted with
about 90 stakeholders.1080 There is no clarity on whether the input from the
stakeholder consultations was incorporated in the report. The composition
of the taskforce was also inclusive and diverse. It consisted of 14 experts
including four women drawn from academia, research institutions, tech
entrepreneurs, consultants and private businesses such as Safaricom, Cisco,
IBM Research Africa, the African Development Bank. The taskforce
membership also includes four women namely: Juliana Rotich, Dr Charity
Wayua, Elizabeth Ondula, and Lesley Mbogo. The Taskforce is led by an
academic and former ICT permanent secretary, Bitange Ndemo.1081 In
addition, there are around 31 AI startups operating in various sectors of
Kenyan economy including agriculture, finance, accounting, health,
communications, education, business development, and law.
In April 2019, the Artificial Intelligence for Development Africa,1082
held a regional conference in Nairobi, Kenya. The theme of the conference
was ”Toward a Network of Excellence in Artificial Intelligence for
Development (AI4D) in sub-Saharan Africa.” It was attended by 60 African
and international experts. The conference was aimed at deepening the
African conversation on AI primarily on: policy and regulations; skills and
capacity building; and the application of AI in Africa. Delegates to the
conference had several aspirations:to have 30 African countries develop AI
specific policies and strategies by 2024; to “create a pipeline of 400 African
PhDs in AI, data science, and other interdisciplinary fields”; to create “a
collective investment of US$ 1 billion dollars in collaborative innovation
and research prioritizing solution areas for sustainable development in
Africa”; to establish an AI Centre of Excellence in each African country by
2030;1083 and to invest in capacity building in AI policy and regulatory
frameworks that are relevant for the African context.1084
C Tanui, The Kenya Blockchain Taskforce Concludes Its Report, Wall Street (Nov. 20,
2018), https://kenyanwallstreet.com/the-kenya-blockchain-taskforce-concludes-report-onblockchain-technology/.
1081
Mpala (n 2).
1082
Artificial Intelligence for Development Africa is tech organization whose object it
“to improve the quality of life for all in Africa and beyond by partnering with Africa’s
science and policy communities to leverage AI through high-quality research, responsible
innovation, and strengthening talent”. See https://africa.ai4d.ai/
1083
https://africa.ai4d.ai/blog-africa-roadmap/
1084
https://tracxn.com/explore/Artificial-Intelligence-Startups-in-Kenya (accessed 24
April 2021).
1080
273
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
However, in the absence of a policy and legal framework, it is difficult to
establish mechanisms for public participation in AI processes.
International Partnership on AI
The 2014 African Union Convention on Cyber Security and
Personal Data Protection is the core regional instrument with relevance to
AI. In a continent widely viewed as a safe haven for cyber criminals, the
Convention aims to harmonise cyber regulation aimed at addressing three
major areas of concern: (1) electronic transactions, (2) personal data
protection, and (3) cyber security and cybercrime.1085 It also empowers
states to regulate the collection, processing, and storage of private data.
1086
Also, article 11 of the Convention calls on member states to establish
national administrative bodies charged with protection of personal data,
while article 24 urges state parties to develop national cyber security
policies. Considering that AI often relies on data, this Convention is crucial
in AI on the African continent. However, Kenya has not ratified the
Convention.1087
In an effort to position Africa for the global shift towards AI, the
African Union has established the AI Working Group to facilitate the
regional approach to AI. The Group held its inaugural meeting in 2019 and
Kenya, a member of the AU with fast growing interest in AI, stands the
opportunity to join this regional initiative on AI. Through this group, the
AU intends to foster collaboration among African states that “could help
countries develop AI strategies, identify other regulatory and governance
issues, and learn from regional best practice."1088
Data Protection
Kenya has a data protection law based on the GDPR. The Data
Protection Bill 2019 make Kenya the third country in East Africa to have
CCDCOE, Mixed Feedback on the ‘African Union Convention on Cyber Security and
Personal Data Protection’
https://ccdcoe.org/incyder-articles/mixed-feedback-on-the-african-union-convention-oncyber-security-and-personal-data-protection/
1086
Article 12 of African Union Convention on Cyber Security and Personal Data
Protection.
1087
African Union, African Union Convention on Cyber Security and Personal Data
Protection (June 27, 2014),
https://au.int/en/treaties/african-union-convention-cyber-security-and-personal-dataprotection
1088
Ngila (n 3 above).
1085
274
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
legislation dedicated to data protection.1089 The Act seeks to: give effect to
Article 31(c) and (d) of the Constitution that contain the right to privacy;
establish the Office of the Data Commissioner; regulate the processing of
personal data; provide for the rights of data ‘subjects’; and create
obligations of data ‘controllers’ and ‘processors.”1090 However, the law
does not provide a right of access to the logic of the processing, as would
be found in the GDPR.
Human Rights
Kenya’s Constitution contains the Bill of Rights under Chapter 4.
The Bill of Rights contains a catalogue of fundamental civil and political
rights on the one hand, and socio-economic and cultural rights on the other.
These rights among others include: Freedom and security of the person,
right to Privacy, Freedom of expression, Freedom of the media, right to
access to information, and right to dignity.1091 Kenya has a Human Rights
Commission which has an oversight mandate on human rights protection
and promotion. Kenya is also a signatory to various regional and
international human rights treaties and conventions including the African
Charter on Human and Peoples Rights, International Convention on Civil
and Political Rights (ICCPR) and International Covenant on Economic,
Social and Cultural Rights (ICSECR).
According to Freedom House, Kenya is “Partly Free” with a score
of 48/100 for political rights and civil liberties.1092 The country’s media and
civil society sectors are vibrant, even as journalists and human rights
defenders remain vulnerable to restrictive laws and intimidation. There are
also concerns on government’s interference with right to privacy of
individuals. In addition, the government frequently uses cybercrime laws to
crackdown online critics of the government.
Evaluation
Despite a global movement towards automation, Kenya is yet to take
off in Artificial Intelligence (AI). It ranked 71 out of 172 nations on the
Government AI Readiness Index 2020. Notably, Kenya does not have AI
Deloitte, Kenya Data Protection Act: Quick Guide (2021),
https://www2.deloitte.com/content/dam/Deloitte/ke/Documents/risk/Kenya%20Data%20
Protection%20Act%20-%20Quick%20Guide%202021.pdf
1090
Kenya Gazette Supplement, Data Protection Act, 2019 (Nov. 11, 2019),
http://kenyalaw.org/kl/fileadmin/pdfdownloads/Acts/2019/TheDataProtectionAct__No24
of2019.pdf
1091
Chapter 4 of the Constitution of the Republic of Kenya
1092
Freedom House, Freedom in the World 2021 – Kenya (2021),
https://freedomhouse.org/country/kenya/freedom-world/2021
1089
275
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
legislation and policy. It has also not ratified the African Union
Convention on Cyber Security and Personal Data Protection. However,
Kenya has established a task force with a mandate to advise the government
on AI development and the report was produced in 2018. There are also
about 39 AI startups in Kenya, which signals the growing interest in AI
interventions in the country. Also, the Data Protection Act, enacted in 2018
provides safeguards for personal data protection, despite not having clear
provisions for AI. In addition, Kenya’s constitution contains a strong Bill
of Rights and has an established Human Rights Commission which has an
oversight mandate over human rights protection and promotion. However,
in the absence of AI legal and policy framework, actual AI practices are
difficult to evaluate.
276
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Korea
National AI Strategy1093
Korea's “National Strategy for Artificial Intelligence” was
announced in December 2019. While its main focus is building a world class
AI technical capacity through ambitious targets such as 'achieving a world
top 3 digital competitiveness by 2030', it also aims to address AI ethics and
algorithmic transparency related issues under the pillar of realizing people
centered AI.
This national strategy materializes the “Presidential Initiative for
AI” that President Moon Jae-in announced on October 28, 2019. During his
remark, he stated “AI is moving beyond scientific and technological
advancements and is approaching us as a new civilization… AI will not only
affect industrial sectors but also solve many issues facing our society: public
health in an aging society, welfare for senior citizens living alone, the safety
of women living by themselves, and the prevention of crimes that are
becoming more sophisticated.”
Korea's “National Strategy for Artificial Intelligence” has nine
major strategies and 100 major tasks in three major areas. The major areas
are (1) fostering a global-leading AI ecosystem, (2) becoming a country
unrivaled for its use of AI, and (3) realizing people-centered AI. This
“National Strategy for Artificial Intelligence” is a result of the cooperation
of the entire Korean ministries and offices including the Ministry of Science
and ICT, the Ministry of the Interior and Safety, and the Ministry of
Education, and the ‘Presidential Committee on the Fourth Industrial
Revolution’ that deliberates upon and coordinates important policy matters
pertaining to the development of AI.1094
In 2021, the Ministry of Science and ICT has shared the consensus
on AI ethics in the annual review meeting and cooperated with UNESCO’s
project by participating in the intergovernmental sessions. At the 11th
Annual Review Meeting, held by UNESCO headquarter and 14 Korean
Ministries, the Korean government presented the results of ongoing projects
supported by various public donors and identified new opportunities for
cooperation. The Ministry of Science and ICT informed their support and
future collaboration for the UNESCO Recommendation.
Ministry of Science and ICT, Policies, National Strategy for Artificial Intelligence
(Mar. 23, 2020)
https://www.msit.go.kr/english/msipContents/contentsView.do?cateId=tst60&artId=2771
576
1094
Presidential Committee on the Fourth Industrial Revolution, About PCFIR (2020)
https://www.4th-ir.go.kr/home/en
1093
277
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In 2021, the Ministry of Science and ICT and Korean Information
Science Development Institute (KISDI) also participated in two
Intergovernmental Meetings and Intersessional Consultations for
intergovernmental negotiation for the Recommendation. Previously, the
Ministry also organized Asia-Pacific Consultation with UNESCO.
Also, the Ministry of Science and ICT provided in-depth comments
on the draft to make the Recommendation actionable to refer to future AI
policy. In that process, the Korean government consented with the purpose
and values of the Recommendation and provided feedback to clarify policy
actions for the draft.
Korea’s National Strategy to Realize Trustworthy AI (2021)
includes several strategies which could be in line with the
Recommendation.1095 In May 2021, the Ministry of Science and ICT
announced the national strategy to build social trust in the era of AI, and it
included strategies for AI impact assessment and AI ethics. We do not have
an English translation of this strategy, but it is in line with the purposes of
‘Part IV. Areas of Policy Action’ in the Recommendation.
The national law for intelligent information society also has
applicable provisions to build the Social Impact Assessment of Intelligent
Information Services framework
AI R&D Strategy
To strengthen its national technological competitiveness, expand
infrastructure, and secure AI talents, Korea announced the AI R&D Strategy
in May 2018 in which the government will invest 2.2 trillion won for five
years (2018 – 2022) in the sectors of brain science, industrial mathematics,
infrastructure, AI technology and talents, and AI service and industry.1096
Furthermore, Korea plans to promote the Next AI R&D Project1097 (2022 –
2026) on the scale of 1 trillion won to go beyond the limitation of the current
AI technology and to become the world’s leading AI technology country.
Above all, it emphasizes the importance of ensuring explainability,
robustness, and fairness of AI R&D activities.
Korea, Artificial Intelligence-Based Policy Division, Announcing Trustworthy AI
Implementation Strategies (May 13, 2021),
https://www.msit.go.kr/bbs/view.do?sCode=user&mId=113&mPid=112&pageIndex=&b
bsSeqNo=94&nttSeqNo=3180239&searchOpt=ALL&searchTxt=
1096
HRST Policy Platform, AI R&D Strategy (May 2018),
https://hrstpolicy.re.kr/kistep/kr/policy/policyPlanKorDetail.html
1097
(footnote #1) National Strategy for Artificial Intelligence, p. 22
1095
278
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Amendments to the Three Major Data Privacy Laws
The Korean government amended the major three data privacy laws
in February 2020 to protect personal information and improve the personal
data protection and privacy governance system in the era of the 4th industrial
revolution.1098 The three laws are the Personal Information Protection Act
(PIPA), the Act on the Promotion of the Use of the Information Network and
Information Projection (“the Network Act”), and the Credit Information
Use and Protection Act (the “Credit Information Act”). The amendments
are the legislative measure that reflects the outcomes of the “Hackathon
agreements” (February and April 2018) in which related ministries, civil
organizations, professionals from industry and the legal circles participated,
and the recommendations of the Presidential Committee on the Fourth
Industrial Revolution (May 2018). It focuses on introducing the concept of
using fictitious names to foster the use of data, reviewing related laws and
unifying the regulatory system, and strengthening the responsibility of the
users.
Personal Information Protection Commission (PIPC)
The amendments to the three major data privacy laws in February
2020 brought together personal information protection functions scattered
across ministries and launched the new organization, PIPC, on August 5,
2020. 1099 The PIPC an integrated supervisory authority with the primary
role of the protection and supervision of personal information. PIPC was
originally under the President as a body run based on compromise and
agreement, but now it has the independent authority for its operation.
The PIPC passed the adequacy decision of the European
Commission, which means that Korea and the EU shared a commitment to
a high level of data protection.1100 Based on the decision, the personal data
will be able to transfer from the EU to Korea safely under the GDPR.
In 2021, the PIPC published AI Personal Information Protection
Self-checklist to provide guidelines for the protection of personal
information gathered and used by artificial intelligence. The checklist
presents 16 specific items to check and 54 items to verify safe handling of
Ministry of Culture, Sports and Tourism, “Data 3 Act” (Mar. 30, 2020)
http://www.korea.kr/special/policyCurationView.do?newsId=148867915
1099
Personal Information Protection Commission, 2019 Personal Information Protection
Policy Performance at-a-glance (Sept. 18, 2020)
http://www.pipc.go.kr/cmt/english/news/selectBoardArticle.do
1100
European Commission, Joint Press Statement by Didier Reynders, Commissioner for
Justice of the European Commission, and Yoon Jong In, Chairperson of the Personal
Information Protection Commission of the Republic of Korea (Dec. 17, 2021),
https://ec.europa.eu/commission/presscorner/detail/en/statement_21_6915
1098
279
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
personal information along with the life cycle of AI: design, development,
and operation of AI. The PIPC also imposed a monetary penalty on an AI
startup (Scatter Lab) for a massive personal data breach in 2021.1101 It was
the first case in Korea that the government has sanctioned the indiscriminate
use of personal information by companies using AI technology.
The Korea Personal Information Protection Commission (PIPC)
signed up for Global Privacy Assembly (GPA) as a regular member since
2012. As a member country, the 34th International Conference of Data
Protection and Privacy Commissioners (ICDPPC) granted accreditation to
the Korean PIPC. In 2018, the PIPC attended the ICDPPC as accredited
members. In 2020, the GPA held the 42nd closed session and adopted a
resolution on the privacy and data protection challenges arising in the
context of the COVID-19 pandemic. The PIPC co-sponsored the resolution
and participated in the GPA COVID-19 taskforce 2020. Furthermore, the
PIPC serves as a member of the GPA’s Policy Strategy Working Group 1:
Global frameworks and standards1102 and Digital Education Working
Group,1103 enhancing data protection and AI accountability in general.
The PIPC is not the only independent agency for AI oversight. The
National Human Rights Commission is a national advocacy institution for
human rights protection.1104 The NHRC was established in 2001 as an
independent agency that does not belong to any legislative, administrative,
or judicial branch. The NHRC continues to advocate the non-discriminatory
use of AI and warn against the risk of bias and deep fake technology.1105
YonhapNews Agency, Developer of AI chatbot service fined for massive personal
data breach (Apr. 28, 2021), https://en.yna.co.kr/view/AEN20210428009500315.
1102
Global Privacy Assembly, Policy Strategy Working Group 1: Global frameworks and
standards (Oct. 2020), https://globalprivacyassembly.org/wpcontent/uploads/2020/10/Day-1-1_2a-Day-3-3_2b-v1_0-Policy-Strategy-WorkingGroup-WS1-Global-frameworks-and-standards-Report-Final.pdf
1103
Global Privacy Assembly, Digital Education Working Group (Oct. 2020),
https://globalprivacyassembly.org/wp-content/uploads/2020/10/DEWG-2019-2020Annual-Report-GPA-20200921-finalannexes_Oct-2020_final-en-211020-1.pdf
1104
National Human Rights Commission of Korea,
https://www.humanrights.go.kr/site/main/index002
1105
National Human Right Commission of Korea,
http://humanrights.go.kr/site/program/webzine/subview?menuid=003001&boardtypeid=1
016&boardid=7605775&searchissue=7605780; Human Rights Commission should put
human rights protection in AI bill... "Severe threats such as 'deep fake porn'" (“The
National Human Rights Commission of Korea officially expressed its opinion that there
is a possibility of discrimination, monitoring, and human rights violations behind the
development of artificial intelligence (AI), and that the National Assembly should reflect
human rights protection regulations when enacting the related fostering law”),
https://view.asiae.co.kr/article/2020060110251892308
1101
280
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Lastly, the Presidential Committee on the Fourth Industrial
Revolution coordinates important policy matters pertaining developing and
acquiring of new science and technology, including AI and data technology.
The PCFIR recommends government promote a trustworthy AI and safe
ecosystem for the AI industry.
Global Partnership – OECD, G20, GPAI, and UNESCO
As a member of the global community, the Korean government has
been actively participating in international cooperation in the AI sector to
promote responsible development and use of AI. Korea endorsed the OECD
Principles in 20191106 as well as the G20 principles and actively participates
in international cooperation to set up the global AI principles. Furthermore,
the Korean government is one of the founding members of the Global
Partnership on AI (GPAI), the world’s first international AI initiative.1107
In July 2020, the Korean government co-organized the Virtual AsiaPacific Consultation on the first draft of the UNESCO Recommendation on
the Ethics of AI with the UNESCO, and led the discussions on values,
principles, and policy tasks regarding the UNESCO Recommendation on
the Ethics of AI.1108
At the Ministerial Council Meeting in October 2021, Korea Ministry
of Science and ICT presented the “Progress over the past two years in
implementing the OECD AI Principles and Future Direction.”1109 A general
director of the AI policy bureau at the Ministry of Science and ICT, Ms.
Kyunhee Song, presented Korea’s progress in accordance with OECD AI
Principles. As part of the effort to implement the OECD AI Principles, the
Korean government also established ‘National Strategy of Artificial
Intelligence (2019)’ and ‘Digital New Deal Strategy (2020) (Data Dam
Projects).
The Korean government has actively participated in the global
discussions on AI ethics, including both the OECD AI principles and the
UNESCO AI recommendation. As result, the Korean government also
OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence
(May 22, 2019) https://www.oecd.org/science/forty-two-countries-adopt-new-oecdprinciples-on-artificial-intelligence.htm
1107
OECD, OECD to host Secretariat of new Global Partnership on Artificial
Intelligence (June 15, 2020) https://www.oecd.org/going-digital/ai/oecd-to-hostsecretariat-of-new-global-partnership-on-artificial-intelligence.htm
1108
Ministry of Science and ICT, Launch of first global AI initiative, GPAI (June 15,
2020)
https://www.msit.go.kr/english/msipContents/contentsView.do?cateId=tst56&artId=2996
961
1109
OECD, Putting the OECD AI Principles into practice: progress and future
perspectives (Oct. 4, 2021), https://oecd.ai/en/mcm
1106
281
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
established Human-centered AI Ethics Standards (2020).1110 Based on the
Ethics Standards, the Korean government implemented action plans for AI
ethics, such as drafting a developer checklist and promoting education for
AI ethics. As part of the Korean government’s effort to implement
trustworthy AI, it plans to disseminate a guidebook that AI developers can
refer to when they confront technological and ethical issues.
AI Ethics
On October 28, 2019, in his “Presidential Initiative for Artificial
Intelligence” President Moon Jae-in stated “The Korean government will
pay special attention to change in the job market and AI-related ethical
issues.”1111 In particular, Korea's “National Strategy for Artificial
Intelligence” includes 'preventing AI dysfunction and establishing AI
ethics' as one of nine major actions and aims to set up AI Ethics Principles
through a public consultation.
As a follow-up action to establish the comprehensive AI ethical
standards which all members of society – developers, providers, and users
– can refer to, from development to use of AI, the Korean government has
formed an AI ethics research team and analyzed the commonalities and
differences in OECD AI Principles and 25 global major AI ethical
principles. As a result, it has drafted the Korean AI ethical standard that
embraces the existing domestic and international ethical principles. The
Korean government plans to announce the Korean AI ethical standards in
December 2020 after hearing opinions of the academia, industry, and civil
organizations.1112
Meanwhile, the Korean government takes a stance that the AI
development should refrain from developing lethal autonomous weapons,
but rather focus on supporting non-weapon systems such as the human
decision-making process and effective management of military supplies. To
this end, it will continue to conduct research activities that correspond to the
1110
Ministry of Science and ICT unveils “National Artificial Intelligence Ethics
Standard” (draft) centered on people,
https://www.msit.go.kr/bbs/view.do?sCode=user&mId=113&mPid=112&pageIndex=&b
bsSeqNo=94&nttSeqNo=3179630&searchOpt=ALL&searchTxt=
1111
Cheong Wa Dae, Remarks by President Moon Jae-in at Korean Artificial Intelligence
Developers Conference “DEVIEW 2019” (Oct. 28, 2019)
https://english1.president.go.kr/Briefingspeeches/Speeches/682
1112
National Strategy for Artificial Intelligence, p.49 (Dec. 2019)
https://www.msit.go.kr/cms/english/pl/policies2/__icsFiles/afieldfile/2020/03/23/Nationa
l%20Strategy%20for%20Artificial%20Intelligence_200323.pdf
282
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
international norms, including discussions on lethal autonomous weapons
at the meetings of the expert group under the UN.1113
Algorithmic Transparency
The Korean government outlined the regulation to secure
accountability, interoperability, and safety of intelligence information
technology in the newly drafted Framework Act on Intelligence
Informatization.1114 The Korean government will determine the details and
level of regulation as a form of ministerial decree after hearing opinions
from the related parties and considering the technology development
progress, AI industry vitalization, and infringement of business’ autonomy.
Korea recently amended the law to reflect new demands of
algorithmic transparency. The Credit Information Use and Protection Act
introduced the right to challenge decisions based on automated processing.
Details regarding the right to challenge decisions based on automated
processing, such as credit extension, and the methods and procedures for
exercising such rights.1115 The Credit Information Act recognizes the data
subject’s right to challenge an automated credit assessment. It defines
“automated credit assessment” as a “credit information company’s or other
act of evaluating a credit information and other data using an information
processing device (such as a computer) without actually involved in the
evaluation of the human individual.” The amendments to the Credit
Information Act do not cover other automated decision-making beyond the
financial sector. However, it is meaningful in the sense that it is the first law
that empowers individual users toward AI transparency.
The PIPC also proposed a bill to amend to Personal Information
Protection Act (PIPA). At the Public institution Affiliated organization
Information Security Conference (PASCON) in 2021, the PIPC proposed
to amend the PIPA.1116 One of the main amendments entails the rights of
data subjects, including rights to request explanation of automated decisions
ZDNet, University boycott ends after ‘KAIST’ confirms no ‘killer robot’ development
(April 10, 2018) https://www.zdnet.com/article/university-boycott-ends-after-kaistconfirms-no-killer-robot-development/
1114
National Law Information Center, (Name of the Law) (June 9, 2020)
https://www.law.go.kr/lsSc.do?section=&menuId=1&subMenuId=15&tabMenuId=81&e
ventGubun=060101&query=%EC%A7%80%EB%8A%A5%EC%A0%95%EB%B3%B4
#undefined
1115
Chambers and Partners, Data Protection & Privacy 2021, South Korea (Mar. 9,
2021), https://practiceguides.chambers.com/practice-guides/comparison/627/6273/1038610395-10401-10406-10414.
1116
https://www.dailysecu.com/form/html/pascon/image/2021/pascon_2021_01.pdf
1113
283
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and rights to object. The proposed bill (2112723) was introduced at the
National Assembly on September 2021, and it is under review.1117
Improvement of Policies and Laws for the Era of AI
In May 2020, the Korean government amended the Framework Act
on National Informatization1118 with the new name of the Framework Act
on Intelligence Informatization to provide fundamental law for the era of
intelligence information. This Act presents the definition of intelligence
information technology and the basis of all regulations that address the
development and use of AI, such as the basic principles of the intelligence
information society, technical requirement, standardization, and personal
data protection. Moreover, the Korean government has introduced the
‘future-oriented legal system’1119 to review and revise regulations to address
the issues of using AI in (1) data, (2) intellectual property, (3)
accountability, (4) regulation of algorithms and trade secret, (5) finance, (6)
platform, (7) labor, (8) healthcare, and (9) welfare.
Human Rights Advocacy
The Korean government has established the National Human Rights
Commission of Korea (NHRCK) in 2001 as a national advocacy institution
for human rights protection.1120 During his congratulatory remark on 2018
Human Rights Day in December, President Moon Jae-in stated “when
human rights are realized in everyday lives, their value is demonstrable…
Human rights are guaranteed through peace, and peace is secured through
human rights.” He also extended his gratitude to NHRCK for “fully
demonstrating the history and significance of the Universal Declaration of
Human Rights.”1121
Legal Business Information, Partial amendment to the Personal Information
Protection Act (draft),
https://www.moleg.go.kr/lawinfo/makingInfo.mo?lawSeq=62160&lawCd=0&&lawType
=TYPE5&mid=a10104010000
1118
National Law Information Center, Framework Act on National Informatization
(2015)
http://www.law.go.kr/lsInfoP.do?lsiSeq=172205&lsId=000028&chrClsCd=010202&url
Mode=engLsInfoR&viewCls=engLsInfoR#0000
1119
(footnote #1) National Strategy for Artificial Intelligence, p. 26
https://www.msit.go.kr/SYNAP/skin/doc.html?fn=14acc067ebaf2780a558e24993a560f0
&rs=/SYNAP/sn3hcv/result/202010/
1120
National Human Rights Commission of Korea, Purpose (2001)
https://www.humanrights.go.kr/site/homepage/menu/viewMenu?menuid=002001001001
1121
Cheong Wa Dae, Congratulatory Remarks by President Moon Jae-in on 2018 Human
Rights Day (December 10, 2018)
https://english1.president.go.kr/Briefingspeeches/Speeches/101
1117
284
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In addition, the Korean government has been an active member in
the seven core international human rights instruments including the
“International Covenant on Civil and Political Rights” and “International
Covenant on Economic, Social, and Cultural Rights.” Especially since its
entry into the United Nations (UN) in 1991 and the Commission of Human
Rights (CHR) in 1993, Korea has been engaged in various international
cooperation activities for improvement of human rights, especially the
rights of the vulnerable and the North Koreans.1122
Evaluation
Korea is one of the leading countries in national AI policies. Korea
has adopted a comprehensive National Strategy for AI and has promoted a
“future-oriented” legal system. Korea has updated national privacy laws,
established a Personal Information Protection Commission, and maintains
a leading role in the defense of human rights. Korea has endorsed the OECD
and the G20 AI principles, and works in cooperation with other countries
on AI policy. While Korea has not yet expressed support for the Universal
Guidelines for AI, elements of the UGAI are reflected in the national AI
policies.
Ministry of Foreign Affairs, Human Rights Diplomacy,
http://www.mofa.go.kr/eng/wpge/m_5648/contents.do
1122
285
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Malaysia
National AI Strategy Overview
The Government of Malaysia does not have a National AI Strategy,
but has embraced AI-enabled technology as part of its future success,
nesting the development of its national AI strategies, including the National
AI Roadmap (AI-Rmap) under the Malaysian whole-of-government
strategy, known as the "Twelfth Malaysia Plan, 2021-2025” (“12 Plan”)
with the aim of “a prosperous, inclusive and sustainable Malaysia,”1123
which was introduced in September 2021.
According to 12 Plan, “national strategies on artificial intelligence
(AI) and blockchain will be developed to guide the growth of every sector
in the economy.” 12 Plan mentions that measures will be implemented to
strengthen guidelines and regulations on personal data protection and data
sharing with the aim of “ensuring data safeguards against cyber-attacks
and unethical uses.” According to the Plan, “an ethical framework and
standards on technology development, deployment and utilisation will also
be introduced to ensure responsible use of technology.”
Background and Related National Policies
In 2017 the Malaysian Government announced plans to develop a
National AI Framework complementary to its National Big Data Analytics
Framework.1124 Digitalization and AI objectives are covered under the
Malaysian Digital Economy Blueprint, operated in coordination with the
Economic Planning Unit (EPU) and various other plans, as illustrated
below.
● Malaysian Digital Economy Blueprint (2021-2031): The
Malaysian government, via its Economic Planning Unit (EPU),
recently launched its digital economy blueprint. MyDIGITAL is a
comprehensive 104-page document that lays the road map to
achieve the country's grand vision to become a regional leader in the
digital economy and attain an inclusive, responsible, and sustainable
socio-economic development, nationally. The intent is to “transform
Government of Malaysia, Twelfth Malaysia Plan (2021)
https://rmke12.epu.gov.my/bm
1124
OpenGOVAsia.com, Plans for cloud-first strategy and national AI framework
revealed at 29th MSC Malaysia Implementation Council Meeting (Oct. 28, 2017),
https://opengovasia.com/plans-for-cloud-first-strategy-and-national-ai-frameworkrevealed-at-29th-msc-malaysia-implementation-council-meeting/
1123
286
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Malaysia into a digitally-driven, high-income nation and a regional
leader in digital economy."1125
● National Fourth Industrial Revolution (2021): The Malaysian
government, via its Economic Planning Unit (EPU), also introduced
its National 4IR Policy, designed to “steer strategic socio-economic
transformation through the ethical use of 4IR policies.” Using a
whole-of-nation approach, the plan intends to: support balanced,
responsible, and sustainable growth for business; provide socioenvironmental well-being for all, and create a fit-for-future
government. Notably, the Policy draws on reference from the World
Economic Forum (WEF), OECD, UN, and the World Bank.1126
● National Big Data Analytics (BDA) Framework (2016-):
Introduced by the Malaysia Digital Economy Corporation (MDEC),
the BDA has four focus areas: architecting a data-driven culture;
identifying roles for a data-driven organization; turning information
into action; accessing tools and partners.1127
● In 2021, the Ministry of Science, Technology, and Innovation
(MOSTI) awarded selected researchers grants to formulate National
Artificial Intelligence (AI) Roadmap for Malaysia.1128
AI-Rmap
AI-Rmap has six overall strategies to execute: establishing AI
governance, advancing AI R&D, escalating a digital infrastructure to enable
AI, fostering AI talents, acculturating AI, enhancing quadruple helix, and
global collaboration.1129 While the strategy of the report does not overtly
address human rights, it does state that it needs to be “human-centric” in
design, and must incorporate the factors of being “explainable, transparent,
and ethical.”1130
Government of Malaysia, Malaysia Digital Blueprint Economy,
https://www.epu.gov.my/sites/default/files/2021-02/malaysia-digital-economyblueprint.pdf
1126
Government of Malaysia, National 4IR Policy (July 1, 2021),
https://www.epu.gov.my/sites/default/files/2021-07/4IR_Presentation.pdf
1127
Malaysia Digital Economy Corporation, Malaysia’s National Big Data Analytics
Initiative (Oct. 2016), https://calabarzon.neda.gov.ph/wp-content/uploads/2016/10/02Big-Data-Analytics-MDEC.pdf
1128
University of Technology, Malaysia, UTM Experts Entrusted by MOSTI to Develop
the National Artificial Intelligence (AI) Roadmap for Malaysia (Dec. 1, 2020),
https://news.utm.my/2020/12/utm-experts-entrusted-by-mosti-to-develop-the-nationalartificial-intelligence-ai-roadmap-for-malaysia/
1129
Id.
1130
Government of Malaysia, Malaysia’s AI Roadmap (Mar. 15, 2021),
https://airmap.my/st1/
1125
287
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI-Rmap has a distinct action plan through 2025, with benchmarks
set for horizon 1 (2021-2022), horizon 2 (2023-2024), and horizon 3 (2050
onwards). Within those benchmarks are four strategic initiatives, with
Strategic Initiative 1.4 being the establishment of an AI code of ethics.
Notably, in horizon 1 for this initiative, Malaysia will “study AI code of
ethics in international organizations and major countries” and “establish AI
ethical standards that are consistent with global norms.” Finalizing the AI
Code of Ethics is not due to be completed until 2023-2024.1131
OECD/G20 AI Principles
Malaysia has not endorsed OECD AI and G20 AI principles.
According to OECD AI Policy Observatory, Malaysia does not have any AI
initiatives.1132
Human Rights
Malaysia is a member of the United Nations and has endorsed the
Universal Declaration of Human Rights. In 2021, Freedom House scored
Malaysia at 51/100 (in 2020, 51) included in the group of partly free
countries. 1133 Freedom House raised concerns that laws, policies, and
practices do not guarantee equal treatment of various segments of the
population. On transparency, Freedom House noted, “the government was
initially open and transparent regarding Malaysia’s COVID-19 status and
the state’s response, however, observers faulted the government’s data
transparency as the late-year coronavirus wave accelerated.”
AI in healthcare
In 2017, the Ministry of Health launched the Malaysian Health Data
Warehouse (MyHDW) as part of a national healthcare information
gathering system. The system is designed to share a patient’s healthcare
records among all public health institutions, ensuring any doctor had full
access to medical records. The Malaysian government also initiated several
public-private collaborations, including the signing of a Memorandum of
Understanding (MoU) between Microsoft Malaysia and CREST
(Collaborative Research in Engineering, Science & Technology) to create a
first-of-its-kind digital health hub, and the creation of Malaysia’s largest
digital health platform, DoctorOnCall, which connects patients with an
Id.
OECD, OECD AI Policy Observatory, https://oecd.ai
1133
Freedom House, “Malaysia: Freedom in the World 2021 Malaysia: Freedom in the
World 2021 Country Report | Freedom House”
1131
1132
288
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
extensive network of over 1,500 Specialists in private hospitals and more
than 100 GP doctors throughout the country.1134
To combat the spread of COVID-19, the Malaysian government led
a multi-agency effort to create the MySejahtera app, which was linked to
the MyTrace app (developed by the Malaysian Ministry of Science,
Technology, and Innovation (MOSTI)) to “enables the identification of
those who have been in close proximity to an infected person using
Bluetooth technology.”1135 As of December 2020, the Malaysian
government cites that there are 24.5 million users of Mysejahtera, with a
strong push by the government for increased usage.1136
MOSTI also plans aggressive use of AI-enabled apps to detect those
who may have been in close contact with a COVID-19 patient, and a
geofencing app to enforce and monitor quarantines.” In May 2021, the
Malaysian government launched the Hotspot Identification for Dynamic
Engagement ("HIDE") system, as an early warning system to preemptively
identify COVID-19 hotspots using predictive technology, big data
analytics, and AI.1137 MOSTI is also planning to add “artificial intelligence
and big data analytics into the HIDE system to produce more accurate
predictions of coronavirus hotspots based on Bluetooth contact tracing.
HIDE currently uses MySejahtera check-in data.1138
Finally, in August 2021, the Malaysian government introduced a
new mobile app, the Vaccine Certificate Verifier app, to combat a rise in
the production and selling of fake certificates in the country.1139 While the
government has rolled out the use of AI-enabled technology to assist the
healthcare industry and COVID-29 response, there has been little
HealthcareITnews.com, An overview of Malaysia’s digital health landscape (July 1,
2020), https://www.healthcareitnews.com/news/asia/overview-malaysia-s-digital-healthlandscape
1135
HealthcareITnews.com, COVID-19: Malaysia’s pandemic approaches and its impact
on telehealth (June 08, 2020), https://www.healthcareitnews.com/news/asia/covid-19malaysia-s-pandemic-approaches-and-its-impact-telehealth
1136
Yahoo News, Health Ministry source: MySejahtera covers 24.5 million users with up
to 30,000 daily downloads despite misconceptions (Dec. 3, 2020),
https://malaysia.news.yahoo.com/health-ministry-source-mysejahtera-covers020809774.html
1137
Mondaq.com, Malaysia: Covid-19: Nowhere To HIDE? (May 24, 2021)
https://www.mondaq.com/government-measures/1071572/covid-19-nowhere-to-hide
1138
Code Blue, Government Plans Bluetooth Covid-19 Contact Tracing, Geofencing SelfQuarantine App (Nov 2, 2021), https://codeblue.galencentre.org/2021/11/02/governmentplans-bluetooth-covid-19-contact-tracing-geofencing-self-quarantine-app/
1139
HealthcareITnews.com, Malaysia launches vaccine certificate verification app (Aug.
24, 2021), https://www.healthcareitnews.com/news/asia/malaysia-launches-vaccinecertificate-verification-app
1134
289
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
communication from the government on a focus on human rights, data
privacy, and algorithmic transparency.
AI System for Surveillance
In 2018, Auxiliary Force Sdn Bhd (AFSB), a member of Royal
Malaysia Police Cooperative Bhd., became the first Malaysian security
force in the country to integrate body-worn cameras with facial recognition
technology.1140 In 2019, the state of Penang launched the first public facial
recognition to help police combat crime. The system: uses AI to identify
faces captured by the CCTV network operated by Penang Island City
Council (MBPP)”, with a plan to have hundreds of new cameras installed
at major roads, intersections, and hotspots for criminal activity.1141 The
Malaysian business sector has also integrated the use of facial recognition
across the nation, including facial recognition “check-ins” at events1142, and
the Malaysian Ministry of Transport introduced a public-private facial
recognition program in 2019, allowing users to take selfies of themselves
for recognition.1143 While there is an increase in facial recognition
technology in Malaysia, there is no overarching AI Policy focused on
human rights and democratic values governing the use of this technology.
Data Protection Laws
Data protection in Malaysia stems from the Personal Data Protection
Act of 2010 (PDPA),1144, =its first comprehensive personal data protection
legislation, by the Malaysian Parliament on June 2, 2010, and came into
force on November 15, 2013. The PDPA seeks to safeguard personal data
and confer certain rights to users regarding personal data. In addition to the
OpenGOVAsia.com, Auxiliary Force of Malaysian Police integrates facial
recognition technology with body-worn cameras (Apr. 16, 2018),
https://opengovasia.com/auxiliary-force-of-malaysian-police-integrates-facialrecognition-technology-with-body-worn-cameras/
1141
Biometricupdate.com, “Malaysian state launches facial recognition to CCTV
network” (Jan 3, 2019), https://www.biometricupdate.com/201901/malaysian-statelaunches-facial-recognition-to-cctv-network
1142
The Malaysian Reserve, Facial recognition tech grows amid concerns” (Dec. 5,
2019), https://themalaysianreserve.com/2019/12/05/facial-recognition-tech-grows-amidconcerns/?__cf_chl_jschl_tk__=DAhg_QRUGomGzvz0vzL53vheCQUq1lvw25Aloim.r
Tw-1636370158-0-gaNycGzNClE
1143
Grab.com, Grab partners with Ministry of Transport to implement facial recognition
technology in Malaysia (Apr 11, 2019), https://www.grab.com/my/press/social-impactsafety/grab-mot-facial-recognition-technology/
1144
Malaysia Ministry of Communications and Multimedia, Personal Data Protection
Act 2010 (June 2010),
https://www.kkmm.gov.my/pdf/Personal%20Data%20Protection%20Act%202010.pdf
1140
290
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
PDPA, five pieces of subsidiary legislation were also enforced on
November 15, 2013.
Additional legislation passed to date include the Personal Data
Protection Regulations 2013 ('the 2013 Regulations'); the Personal Data
Protection (Class of Data Users) Order 2013 ('the Order'); the Personal Data
Protection (Registration of Data User) Regulations 2013 ('Registration
Regulation'); the Personal Data Protection (Fees) Regulations 2013; the
Personal Data Protection (Compounding of Offences) Regulations 2016
('Compounding of Offences Regulations'); the Personal Data Protection
(Class of Data Users) (Amendment) Order 2016 ('the Order Amendment');
and the Personal Data Protection (Appeal Tribunal) Regulations 2021.1145
The PDPA imposes strict requirements on any person who collects
or processes personal data (data users) and grants individual rights to 'data
subjects', and is enforced by the Commissioner of the Department of
Personal Data Protection (the Commissioner). It is observed that the PDPA
is similar in sense to the Data Protection Directive 95/46/EC of the
European Union (EU), leading the PDPA to be described as European-style
privacy law.1146
Issues with the PDPA are that it does not apply to federal and state
governments,1147 exempts the processing of information by a credit
reporting agency and the PDPA does not constrain government access to
data. part of an ongoing review of the PDPA, the Personal Data Protection
Commissioner of the Ministry of Communications and Multimedia
Malaysia has issued Public Consultation Paper No. 01/2020 – Review of
Personal Data Protection Act 2010 (PC01/2020) dated February 14, 2020,
to seek the views and comments of the public on 22 issues set out in
PC01/2020.
Autonomous vehicles
According to the Malaysian Investment Development Authority
(“MIDA”), Malaysia supports efforts to become a regional leader in
manufacturing, engineering, and technological innovation. To this end, the
National Automotive Policy 2020 (NAP 2020) aims to develop the
Malaysian auto industry “through research and development of new
technologies, especially in the areas of Next-Generation Vehicles (NxGV),
DataGuidance.com, “Personal Data Protection Act 2010”, Jun 2021,
https://www.dataguidance.com/notes/malaysia-data-protection-overview
1146
The Law Reviews, The Privacy, Data Protection and Cybersecurity Law Review:
Malaysia (Nov 5, 2021), https://thelawreviews.co.uk/title/the-privacy-data-protectionand-cybersecurity-law-review/malaysia
1147
Id.
1145
291
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Industrial Revolution 4.0 (IR 4.0) and mobility -as a service (MaaS)”.1148
According to the MIDA, the first autonomous vehicle (AV) testing route of
Malaysia, which will allow tech companies to test the capabilities of their
vehicles, has been approved by the Ministry of Transport of Malaysia and
the Sepang Municipal Council.1149
Besides, as mentioned by the Government of Malaysia on its official
website, AI will be a key technology in the development of autonomous or
self-driving vehicles. The government of Malaysia mentions that “not only
is AI capable of collecting and analysing data through sensors and cameras
but it is also capable of adapting to situations and learning through machine
learning.”1150
Evaluation
Malaysia has rapidly introduced technology into its public and
private sectors and has constructed numerous digital policies and taken
action to map out and support its digital economy and support business
growth. While there is no national AI framework currently, the AI-Rmap is
ongoing and part of the overall 12 Plan. In October 2021, the Malaysian
Ministry of Science, Technology, and Innovation (Mosti) announced it is
developing 17 technology roadmaps in line with the 10 science and
technology drivers, and the 10 socio-economic drivers in the 10-10
Malaysian Science, Technology, Innovation, and Economy (MySTIE)
Framework. This includes “artificial intelligence (AI), blockchain,
electrical and electronics (E&E), advanced materials, robotics, and
vaccines.”1151 Malaysia has rolled out facial recognition in its public sector,
but the absence of an overall strategy for use and guidelines for humanbased values is troubling. Finally, Malaysia has rolled out several public
and private sector collaborative efforts to support the healthcare industry
and COVID-19 response but has provided little insight on efforts to ensure
data privacy, human rights, and algorithmic transparency.
1148
Malaysia Industrial Development Authority, Malaysia steering towards autonomous
vehicle technology, https://www.mida.gov.my/malaysia-steering-towards-autonomousvehicle-technology/
1149
Id.
1150
Government of Malaysia, Capacities for Digital Transformation, Autonomous Car,
https://www.malaysia.gov.my/portal/content/30713
1151
The Malaysian Reserve, Dr Adham: 17 tech roadmaps in pipeline (Oct. 18, 2021),
https://themalaysianreserve.com/2021/10/18/dr-adham-17-tech-roadmaps-in-the-pipeline/
292
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Mexico
National AI Strategy
In 2018, Mexico became the first Latin American country to
announce a national AI strategy. Towards an AI Strategy in Mexico:
Harnessing the AI Revolution was commissioned by the United Kingdom’s
Embassy in Mexico City in collaboration with the Office of the Mexican
President under the Peña Nieto administration.
The AI Strategy report provides a preliminary outline of how
Mexico should become a leader in AI. Mexico’s manufacturing-centric
economy, the report argues, faces grave risks amid growing automation and
should take a clear, strategic position in developing AI. The AI Strategy sets
out six thematic areas: governance, government, and public services;
research and development; capacity, skills, and education; data
infrastructure; and ethics and regulation. Within the category of ethics and
regulation, the report recommends that the Mexican government bring data
assets inside the scope of Mexican competition law (COFECE) in
recognition of the fact that data is a competitive asset. The report also called
for the creation of an AI Ethics Council which would “set guidelines and
limits which reflect Mexican Values” and “award a quality mark for AI
companies who abide by the standards.” 1152
The AI Strategy report set out five key actions for the Mexican
government: develop an inclusive governance framework; identify the
needs of AI in industry; open the recommendations of the Policy Report for
public consultation; support Mexico’s AI leadership in international
forums; and promote continuity through changing administrations, by
working with all interested stakeholders towards an official AI National
Policy.1153
Implementation of the National AI Strategy
The AI strategy for Mexico was initially published in 2018 under
the former Peña Nieto presidential administration, which ended that same
year. The current government administration under President López
Obrador (2018-2024) has neither implemented the OECD AI principles nor
any of the goals originally outlined in the AI strategy of 2018. There are no
current government metrics or identifiable progress made on the
achievement of AI policies under the current government administration.
Oxford Insights, Towards an AI Strategy in Mexico: Harnessing the AI Revolution
(June 2018) https://www.oxfordinsights.com/mexico
1153
CAF- Development Bank of Latin America, Mexico: the story and lessons behind
Latin America’s first AI strategy (June 2020) https://www.cminds.co/reports
1152
293
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
No recent information regarding AI is available on any official website of
the federal government. The only publicly available information from the
government is a blog post announcing the publication of the National
Strategy posted during the previous administration.1154 The federal
government has reoriented its policy priorities away from AI and has cut
the funding allotted to the implementation of the National Strategy.
Public Participation
The effort to maintain momentum for increased AI development and
policymaking is spearheaded by Coalition IA2030Mx, a multisectoral
coalition made up of professionals, academic institutions, companies,
startups, public agencies and other key actors of the digital ecosystem and
Artificial Intelligence in Mexico.1155 The IA2030Mx goals include (1) the
participation of all member states of Mexico, (2) the creation of a National
AI agenda for 2030, and (3) the promotion of the OECD AI Principles. The
IA2030Mx said “The members of this movement have been working since
the beginning of 2018 under a philosophy of co-responsibility of
government, academia, industry and civil society, seeking that Mexico does
not lag behind in the 4th Industrial Revolution, strategically take advantage
of the benefits of AI and mitigate the possible ethical and social risks.”
In 2019, the IA2030 coalition conducted a mass survey to determine
the major areas of concern in Mexico regarding AI.1156 This knowledge was
then mobilized in the creation of a 2020 National Agenda for AI. The central
themes of the agenda are data, digital infrastructure, and cybersecurity;
ethics; governance, government, and public services; capabilities and
education; and the collaboration of Mexicans outside of the Republic. This
agenda was created with the input of over 400 different actors but had no
collaboration by the federal government. Different levels of government,
like the state of Jalisco and some Senatorial committees, have expressed
interest in AI governance and have participated in projects with C Minds on
the matter; however, non-state actors have been the major participants in
shaping the future of AI in Mexico.
In December 2021, Centro LATAM Digital and Iniciativa
Latinoamericana por los Datos Abiertos (ILDA) with the financial support
Enrique Zapata, Estrategia de Inteligencia Artificial MX 2018 (Mar. 18, 2018)
https://datos.gob.mx/blog/estrategia-de-inteligencia-artificial-mx-2018
1155
IA2030Mx, Artificial Intelligence in Mexico: A National Agenda (Nov. 2020)
(English translation), https://www.ia2030.mx/
1156
IA2030Mx, Artificial Intelligence in Mexico: A National Agenda (Nov. 2020)
(English translation), https://www.ia2030.mx/
1154
294
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of the International Development Research Centre (IDRC) and the InterAmerican Development Bank (IADB) published a report on public policy
on AI of Mexico. The main purpose of this report is to identify the main
challenges and obstacles for the design of public policies on AI that includes
a human rights-based approach and that may serve to support and resume
Mexico’s national AI strategy by the government of AMLO. The report
contains a section of general conclusion and recommendations. Among the
main recommendation of the report are (i) to create a national strategy on
AI that is multi-stakeholder and multi-disciplinary that could contribute to
the development of public policies and mechanism for measurement and
assessment; (ii) put a strong emphasis on the protection of personal data
related with the development of AI technologies; (iii) implement the OECD
principles on AI; (iv) include efficient government coordination
mechanisms to implement a national plan on AI that could identify the main
actors and stakeholders of different areas; and (v) to work closely with the
different commissions and groups of National Congress to make them
aware of the benefits and risks of AI that may support the drafting of flexible
regulations on AI.1157
Global Partnerships: OECD, G20, GPAI, and COMEST
Mexico has taken an active role in pursuing international
cooperation for the ethical development of AI. The Mexican government
endorsed the OECD AI Principles in 2019 as well as the G20’s and is one
of the founding members of the Global Partnership on AI (GPAI), the
world’s first international AI initiative.1158 Mexico is also represented in
UNESCO’s World Commission on the Ethics of Scientific Knowledge and
Technology (COMEST). COMEST produced a preliminary study on the
ethics of AI, which has now become the foundation of UNESCO’s
Recommendation on the Ethics of AI which will be elaborated between
2019 and 2021.1159
empatIA, Reportes de política publica: México (Dec. 2021),
https://secureservercdn.net/192.169.220.85/dxc.177.myftpupload.com/wpcontent/uploads/2021/12/Policy-report-Mexico-version-final.pdf
1158
OECD, OECD to host Secretariat of new Global Partnership on Artificial
Intelligence (June 15, 2020) https://www.oecd.org/going-digital/ai/oecd-to-hostsecretariat-of-new-global-partnership-on-artificial-intelligence.htm; Gobierno de México,
Declaración Conjunta de los miembros fundadores de la Alianza Global sobre la
Inteligencia Artificial (June 15, 2020) https://www.gob.mx/sre/prensa/declaracionconjunta-de-los-miembros-fundadores-de-la-alianza-global-sobre-la-inteligencia-artificial
1159
UNESCO, Elaboration of a Recommendation on the ethics of artificial intelligence
(2020) https://en.unesco.org/artificial-intelligence/ethics#recommendation, COMEST:
https://en.unesco.org/themes/ethics-science-and-technology/comest/members
1157
295
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection
The National Institute for Transparency, Access to Information and
Personal Data Protection (INAI) was one of the national DPA’s that
sponsored the Global Privacy Assembly’s Resolution on Accountability in
the Development and Use of Artificial Intelligence of October 2020.
The Ibero-American Data Protection Network (RIPD),1160 a group
of experts on data protection and access to information currently, is
composed of 34 entities, including 14 federal and state access to information
and data protection authorities of Latin America and countries of the
Caribbean, which includes Mexico’s INAI. In 2019, the RIPD adopted
“Specific Guidelines for Complying with the Principles and Rights
Governing Personal Data Protection in Artificial Intelligence Projects.”1161
The AI Guidelines provide a common framework for the entities in the
RIPD. The AI Guidelines are based on the Standards for Personal Data
Protection for the Iberoamerican States approved in 2017.1162 The AI
Guidelines provide recommendations for the processing of personal data for
AI systems.
Although, there is now the RIPD reference framework for the
processing of personal data for AI systems, the INAI has not yet developed
national policies for the protection of personal data in AI systems. INAI is
currently part of an ongoing initiative sponsored by Facebook, C-Minds, the
Interamerican Development Bank (IDB) and the BID LAB, which will
gather a number of companies that currently use AI in their products and
services across Mexico. The main purpose of this initiative is to facilitate
and test public policies for the governance of AI systems and provide for
transparency and accountability practices for data protection during 2020
and 2021. The outcome of this initiative will be a report with public policy
recommendations for INAI and other data protection agencies in Latin
America. Likewise, the report will serve as a basis for the development of a
Framework and Manual of T&E of AI Systems for Mexico and will be
1160
The Ibero-American Data Protection Network (Red Iberoamericana de Protección de
Datos (RIPD), https://www.redipd.org/
1161
Red Iberoamericana de Protección de Datos (RIPD), Orientaciones Específicas para
el Cumplimiento de los Principios y Derechos que Rigen la Protección de los Datos
Personales en los Proyectos de Inteligencia Artificial (June 21, 2019),
http://inicio.inai.org.mx/nuevo/RIPD_orientaciones_especificas_de_proteccion_de_datos
_en_ia.pdf
1162
Red Iberoamericana de Protección de Datos (RIPD), Estándares de Protección de
Datos Personales para los Estados Iberoamericanos. (June 20, 2017),
https://www.redipd.org/sites/default/files/inlinefiles/Estandares_Esp_Con_logo_RIPD.pd
f
296
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
presented by the INAI in collaboration with the PPP of Mexico, C-Minds
and the IDB Group during 2021.1163
Algorithmic Transparency
In November 2018, the Mexican government published general
principles for the development and use of systems based on AI in Mexico’s
federal government.1164 The key principles are:
• Monitor and evaluate the impacts of AI systems in order to
ensure that they achieve the expected results
• Promote transparency, by explaining to the users that interact
with AI systems the decision process taken by such systems,
the expected benefits as well as the potential risks derived from
using such systems
• Protect privacy, by incorporating mechanisms of control and
consent for the use of personal data during the design of AI
systems
• Foster equality, by reducing risks of discriminatory biases
derived from the utilized data
• Due process, by allowing individuals to dispute decisions made
by AI systems.
The US Library of Congress noted this summer that “the presidential
administration that adopted this strategy and its guiding principles ended on
November 30, 2018. No information could be located on whether the new
administration (which commenced on December 1, 2018) will continue
with this strategy and its principles or initiate a similar effort.”1165
Human Rights
Freedom House gives Mexico a “partly free” (62/100) rating for
political rights and civil liberties.1166 According to Freedom House,
“Mexico has been an electoral democracy since 2000, and alternation in
power between parties is routine at both the federal and state levels.
CMINDS, Prototipo de Políticas Públicas. Transparencia y explicabilidad de
sistemas de IA, https://www.cminds.co/prototipo-politica-ia
1164
Principles for the Administraion of Artificial Intelligence (Nov. 2018),
https://www.gob.mx/cms/uploads/attachment/file/415644/Consolidado_Comentarios_Co
nsulta_IA__1_.pdf
1165
US Library of Congress, Regulation of Artificial Intelligence: The Americas and the
Caribbean (July 24, 2020), https://www.loc.gov/law/help/artificialintelligence/americas.php
1166
Freedom House, Freedom in the World 2021 – Mexico (2021),
https://freedomhouse.org/country/mexico/freedom-world/2021
1163
297
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
However, the country suffers from severe rule of law deficits that limit full
citizen enjoyment of political rights and civil liberties.”
Evaluation
Mexico launched an ambitious strategy for AI in 2018 and identified
key ethical considerations for the deployment of AI systems. Mexico also
endorsed the OECD and the G20 AI Principles, is a founding member of
the GPAI, and worked with UNESCO on the development of
recommendations for AI. But there has been little activity on AI policy since
2018. Consistent with its international commitments, Mexico should
develop the legal frameworks necessary for AI oversight prior to the
deployment of AI systems.
298
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Netherlands
Strategic Action Plan for AI
In September 2019, the Dutch government set out The Strategic
Action Plan for Artificial Intelligence.1167 The AI Plan sets out three broad
themes – “Capitalising on societal and economic opportunities,” “Creating
the right conditions,” and “Strengthening the Foundations.” The key goals
are public-private partnerships, international cooperation, an “inclusive
approach that puts people first,” and “a country that is at the forefront of AI
applications which serve the interests of people and society.” Under this
plan, government commits to protect public values and human rights,
further defined as prohibition of discrimination, protection of privacy,
freedom of speech, human dignity and autonomy, the right to a fair trial,
and human rights.
The Dutch AI strategy follows the Dutch Digitalization Strategy
(2018), the first Cabinet-wide effort to formulate key priorities for
digitalization, data and AI. Within the Digitalization Strategy the
government “supports and endorses the guidelines established in the EU’s
recent communication on ‘Ethics guidelines for trustworthy AI.”1168 The
government also commits to creating a “responsible innovation toolbox
(including impact assessments, handbooks and guidelines)” and making
knowledge available in the areas of transparency, explainability and
accountability. Through the Transparency Lab initiative, the “government
is working with businesses and supervisory bodies to assess how algorithms
and their practical applications can be made more transparent and
verifiable.” The government seeks to “ensure that as many Dutch
companies and public organizations as possible actively participate in the
pilot phase of the High-Level Expert Group’s ethical guidelines for AI.”
The Government has published an updated Dutch Digitalization Strategy
2021.1169 The ministries that coordinate the efforts are advised by the
Digital Netherlands Council, which joins the deliberation process on
digitalisation policy and consists of domain experts. The Strategy commits
The Strategic Action Plan for Artificial Intelligence (2019),
https://www.government.nl/binaries/government/documents/reports/2019/10/09/strategicaction-plan-for-artificialintelligence/Strategic+Action+Plan+for+Artificial+Intelligence.pdf
1168
Dutch Digitalization Strategy (2018),
https://www.nederlanddigitaal.nl/documenten/publicaties/2019/11/13/english-version-ofthe-dutch-digitalisation-strategy-2.0
1169
Dutch Digitalization Strategy (2021), https://www.nederlanddigitaal.nl/english/thedutch-digitalisation-strategy-2021
1167
299
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the Netherlands to European and international cooperation within
appropriate ethical and legal frameworks.
Across the Dutch documents reviewed, the concept of Fairness is
mentioned only in reference to GDPR & AI High-Level Expert Group
Ethical Guidelines. Rule of Law is mentioned in reference to AI developed
within legal and ethical frameworks and the work of The Minister for Legal
Protection. Fundamental Rights are defined as “privacy, non-discrimination
and autonomy.” In addition to these goals, Accountability and Transparency
appear in multiple times in all government documents.
The Netherlands Organisation for Science Research (NWO) and the
Ministry of Economic Affairs & Climate Policy jointly developed a publicprivate research programme that funds research on explainable, socially
aware and responsible AI.1170 The Special Interest Group of AI, SIGAI,
representing all computing science academic institutes and researchers in
the Netherlands that perform AI research also published Dutch Artificial
Intelligence Manifesto highlighting the importance of socially aware,
explainable and responsible AI.1171
The public administration within the Kingdom is layered and
sometimes disconnected. However, the government organizations and the
Association of the Netherlands Municipalities (VNG), the Ministry of the
Interior and Kingdom Relations commit to focusing on “ethics in, by and
for design and the transparency of algorithms when government
experiments with AI for public tasks.” A study commissioned by Ministry
of Interior proposed AI guidelines that apply to the public and private
sectors. The guideline, prepared by researcher at VUB, Tilburg University,
Eindhoven University of Technology and the National Human Rights
Institute of the Netherlands, is a result of Dutch parliament stating that
“racism must be ended as soon as possible, not least by stopping the use of
discriminatory algorithms.”1172
NWO, First national research agenda for Artificial Intelligence (Nov. 21, 2019),
https://www.nwo.nl/en/news-and-events/news/2019/11/first-national-research-agendafor-artificial-intelligence.html
1171
Dutch Artificial Intelligence Manifesto (2008), http://bnvki.org/wpcontent/uploads/2018/09/Dutch-AI-Manifesto.pdf
1172
VUB Today, New guidelines aim to correct discriminatory algorithms: VUB
researchers help to create AI rules for government organisations and companies (July
15, 2021), https://today.vub.be/en/article/new-guidelines-aim-to-correct-discriminatoryalgorithms
1170
300
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In 2020 the Dutch Data Protection Authority (Dutch DPA) approved
the first ‘code of conduct’ in the Netherlands, the Data Pro Code1173 drafted
by NL Digital, the Dutch industry association for organizations in the ICT
sector. In 2021, researchers from VUB, Tilburg University, Eindhoven
University of Technology and the National Human Rights Institute of the
Netherlands developed AI guidelines that apply to the public and private
sectors. The Guideline is a result of Dutch parliament stating that “racism
must be ended as soon as possible, not least by stopping the use of
discriminatory algorithms” and was funded by the Ministry of the
Interior.1174
Predictive Policing
Dutch Police, in collaboration with Utrecht University and the
University of Amsterdam, established the National Police Lab AI to
develop “AI techniques to improve the safety in the Netherlands in a
socially, legally and ethically responsible way.”1175 In alignment with the
government’s commitment to experiment with technology to solve social
issues, Dutch Police has launched pilot projects with predictive policing to
anticipate and prevent crime that might be committed by a certain person or
at a certain location.
The first is the Sensing Project in Roermond where police uses
cameras and other sensors to systematically monitor all people driving in
and around Roermond and create a risk score, effectively transforming the
“city into a living lab where every person travelling by car is subjected to
mass surveillance and other human rights violations”.1176 The project
violates the principles of human rights, informed consent, right to privacy
and data protection, right to due process and non-discrimination. Amnesty
International calls on the Dutch government to “halt the Sensing project and
comparable ‘experimental’ predictive policing projects” and to “implement
a mandatory and binding human rights impact assessment requirement
applicable to the public sector.” 8
Wanbound BV, Data Processing Agreement (Apr. 2018),
https://www.wanbound.com/wp-content/uploads/2018/05/Nederland-ICT-Dataprocessing-agreement-UK-Part-2.pdf
1174
VUB Today, New guidelines aim to correct discriminatory algorithms (July 15,
2021), https://today.vub.be/en/article/new-guidelines-aim-to-correct-discriminatoryalgorithms
1175
Innovation Center for Artificial Intelligence, Police Lab AI, https://icai.ai/police-labai/
1176
Amnesty International, We Sense Trouble (2020),
https://www.amnesty.org/download/Documents/EUR3529712020ENGLISH.PDF
1173
301
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Second predictive policing project is Criminaliteits Anticipatie
Systeem (Crime Anticipation System or CAS) implemented nationwide in
2017. The use of CAS to predict crime locations makes the Netherlands the
first country in the world to deploy predictive policing on a national
scale.1177 To date, none of the systems in use by Dutch police have been
subjected to a comprehensive human rights evaluation.
Automated Profiling Fiasco
In early 2020, a Dutch court ruled that the System Risk Indication
algorithm (SyRI) algorithm, used to combat fraud in government programs,
violated Article 8 of the European Convention on Human Rights.1178 In the
landmark decision, the Court ruled that the principle of transparency was
not observed, because there is no insight into the risk indicators and the
operation of the risk model. The Court also advised that there is a risk that
inadvertent links are established with the use of SyRI on the basis of bias,
such as a lower socio-economic status or an immigration background.
The UN Special Rapporteur on Extreme Poverty and Human Rights
explained that the decision challenged the “systematic, legislatively
sanctioned, used of digital technologies in welfare state on human rights
ground.”1179 In a filing with the court, the Special Rapporteur voiced
concern that “SyRI has consistently been rolled out in poorer and more
vulnerable areas of municipalities”, and that the Dutch government has
denied access to information about the data and ‘risk models’ used in the
algorithm.” The Special Rapporteur called the decision, “a clear victory for
all those who are justifiably concerned about the serious threats digital
welfare systems pose for human rights.”1180 In April 2020, Data Processing
by Partnerships Act was introduced by the government. Where SyRI was
1177
Strikwerda, Litska (Aug. 2020), “Predictive Policing: The Risks Associated with Risk
Assessment.” The Police Journal. https://doi.org/10.1177/0032258X20947749.
1178
Europe Limits Government by Algorithm. The US, Not So Much, Wired (Feb. 7,
2020), https://www.wired.com/story/europe-limits-government-algorithm-us-not-much/
1179
UN HROHC, Brief by the UN Special Rapporteur on extreme poverty and human
rights as Amicus Curiae in the case of NJCM c.s./De Staat der Nederlanden (SyRI)
before the District Court of The Hague (2019),
https://www.ohchr.org/Documents/Issues/Poverty/Amicusfinalversionsigned.pdf
1180
UN HROHC, Landmark ruling by Dutch court stops government attempts to spy on
the poor – UN expert (Feb. 5, 2020),
https://www.ohchr.org/en/NewsEvents/Pages/DisplayNews.aspx?NewsID=25522&LangI
D=E
302
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
related to public data sharing, this bill expands the data surveillance and
sharing to all data stored in public and private storage.1181
A government report states that “Human dignity and security are not
elements of the Dutch Constitution but are referred to in international
treaties such as the Charter of Fundamental Rights of the European Union
and the Universal Declaration of Human Rights.” The same report warned
that there is a clear risk that “fundamental rights and ethics in the digital
age” as the final section in the Dutch National Digitalisation Strategy, could
literally be an afterthought in the digitalisation agenda.1182
In January 2021, the Dutch government resigned after it became
clear that thousands of families, disproportionately of ethnic minority
backgrounds, were wrongly accused of child welfare fraud by a
discrimnatory algorithm and told to pay money back.1183
AI Registry
In September 2020, Amsterdam launched an AI registry in beta
version to detail how city government uses algorithms to deliver services.
“Each algorithm cited in the registry lists datasets used to train a model, a
description of how an algorithm is used, how humans utilize the prediction,
and how algorithms were assessed for potential bias or risks. The registry
also provides citizens a way to give feedback on algorithms their local
government uses and the name, city department, and contact information
for the person responsible for the responsible deployment of a particular
algorithm.”1184
Public Participation
The Electronic Announcement Act requires national governments to
publish official publications on the internet rather than on paper.1185 All AI
AlgorithmWatch, Automating Society Report 2020,
https://automatingsociety.algorithmwatch.org/wp-content/uploads/2020/10/AutomatingSociety-Report-2020.pdf
1182
Kool, L., E. Dujso, and R. van Est (2018). Directed digitalisation – Working towards
a digital transition focused on people and values – The Dutch approach. The Hague:
Rathenau Instituut.
1183
Gabriel Geiger, How a Discriminatory Algorithm Wrongly Accused Thousands of
Families of Fraud, Vice (March 1, 2021) https://www.vice.com/en/article/jgq35d/how-adiscriminatory-algorithm-wrongly-accused-thousands-of-families-of-fraud
1184
Khari Johnson, Amsterdam and Helsinki launch algorithm registries to bring
transparency to public deployments of AI, VentureBeat (Sept. 28, 2020),
https://venturebeat.com/2020/09/28/amsterdam-and-helsinki-launch-algorithm-registriesto-bring-transparency-to-public-deployments-of-ai/
1185
European Commission, Digital Government Factsheet 2019 – Netherlands (2019),
https://joinup.ec.europa.eu/sites/default/files/inlinefiles/Digital_Government_Factsheets_Netherlands_2019_0.pdf
1181
303
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
policies are accessible by public via the websites of The Dutch Parliament
and the Digital Government.1186 The government also commits to
improving the basic digital skills of all citizens and ensuring the
accessibility of government services and information and organizing
assistance for those with less digital skills.
Participation in the development of Dutch digitalization plan and
strategic action plan is geared more towards public agencies, private
companies, universities and research institutes than the citizens directly.
Taskforce AI that created the initial AI report is a public-private partnership,
and its new initiative “Dutch AI Coalition (NL AIC) is a cooperation
between the different research centers. 1187 National Innovation Centre for
AI (ICAI) is also a national network between knowledge institutions,
industry and government.1188
Fundamental Rights and OECD AI Principles
The Netherlands has endorsed Universal Declaration of Human
Rights, The European Union (EU) Charter of Fundamental Rights and The
European Convention on Human Rights (ECHR).1189 However, there are
differences in the legislative and institutional frameworks in the four
constituent countries and The Office of the High Commissioner for Human
Rights (OHCHR) recommends that the State party “establish a national
human rights institution in Aruba, Curaçao and Sint Maarten.”1190 OHCHR
is also “concerned that the anti-discrimination provisions of the
Netherlands, including the Equal Treatment Act 1994, do not prohibit
discrimination based on all grounds, including color, language, social
origin, property, birth or other status.”18
The Netherlands has endorsed the OECD AI Principles. “The
Netherlands is following the European approach to responsible AI and
wants European values and standards to be embedded in AI applications at
1186
Netherlands, House of Representatives, https://www.houseofrepresentatives.nl;
Netherlands, Digital Government Agenda, https://www.nldigitalgovernment.nl/digitalgovernment-agenda/
1187
HSD Foundation, New Dutch AI Coalition Demands National Approach (July 23,
2019), https://www.thehaguesecuritydelta.com/news/newsitem/1329-dutch-ai-coalitiondemands-national-approach
1188
Innovation Center for Artificial Intelligence, https://icai.ai/
1189
The European Union (EU) Charter of Fundamental Rights in the Netherlands
https://fra.europa.eu/sites/default/files/fra_uploads/fra-2019-eu-charter-innetherlands_en.pdf
1190
UN HROHC, UN Treaty Body Database,
https://tbinternet.ohchr.org/_layouts/15/treatybodyexternal/Download.aspx?symbolno=C
CPR/C/NLD/CO/5&Lang=En
304
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
an early stage (in the design and development phase.”1191 Netherlands
joined GPAI in December 2020. 1192
The Dutch government agrees with the conclusions of the Joint
Committee of the Advisory Council on International Affairs (AIV) and the
Advisory Committee on Issues of Public International Law (CAVV)
advisory report that meaningful human control is required in the
deployment of autonomous weapon systems and responsibility and
accountability attribution needs to be taken into account in the design stage
of weapon systems.1193 Government also views that a moratorium on fully
autonomous weapon systems to be currently unfeasible.
In October 2021, the Netherlands Scientific Council for
Government Policy recommended the development of appropriate
regulatory frameworks that safeguard fundamental rights and values in the
long-term use regulation to actively steer developments of surveillance and
data collection, the concentration of power, and the widening gap between
the public and private sector in the digital domain.1194
Algorithmic Transparency
The Netherlands is subject to GDPR, and the government advocates
that a European regulator should be able to “impose ex-ante obligations on
large digital platforms with a gatekeeper role.”1195 In alignment with GDPR
requirements, the Dutch Data Protection Authority (Dutch DPA) is
established. Dutch DPA advised that it is concerned with lack of
transparency and poor data security practices in the public sector, policing,
criminal justice, and that digital government will be one of its three core
focus areas for 2020-23.1196
Despite all the actions taken already in the Netherlands, there are
also several concerns about these actions and how the government proceeds
The Strategic Action Plan for Artificial Intelligence (2019)
Global Partnership on AI. https://gpai.ai/.
1193
Advisory Council on International Affairs Government response to AIV/CAVV
advisory report no. 97, Autonomous weapon systems: the need for meaningful human
control (Mar. 2, 2016), https://perma.cc/J37M-UQ33
1191
1192
Netherlands Scientific Council for Government Policy (wrr). Mission ai – The New
System Technology (October 2021);
https://english.wrr.nl/publications/reports/2021/11/11/summary-mission-ai
1195
Government of Netherlands, Dutch position on competition policy in relation to
online “platforms” (Nov. 10, 2019),
https://www.government.nl/documents/publications/2019/10/11/dutch-position-oncompetition-policy
1196
Dutch Data Protection Authority, Focus Dutch Data Protection Authority 2020-2023,
https://autoriteitpersoonsgegevens.nl/sites/default/files/atoms/files/apdataprotectie_in_een_digitale_samenleving_-gb_wtk.pdf
1194
305
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
with its priority goals. The findings in the report of the Temporary
Committee on the Digital Future (TCDT) of the Dutch House of
Representatives (published in May 2020) concludes that the House of
Representatives has only discussed digitisation in a fragmented way; not all
of the risks, opportunities and public values have been addressed; and many
laws and rules in the field of digitisation are determined in the European
Union.1197 The TCDT proposes that a standing committee for Digital
Affairs be established after the elections. The government has set up an ‘AI,
public values and human rights’ knowledge platform where ministries
exchange knowledge and develop policy on public values and human rights
in AI applications.1198
Although Dutch government publicly commits to human rights and
OECD AI Principles, the eagerness of the Kingdom to combine and then
share all its data, as well its enthusiasm for techno-solutionism especially in
the fields of justice administration and law enforcement are reasons for
concern. The Netherlands government has strong commitment to advance
the AI capabilities and applications. The main public-private partnership,
Dutch AI Coalition (NL AIC), developed the AiNed programme: an agenda
for the period 2020–2027 to position the Netherlands internationally as the
best testing ground for the introduction of responsible AI application.. The
Netherlands government is creating the tools and space required in
collaboration with private companies and knowledge institutes. However,
the citizens who are impacted by the government’s data-sharing practices
and experiments in social arena are not meaningfully included in the
discussions. The reality of cases like SyRI and CAS clashes with the public
commitment to ethical AI principles in strategy documents. The
Netherlands has not endorsed the Universal Guidelines for AI,1199 or the
GPA Resolution on AI Accountability.1200 The Netherlands Court of Audit
1197
House of Representatives, The temporary committee on the Digital Future (TCDT),
Summary of the report Update required. Towards greater parliamentary control of
digitisation, https://www.houseofrepresentatives.nl/membersparliament/committees/temporary-committee-digital-future/summary-report-updaterequired
1198
Artificial Intelligence and Public Values;
https://kennisopenbaarbestuur.nl/thema/artifici%C3%ABle-intelligentie-en-publiekewaarden
1199
The Public Voice, Universal Guidelines for AI Endorsement,
https://thepublicvoice.org/AI-universal-guidelines/endorsement/
1200
Global Privacy Assembly, Adopted Resolution on Accountability in the Development
and Use of Artificial Intelligence (October 2020) https://globalprivacyassembly.org/wpcontent/uploads/2020/10/FINAL-GPA-Resolution-on-Accountability-in-theDevelopment-and-Use-of-AI-EN-1.pdf
306
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
investigating the Dutch government’s use of algorithms concluded that
“government interests are central to algorithms, not private citizens’
interests; that the government recognises the importance of privacy but
takes little account of ethical aspects; and warns against government’s use
of algorithms becoming dependent on external suppliers.”1201
Evaluation
The Netherlands has taken positive steps towards the rights-based
deployment of AI with endorsement of OECD AI Principles, GDPR, and
well-established protections for personal data. The country is expected to
expand algorithmic transparency with the example set by Amsterdam AI
registry initiative. And the Dutch court should be credited with a landmark
decision concerning the use of secret algorithms in government services.
Still the rise of predictive policing and biometric databases, as well as riskbased systems that may adversely impact minority and vulnerable groups
remains a concern. The Netherlands is a member of the Global Partnership
on AI. As the Netherlands seeks to lead on AI testing and responsible AI
applications, future public adoption of AI systems is expected to be aligned
with responsible and human-centric development and use of AI, respecting
human rights and fundamental freedoms.
The Netherlands Court of Audit (January 2021). Understanding algorithms.
https://english.rekenkamer.nl/publications/reports/2021/01/26/understanding-algorithms
1201
307
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Nigeria
National AI Strategy
Nigeria does not have a national (AI strategy and policy that can
help promote and attract investments in AI into Nigeria) but has a National
Digital Economy Policy and Strategy,1202 published by the country’s
Ministry of Communications and Digital Economy. Nigeria is among the
African countries that is working on a unified African AI Strategy with the
African Union (AU) in a formal AI working group, the AU AI Working
Group. The Working Group held its first meeting in December 2019, with
Egypt, Algeria, Uganda, Cameroon, and Uganda. Discussions included
work on an AI Observatory and an AI Readiness Indicator.1203
Nigeria is a member of UNESCO, and participated in the UNESCO
AI Recommendations process. Nigeria also participates in The African
Forum for Ethics and Governance of Artificial Intelligence (AFEGAI),
which was created in 2019 following UNESCO’s and African Member
States’ recommendation to establish an African Forum of Association of
Artificial Intelligence (AI).1204 AFEGAI brings together AI constituencies
in Africa to support the ethical development of Artificial Intelligence.
AFEGAI coordinates AI Governance Forums in Africa.1205
In November 2020, the Nigerian government launched1206 the
country’s first Artificial Intelligence and Robotics Centre in a bid to position
the country for the Fourth Industrial Revolution that is anchored on
emerging technologies. The Centre’s focus is to serve as the digital
laboratory for advancing skills development and innovation in emerging
technologies with emphasis in AI and Internet of Things (IoT).
Earlier, Nigeria announced that it was working with the United Arab
Emirates on Solar Energy and Artificial Intelligence1207 after a meeting held
1202
Nigerian Communications Commission ‘The National Digital Economy Policy and
Strategy’ (June 1, 2020) https://www.ncc.gov.ng/docman-main/industrystatistics/policies-reports/883-national-digital-economy-policy-and-strategy/file.
1203
African Union AI Working Group holds first session, (Dec. 2019), https://mcit.gov.eg
1204
Forum on AI in Africa Summit, The Benguerir Declaration (Dec. 13, 2018),
https://en.unesco.org/sites/default/files/ai_outcome-statement_africa-forum_en.pdf.
1205
UNESCO, Multistakeholder group discusses ten building blocks towards creating
inclusive AI policies (Jan. 24, 2022), https://en.unesco.org/news/multistakeholder-groupdiscusses-ten-building-blocks-towards-creating-inclusive-ai-policies
1206
Ogunfowoke A, Innovation Village (Nov. 20, 2020); FG Launches Nigeria’s first
Artificial Intelligence and Robotics Centre https://innovation-village.com/fg-launchesnigerias-first-artificial-intelligence-and-robotics-centre/.
1207
O Shogbola, Nigeria and UAE to cooperate on Artificial Intelligence and Solar
Energy (Oct. 24, 2018), https://perma.cc/GN67-8M8R.
308
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
in Nigeria between Nigeria’s Minister of Science and Technology, and the
Ambassador of the United Arab Emirates to Nigeria. It said it would work
through one of its agencies, National Agency for Science and Engineering
Infrastructure, to develop AI in Nigeria. It is unclear if this engagement has
a relationship with the establishment of Nigeria’s Robotics and Artificial
Intelligence Centre.
Nigeria Communications Commission (NCC), the telecoms
regulator, also announced the establishment of a new department on digital
economy under the Office of the Executive Vice Chairman/CEO to focus
on implementing programmes and policies that are aimed at fully
supporting and promoting the national digital economy agenda of Nigeria's
Federal Government. The Commission says the department is domiciled
under the office of the Chief Executive Officer to indicate the seriousness
attached to the objective. Prior to this, Nigeria's Ministry of Science and
Technology announced a goal for ensuring that Nigeria is well suited to the
AI economy as critical to driving knowledge and Innovation and to create
more job opportunities for Nigeria.1208
Public Participation
There’s no evidence of any activity organized by the Nigerian
government towards consulting with the public on AI. However, Google1209
and Microsoft1210 released research documents that proposed policy
direction for the Nigerian government. The Microsoft research document
highlighted that Nigeria, like most emerging economies, is yet to capitalize
on AI despite its inherent opportunities because of low awareness. It
proposed the need for a national AI policy and strategy to provide a cohesive
policy roadmap for channeling the potential of AI in Nigeria. It also
emphasized the need for digital AI policy capacity building. Also, the
document stressed the need to optimize its data ecosystem to leverage AI
and machine learning opportunities for social good in areas such as financial
inclusion, universal healthcare and food security.
1208
NCC, Press Statement: NCC Creates New Department to Accelerate FG’s Digital
Economy Agenda (July 7, 2020), https://www.ncc.gov.ng/media-centre/newsheadlines/839-press-statement-ncc-creates-new-department-to-accelerate-fg-s-digitaleconomy-agenda
1209
Courtney Heldreth et al, AI in Nigeria (2019),
https://research.google/pubs/pub48985/
1210
Microsoft, Enabling a Digital Nigeria; A Position Paper of Microsoft’s Vision for
Digital Transformation and a Digital Economy that Works for Everyone (2020),
https://info.microsoft.com/rs/157-GQE-382/images/EN-CNTNT-WhitepaperSRGCM3460.pdf
309
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Research and Development
According to the director-general of Nigeria's National Information
Technology Development Agency (NITDA), the National Centre for
Artificial Intelligence and Robotics (NCAIR) is also expected to be a
research development center.1211 The Communications Commission
commissioned1212 a study to assess the ethical and societal impact of AI to
achieve economic diversification in an inclusive and sustainable way. The
study concluded that it is the duty of regulators to govern artificial power
while it is also the responsibility of programmers and engineers to ensure
that ethical and security concerns are addressed during initial design of these
systems. It urges the NCC to step forward and develop frameworks for AI
in Nigeria.
Privacy
Section 37 of the Nigerian Constitution provides for the right to
privacy. Beyond this, Nigeria is yet to enact a data protection law to cater
for data privacy issues or privacy as it may apply in the digital age.
Currently, there is a National Data Protection Regulation (NDPR)1213 and a
Data Protection Bill has been drafted.1214 The scope of the NDPR applies to
all transactions intended for the processing of personal data of natural
persons residing in Nigeria or Nigerian citizens residing in foreign
jurisdictions. Data processing under the NDPR includes the collection,
recording, storage, retrieval, use, disclosure, transmission, erasure, and
destruction of personal data. The stated objective of the NDPR includes;
1) To safeguard the rights of natural persons to data privacy;
2) Fostering of safe conduct for transactions involving the
exchange of Personal Data;
3) To prevent manipulation of Personal Data;
1211
Synced, Nigerian Government to Set Up the Nation’s Centre for AI and Robotics to
Empower Students (Sept. 3, 2020),‘https://syncedreview.com/2020/09/04/nigeriangovernment-to-set-up-the-nations-centre-for-ai-and-robotics-to-empower-students/
1212
NCC; Ethical and Societal Impact of Artificial Intelligence
https://www.ncc.gov.ng/technical-regulation/research/919-ethical-societal-impact-ofartificial-intelligence-ai
1213
HA Kurth, Nigeria Issues New Data Protection Regulation (Aug. 5, 2019),
https://www.huntonprivacyblog.com/2019/04/05/nigeria-issues-new-data-protectionregulation/ (Accessed 25 May 2021).
1214
ITedge, Nigerian Government Seeks Your Input On Draft Bill On Data Protection
(Aug. 26, 2021), https://itedgenews.ng/2020/08/26/nigerian-government-seeks-yourinput-on-draft-bill-on-data-protection/ (Accessed 16 May 2021).
310
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
4) To ensure that Nigerian businesses remain competitive in
international trade through the safe-guards afforded by a sound
data protection regulation.
On the other hand, the Data Protection Bill1215 proposes to establish
and provide an efficient regulatory framework to protect personal data,
regulate the processing of information relating to data subjects, and
safeguard their fundamental rights and freedoms as guaranteed under the
Nigerian Constitution. The Bill also seeks to promote a code of practice that
ensures the privacy and protection of data subject’s data without unduly
undermining the legitimate interests of commercial organisations and
government security agencies for such personal data. It also seeks to
minimise the harmful effect of personal data misuse or abuse on data
subjects and other victims. In addition, the Bill provides for the
establishment of an impartial, independent and effective regulatory
authority that will coordinate data protection and privacy issues. The
regulatory authority is expected to superintend over data controllers and
data processors within the private and public sectors; and ensure that
personal data is processed in accordance with the data protection principles.
Digital Rights in Nigeria
In 2016, Nigeria joined the United States, Canada, Australia, United
Kingdom, and some European Union states to sponsor a United Nations
Resolution that affirms that rights that apply offline must also apply
online.1216 Section 37 of the Nigerian Constitution provides for the right to
privacy while section 37 makes provision for the right to freedom of
expression. Nigeria is a party to the African Charter on Human and Peoples’
rights1217 and the International Covenant on Civil and Political rights. In
2019, the Nigerian parliament passed a Digital Rights and Freedom Bill1218
into law but the president declined assent to the Bill, citing possible
duplication with other proposed bills and that the Bill was too technical.1219
1215
Clause 1 Draft Data Protection Bill 2020.
UN Human Rights Council, The Promotion, Protection and Enjoyment of Human
Rights on the Internet: Resolution Adopted by the Human Rights Council, 18 July 2016,
A/HRC/RES/32/13, https://www.article19.org/data/files/Internet_Statement_Adopted.pdf
1217
UN Human Rights, Ratification Status for Nigeria (May 30, 2021),
‘https://tbinternet.ohchr.org/_layouts/15/TreatyBodyExternal/Treaty.aspx?CountryID=12
7&Lang=EN
1218
Digwatch, Nigerian Senate passes Digital Rights and Freedom Bill (Mar. 13, 2018),
https://dig.watch/updates/nigerian-senate-passes-digital-rights-and-freedom-bill
1219
S Fowowe (Mar. 20, 2019).
1216
311
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
This raises concerns with respect to Nigeria’s readiness to provide an
enabling regulatory environment for new and emerging technologies such
as AI, given their human rights implications. The proposed bill sought to
codify the application of human rights within today’s digital realities by
guaranteeing the application of human rights for users of digital platforms
and or digital media and for related matters1220 and could have served to
mitigate some of the human rights concerns associated with application of
AI systems.
In a report on digital rights and privacy in Nigeria,1221 it was
observed that the Nigerian State conducts surveillance activities without
judicial oversight and a comprehensive framework for data protection. The
report recommended the enactment of a comprehensive framework for data
protection and privacy and judicial oversight over surveillance. The 2020
Freedom on the Net report rates Nigeria as partly free stating among other
things that ‘several legal provisions may allow the government to conduct
surveillance without respect for the necessity and proportionality principles,
and international guidelines that apply human rights law to monitoring
technologies.1222
COVID-19 Pandemic
The Nigerian Governors forum said it was collaborating with
MTN,1223 the leading telecommunications company, “to mitigate the effect
of the COVID-19 pandemic by mining its users’ data to profile the states’
vulnerability to the spread of the coronavirus.” This is an indication of
efforts to leverage big data for governance. The outcome of that
collaboration has not been publicly communicated. Also, MTN Nigeria
Buhari declines assent to Digital Rights and Freedom Bill, four others’
https://guardian.ng/news/buhari-declines-assent-to-digital-rights-and-freedom-bill-fourothers/
1220
Pavestone Legal, The Digital Rights and Freedom Bill (2019)
https://pavestoneslegal.com/the-digital-rights-and-freedom-bill-2019/
1221
A Adegoke (2020), Digital Rights and Privacy In Nigeria (2020),
https://paradigmhq.org/wp-content/uploads/2021/05/Digital-Rights-and-Privacy-inNigeria_0.pdf
1222
Freedom House (2020) ‘Freedom of the Net 2020
https://freedomhouse.org/country/nigeria/freedom-net/2020
1223
NGF; ‘Governors, MTN partner to use data to halt spread of COVID-19’
http://www.ngf.org.ng/index.php/73-featured-news/1564-governors-mtn-partner-to-usedata-to-halt-spread-of-covid-19
312
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
denied sharing identifiable user’s data.1224 A World Bank recommendation
to the Nigerian government proposes investment in AI enabled Geo-spatial
data and that such data can help policymakers in Nigeria respond to the
COVID-19 crisis and build back better.1225 Meanwhile, Nigeria has multiple
data sets, captured across multiple platforms such as Bank Verification
Number (BVN), Voter’s card, International passport, Road Safety and
National Identification Number (NIN). Reports suggest that the country
loses up to US$2 billion annually on biometric data collection
duplication.1226 This shows a lack of uniformity or harmonization of
captured data, posing a clear challenge to the deployment of AI.
Facial Recognition
Nigeria plans to use facial recognition technology as an essential
part of its digital identification scheme. According to the Director-General
of the Nigerian Identity Management Commission (NIMC), the agency
has plans to capture the iris of prospective enrollees and that the agency
already has the capacity for capturing the iris at its backend.1227 Also,
Access Bank, one of Nigeria’s leading financial institutions announced
plans to launch a facial recognition payment system which will enable it to
verify its customers and to perform transactions without a password.1228 In
addition to this, one of NIMC’s licensees for Identity verification in
Nigeria says it offers “AI powered facial recognition technology for ecommerce ID authentication”.1229
Selected AI initiatives in Nigeria
The private sector and the tech ecosystem are taking the lead with
respect to AI initiatives in Nigeria. Most of the initiatives that may be
1224
Communications Week ‘MTN Nigeria Denies Sharing Identifiable User Data’ (02
May 2020), https://www.nigeriacommunicationsweek.com.ng/mtn-nigeria-deniessharing-identifiable-user-data/ (Accessed 07 June 2021).
1225
J Blummenstock et al; ‘Using Big Data and machine learning to locate the poor in
Nigeria’ (21 February 2021) https://blogs.worldbank.org/opendata/using-big-data-andmachine-learning-locate-poor-nigeria (Accessed 30 June 2021).
1226
J Lee ‘Nigeria loses $2b annually to agencies’ duplication of biometrics data
collection’ (14 August 2021); https://www.biometricupdate.com/201708/nigeria-loses2b-annually-to-agencies-duplication-of-biometrics-data-collection (30 June 2021).
1227
ID4Africa,(23 September 2021); ‘EP7; NIgeria’s Identity Ecosystem’
https://www.youtube.com/watch?v=OgcKzQ8I7_U&t=4605s Watch from 1:18:00.
(Accessed 24 May 2021).
1228
The Payper(03 March 2020) ‘Access Bank to launch a facial recognition payment
system in Nigeria’ https://thepaypers.com/mobile-payments/access-bank-to-launch-afacial-recognition-payment-system-in-nigeria--1240957# (30 May 2021)
1229
VerifyMe https://verifyme.ng/ (Accessed 22 May 2021).
313
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
identified are private player driven. Meanwhile, the Nigerian government
has promoted or adopted some AI-driven technologies. For example, it
launched a homegrown plagiarism detection software codenamed
EagleScan.1230 The software, which comes with a very robust artificial
intelligence-driven data analytics and visualisation engine, is currently
hosted on www.eaglescan.ng.
The NASS-AI project uses AI to classify parliamentary bills from
the national assembly, with the aim of making them more accessible. The
system is anchored on the principle of open data, to ensure that legislativerelated documents are complete, accessible, and machine-processable,
amongst other requirements.1231
In December 2020, the Ministry of Science and Technology
announced its plans to deploy AI for mining safety and efficiency. In the
announcement, the Minister of Science and Technology said the
deployment of relevant technology to the solid mineral sector will enhance
economic diversification in an inclusive and sustainable way.1232 He added
that science, technology and innovation are of strategic importance in fully
exploiting the solid mineral industry, to increase wealth and create more
jobs for citizens.
Evaluation
Nigeria does not yet have a national strategy for AI. It has not
endorsed the OECD/G20 AI Principles and has not promoted public
participation in the development of AI policy. Nigeria has secondary
legislation on data protection (the NDPR) and is yet to adopt a
comprehensive law for data protection. Nigeria’s human rights record is not
impressive. The 2021 Human Rights Watch 2021 described Nigeria as
“partly free” with a score of 45/100 for the protection of political rights and
civil liberties, down slightly from 2020.1233 The private sector and young
people are driving the adoption of AI in Nigeria. Government is yet to assert
1230
H Tyoemba, Federal Govt Launches Home-grown Plagiarism Detection Software,
Vanguard https://leadership.ng/federal-govt-launches-home-grown-plagiarism-detectionsoftware/.
1231
T Atoyebi, NASS AI: Revolutionizing access to Nigeria’s legislative bills,
Technopreneur (Feb. 7, 2020), https://technopreneur.com.ng/2020/02/07/nass-airevolutionizing-access-to-nigerias-legislative-bills/
1232
C Uchechukwumgemezu, Minister: AI to be deployed for mining safety, efficiency’
published by TodayNG (Dec. 9, 2020), https://www.today.ng/technology/ministerdeployed-mining-safety-efficiency-331732
1233
Freedom in the House, Freedom in the World 2021 – Nigeria (2021),
https://freedomhouse.org/country/nigeria/freedom-world/2021
314
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
itself through development of regulations and policy around the subject of
AI in Nigeria.
315
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Norway
National AI Strategy
The Norwegian Government presented its National AI strategy in
January 2020.1234 The objective of the strategy is to outline the policy
actions for the coming years in order to maximise the opportunities that AI
can bring along for Norwegian individuals, for businesses and industry, and
for the public sector. 1235 To achieve this outcome, the national AI strategy
highlights the following policy initiatives:
• Expanding the offer of education programmes and workplace
trainings in the field of AI in order to create a solid basis of digital
skills and capabilities;
• Strengthening the Norwegian research in AI;
• Enhancing the innovation capacity in AI in both the private and
public sector;
• Outlining ethical principles for AI in order to allow fair, reliable
and trustworthy AI-related developments;
• Establishing digitalisation-friendly regulations as to define the
legislative framework in which AI developments take place;
• Constructing a strong data infrastructure ensuring open data and
data sharing across sectors and business areas. Dedicated
opportunities for language data resources are established
through The Norwegian language bank at the National library;
• Deploying a telecommunication infrastructure that provides highcapacity connectivity and computing power, and that ensures
security in AI-based systems.
The report also includes a section on ethics which focuses on
adopting the EU ethics guidelines for trustworthy AI1236 and the AI
principles from the OECD1237. The work on Norway’s AI national strategy
was founded upon the agreement from April 2018, when 25 European
The Government of Norward, Ministry of Local Government and Modernization, The
National Strategy for Artificial Intelligence (Jan. 14, 2020), The National Strategy for
Artificial Intelligence
1235
European Commission, AI Watch (2020), https://knowledge4policy.ec.europa.eu/aiwatch/norway-ai-strategy-report_en.
1236
European Commission, EU Ethics guidelines for trustworthy AI, April 2019
https://wayback.archive-it.org/12090/20201227221227/https://ec.europa.eu/digitalsingle-market/en/news/ethics-guidelines-trustworthy-ai
1237
OECD, AI Principles, https://www.oecd.org/going-digital/ai/principles/
1234
316
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
countries, including Norway, signed a Declaration of cooperation on
Artificial Intelligence (AI).1238
Other efforts were made for the publication of the national strategy.
For example, the Norwegian Data Protection Authority (DPA) published a
report regarding the artificial intelligence and privacy.1239 Another report
Digital21 focused on national collaboration and encouraged engagement
from industry, academia and government.1240 The Norwegian Board of
Technology also published a report, Artificial Intelligence: Opportunities,
Challenges and a Plan for Norway.1241
The national AI strategy emphasizes that implementation and
progress will be closely reviewed and, if necessary, changed with further AI
policies. The frequency with which revised strategy reports will be provided
is not specified in the plan.
Public Participation (A)
The Norwegian Ministry of Local Government and Modernisation
published the National Strategy for Artificial Intelligence on January 14th,
2020.1242 Throughout 2019, the Minister of Digitalisation travelled around
Norway to listen to suggestions from different communities to inform the
national strategy. Additionally, there were close to 50 written statements
from a variety of businesses and government agencies that provided
comments on the content of the strategy. Norway has a wide range of policy
documents pertaining to the field of artificial intelligence. These do not only
come from the Norwegian government, but also from a range of actors in
civil society.
Norway also developed centralised, accessible repositories of open
public data. In Norway, the Brønnøysund Register Centre and the
Norwegian Digitalisation Agency have established a national directory of
data held by different public agencies, their relationships, what they mean
European Commission, EU Declaration on Cooperation on Artificial Intelligence
(Apr. 10, 2018),
https://ec.europa.eu/jrc/communities/en/community/digitranscope/document/eudeclaration-cooperation-artificial-intelligence
1239
Datasylnet, Report on the regulation of privacy and AI (June 2018),
https://iapp.org/media/pdf/resource_center/ai-and-privacy.pdf
1240
Digital 21, https://digital21.no
1241
Teknologirådet, Artificial Intelligence: Opportunities, Challenges and a Plan for
Norway (Nov. 26, 2018), https://teknologiradet.no/en/publication/ai-and-machinelearning-possibilites-challenges-and-a-plan-for-norway/
1242
European Commission, AI Watch (2020), https://knowledge4policy.ec.europa.eu/aiwatch/norway-ai-strategy-report_en.
1238
317
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and whether data can be shared and on what terms. Portugal also plans to
create a centralised repository for administrative data.1243
Global Partnerships
Whilst Norway has not joined the Global Partnership on AI, it has
been actively collaborating, especially with its neighbors. The NordicBaltic Region is one of the few regions that have declared a collaboration
on artificial intelligence.1244 The declaration was released in May 2018 as
AI in the Nordic-Baltic Region. The collaboration includes focus areas such
as ‘’improving opportunities for skills development’’, ‘’developing ethical
and transparent guidelines to guide when and how AI applications should
be used’’, ‘’ensuring that AI gets a prominent place in the European
discussion’’ and ‘’utilize the structure of Nordic Council of Ministers to
facilitate the collaboration in relevant policy areas.’’ Under this new
resolution, governments will apply a joint approach to taking advantage of
and further elevate the Nordic region’s already high status as a leader in the
development and use of AI and digital technologies.
Norway has worked to strengthen national funding for research and
innovation in artificial intelligence, and also to substantially increase their
research collaborations. With the Norwegian Artificial Intelligence
Research Consortium (NORA.ai), Norway has taken important steps to
support the European ambition of increased cross-border co-operation in AI
research.1245
On November 24 the Supreme Audit Institutions of Finland,
Germany, the Netherlands, Norway and the UK jointly published a
whitepaper called Auditing machine learning algorithms for public
auditors. This paper discussed in detail audits of machine learning (ML)
algorithms by the Supreme Audit Institution with project management, data,
model development, model in production and evaluation. The project
auditingalgorithms.net is maintained by The Office of the Auditor General
of Norway (Riksrevisjonen).
OECD (2021), State of implementation of the OECD AI Principles: Insights
from national AI policies (Jun 18, 2021), https://doi.org/10.1787/1cd40c44-en.
1244
Government of Sweden, AI in the Nordic Baltic Region (May 14, 2018),
https://www.regeringen.se/49a602/globalassets/regeringen/dokument/naringsdepartement
et/20180514_nmr_deklaration-slutlig-webb.pdf
1245
Norwegian Artificial Intelligence Research Consortium (NORA), About NORA,
https://www.nora.ai/about/
1243
318
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD AI Principles
Norway is a member of the OECD and endorsed the OECD AI
Principles.1246 In terms of ethical principles, the Norwegian government
seeks to encourage responsible, accountable, transparent, and trustworthy
AI while protecting integrity and privacy.1247
Additionally, Norway is supporting the OECD AI principles by
fostering a digital ecosystem for AI with the National Data Catalogue.1248
This is a public website providing an overview of descriptions of datasets,
concepts, APIs and information models. Both the public and private sectors
are involved in this project. This project was established by the
Brønnøysund Register Centre and the Norwegian Digitalisation Agency.
To further the outreach of the overall project to the private sector the
government funded the creation of the Data Factory, a new interface of the
catalogue to make navigation easier for the public and companies.1249
Within the Data Factory, a community named Data Village has also been
built as a community of practice in various dataset categories. NORA.ai also
created the Norwegian AI Directory to map out all the activities within the
field of AI in Norway1250. In addition, NORA.startup has been established
as an ecosystem of new companies in the field of AI that has gone through
a quality assurance process to ensure the startups are active in researchbased innovation. These startups are registered as part of a larger ecosystem
that Norway is part of called the European AI Startup Landscape together
with France, Germany and Sweden NORA.ai, the Norwegian Open AI Lab
(NAIL) and Cluster for Applied AI in Halden jointly contribute to this
digital ecosystem.
Human Rights
Norway was one of the 193 countries that signed the Universal
Declaration of Human Rights. This has also laid the foundation for other
treaties ratified by Norway, such as the European Convention of Human
Rights and the International Covenant on Civil and Political Rights.1251
1246
OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence,
(May 22, 2019), https://www.oecd.org/science/forty-two-countries-adopt-new-oecdprinciples-on-artificial-intelligence.htm
1247
Norway, National Strategy for Artificial Intelligence (Jan. 14, 2020)
https://www.regjeringen.no/en/dokumenter/nasjonal-strategi-for-kunstigintelligens/id2685594/
1248
Norge.no, National Data Catalogue, https://www.norge.no/en/service/national-datacatalog
1249
Datafabrikken, https://datafabrikken.norge.no
1250
Norwegian AI Directory, https://aidirectory.no
1251
NIM, The Human Rights Framework in Norway
https://www.nhri.no/en/2019/the-human-rights-framework-in-norway/
319
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Among these, there is also Convention 108+, the Modernized Convention
for the protection of individuals with regard to the processing of personal
data.
On 21 May 1999, the Norwegian Parliament passed the Human
Rights Act, and this elevates five key human rights conventions to a special
status in Norwegian law. These conventions prevail in the event of a conflict
with regular domestic legislation. To further strengthen human rights, the
Norwegian Constitution was amended as part of its bicentennial anniversary
in May 2014 to add several human rights related to the environment and the
Sámi people. In 2018, human rights in Norway were further implemented,
when the International Convention on the Elimination of All Forms of
Racial Discrimination (ICERD) was implemented in Norwegian law by the
Equality and Anti-Discrimination Act. Human rights and democracy are
emphasized on throughout the Norwegian national AI strategy, with clear
commitment to fundamental values and ethical principles.
Norway typically ranks among the top nations in the world for the
protection of human rights and transparency. According to Freedom House,
Norway’s Global freedom score is 100/100.1252 In the latest Freedom
House’s Country Report, Norway is described as one of the most “robust
democracies in the world.”1253 Elections are free and fair, and power
regularly rotates between parties. Civil liberties are respected, with
independent media and civil society actors holding the government to
account.
Furthermore, diversity in AI is valued in Norway through
prioritising the development of language technology systems that support
communications in Norwegian, Sámi and smaller dialects. In this way
ensuring that the indigenous rights of the Sámi people, a Finno-Ugricspeaking people inhabiting the region of Sápmi that Norway is part of, are
included within applications in the field of AI. This focus on the Sámi
language is also included in the Norwegian National AI Strategy.
Oversight: Data Protection Authority
The Norwegian Data Protection Authority (DPA or Datatilsynet) is
the national Data Protection Authority for Norway. It resides in Oslo and is
in charge of enforcing the GDPR in Norway. The GDPR is enshrined in
Norwegian law in the form of a Personal Data Act which came into force
Freedom House, Freedom in the World 2021 – Norway (2021),
https://freedomhouse.org/country/norway/freedom-world/2021.
1253
Freedom House, Freedom in the World 2021 – Norway (2021),
https://freedomhouse.org/country/norway/freedom-world/2021
1252
320
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
on May 25 2018.1254 Together with other EEA countries (Iceland and
Liechtenstein), the Norwegian Data Protection Authority became members
of the European Data Protection Board (EDPB), however without voting
rights and without the right to be elected as chair and vice-chair, for GDPRrelated matters.1255
The Norwegian National AI Strategy, highlights that the Norwegian
Data Protection Authority (DPA) is an important entity. Datatilsynet has
been very active in both enforcement and publication of guidelines on a
wide range of significant data protection concerns, including codes of
conduct, CCTV surveillance and software development with Privacy by
Design and by Default. The Norwegian Data Protection Authority has made
a list of processing activities that they believe are likely to pose a significant
risk to data subjects' rights and freedom available, and will always
necessitate a Data Protection Impact Assessment such as: 1256
● processing of biometric data for identification purposes on a large
scale
● collecting and combining personal data from third parties in order
to decide whether the data subject shall be offered, continue to
receive, or shall be denied a product, service or offer
● monitoring the employees internet activity, electronic
communication or camera surveillance for the purposes of
employee monitoring
● systematic monitoring, including camera surveillance, on a large
scale, in areas accessible by the public
Interestingly, following a data protection impact assessment of
Facebook, Datatilsynet, announced that it will no longer communicate via
the social media network. According to the DPA, parties processing
personal data must follow the EU General Data Protection Regulation,
which applies when an organization utilizes social media. However,
Facebook is not sufficiently complying with Article 26. "We believe the
risks to the users’ rights and freedoms associated with the processing of
personal data through a Page on Facebook are too high," stated by DPA
Director-General Bjorn Erik Thon.1257 This is not the first time the DPA is
Datasylnet, Artificial Intelligence and Privacy (January 2018),
https://www.datatilsynet.no/globalassets/global/english/ai-and-privacy.pdf
1255
Data Protection Law of the World, Norway Data Protection Authority (January
2021),
https://www.dlapiperdataprotection.com/index.html?t=authority&c=NO
1256
Datasylnet, Processing operations subject to the requirement of a data protection
impact assessment. https://www.datatilsynet.no/globalassets/global/dokumenter-pdferskjema-ol/regelverk/veiledere/dpia-veileder/dpialist280119.pdf
1257
Datasylnet, Norwegian Data Protection Authority choose not to use Facebook(2021),
1254
321
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
opposing Facebook. Back to 2012, the Norwegian DPA raised important
concerns regarding Facebook’s facial recognition tool. Bjorn Erik Thon
stated that "It's a very powerful tool Facebook has and it's not yet clear how
it all really works," and "They have pictures of hundreds of millions of
people. What material Facebook has in its databases is something we need
to discuss with them."1258
The Norwegian DPA also published a report on the regulation of
privacy and AI in June 2018. The report showed how imperative to further
knowledge about the privacy implications of artificial intelligence, and “not
only in order to safeguard the right to privacy of the individual, but also to
meet the requirements of society at large’’.1259 In this report, the DPA
provides greater technical detail in describing artificial intelligence, while
also taking a closer look at four relevant AI challenges associated with the
data protection principles embodied in the GDPR: fairness and
discrimination; purpose limitation; data minimization; and transparency and
the right to information. A strong emphasis lays on bringing awareness to
the ethical and privacy consequences of AI systems, as well as ensuring that
the deployed systems respect privacy by design and meet the legislative
requirements. The DPA is committing extensive work in ‘’finding a balance
between considerable social advances in the name of AI and fundamental
privacy rights.’’1260
The creation of a sandbox in the field of artificial intelligence was
also included in the national AI strategy of Norway. This regulatory
Sandbox for responsible artificial intelligence, issued by the Norwegian
DPA, has gone through several projects openly discussing these with the
public and as such improving transparency. Providing free guidance to a
handful of carefully selected companies across different sectors, the goal is
promoting the development of innovative artificial intelligence solutions
https://www.datatilsynet.no/en/news/2021/norwegian-data-protection-authority-choosenot-to-use-facebook/
1258
Bloomberg, Facebook faces facial recognition fight in Norway (Aug. 4, 2012),
https://www.bloomberg.com/news/articles/2012-08-02/facebook-faces-norway-probeover-facial-recognition-photo-tags
1259
Datasylnet, Report on the regulation of privacy and AI (June 2018),
https://iapp.org/media/pdf/resource_center/ai-and-privacy.pdf
1260
Datatilsynet, Report on the regulation of privacy and AI (June 2018),
https://iapp.org/media/pdf/resource_center/ai-and-privacy.pdf
322
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
that are both responsible and ethical.1261 The Sandbox will continue with
new projects in 2022.1262
Data Protection Laws and Policies
The Norwegian government recently proposed the adoption of a
new Electronic Communication Act, even though it might be short-lived
due to the possible adoption of a new ePrivacy Regulation at the EU level
soon. Among other things, the proposed act could entail changes regarding
the rules on consent for the use of tracking technologies (e.g., cookies),
which under the current legal regime may be given through web browser
settings.
Furthermore, in December 2021, Datatilsynet issued its highest fine
so far — 65 million NOK (around 6.5 million euros) — against Grindr for
failing to comply with the consent requirements under the GDPR. Grindr is
likely to appeal the fine before the Norwegian Privacy Board of Appeals,
Personvernnemnda, which could issue its decision in 2022.1263
Evaluation
Norway has a full-fledged official national AI strategy and AI ethics
is a central topic. Norway endorsed the OECD AI Principles and is
committed to developing trustworthy AI. Norway ranks at the top among
nations for the protection of political rights and civil liberties and has an
active data protection agency. Norway has endorsed and is implementing
the Universal Declaration of Human Rights. Human rights and democracy
is also present throughout the Norwegian national AI strategy. The
Norwegian strategy was shaped by a wide participation both with physical
meetings across the country and comments from a range of actors in society.
Norway has materials available to the public from the strategy and within
other parts of the Norwegian government, ranging from the DPA to the
Office of the Auditor General of Norway. The country has independent
oversight of AI as well as goals of fairness, transparency and accountability.
Norway has rights to algorithmic transparency through the GDPR and this
is monitored by the Norwegian DPA. Overall Norway has a strong
Datatilsynet, Sandbox for responsible artificial intelligence 2021
https://www.datatilsynet.no/en/regulations-and-tools/sandbox-for-artificial-intelligence/
1262
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
1263
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
1261
323
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
commitment to democratic values in the development and implementation
of artificial intelligence.
324
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Philippines
National AI Strategy
In May 2021, the Philippines’ Department of Trade and Industry
(DTI) launched the country’s national AI roadmap,1264 which aims to make
the Philippines a regional AI “powerhouse”, accelerating “the adoption and
utilization of AI in the country to advance industrial development, generate
better quality entrepreneurship, and higher-paying opportunities for
Filipinos.”1265 The roadmap was developed by the Philippines Council for
Industry, Energy and Emerging Technology Research and Development
(DOST-PCIEERD) under the country's Department of Science and
Technology.1266 The roadmap identifies various intended applications of AI,
including its use in “real estate, banking and financial services, surveillance,
retail and e-commerce, education, space exploration, agribusiness, urban
planning, manufacturing, healthcare, and logistics and transportation.”1267
The national AI roadmap describes the four important dimensions
for AI readiness for implementation, infrastructure and investment,
namely:
1) Digitization and Infrastructure;
2) Research and Development;
3) Workforce Development; and
4) Regulation.1268
Central to the DTI’s roadmap is the establishment of the National
Center for AI Research (NCAIR), whose full-time scientists and research
engineers will assist micro, small, and medium enterprises (MSMEs)
interested in using AI technology — an element of the country’s approach
to inclusive AI development. The roadmap also includes plans to build a
national data center (NDC).1269
Philippines’ Department of Trade and Industry, ARTIFICIAL INTELLIGENCE
ROADMAP, http://innovate.dti.gov.ph/resources/roadmaps/artificial-intelligence/
1265
Kris Crismundo, DTI eyes PH as AI powerhouse in region,(May 5, 2021),
https://www.pna.gov.ph/articles/1139198
1266
Philippines Council for Industry, Energy and Emerging Technology Research and
Development, Artificial Intelligence and Information & Communications Technology:
Roadmapping Executive Report, http://projects.pcieerd.dost.gov.ph/roadmaps/AIICT.pdf
1267
Kris Crismundo, DTI eyes PH as AI powerhouse in region (May 5, 2021),
https://www.pna.gov.ph/articles/1139198
1268
Ibid.
1269
Philippines’ Department of Trade and Industry, ARTIFICIAL INTELLIGENCE
ROADMAP, http://innovate.dti.gov.ph/wp-content/uploads/2021/05/National-AIStrategy-infographic-2048x1105.jpg
1264
325
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Furthermore, in February 2020, the DOST launched the Smarter
Philippines through Data Analytics, Research and Development, Training
and Adoption (SPARTA) initiative to retool and upgrade the skills of an
initial 30,000 workers by 2022.1270 There is little information available
concerning the program’s uptake and effectiveness.
Public Participation
There is some evidence of a public consultation process in the
development and implementation of the country’s national AI policy. The
roadmap mandates the establishment of a “committee of experts in data and
AI ethics who will guard against abuse/misuse of data and AI algorithms,”
but it is unclear whether this committee has been formed and if it will have
a mechanism for public consultation.1271 On October 28th, 2021, pursuant to
the DTI’s Inclusive Innovation Strategy, the DTI hosted “Artificial
Intelligence: Opportunities and Challenges for Philippine Industries,” a
virtual session open to the public intended to “bring together participants
from the government, industries and academe not only to deliberate on
matters of common interest concerning AI, but more so to bring everyone
into the fold as we embrace and adapt to our new economic realities.”1272
The National Privacy Commission’s website includes opportunities for the
public to vocalize data-related concerns through its “AskPRIVA” tool, and
includes contact information to file complaints or to contact relevant
authorities.1273
OECD/G20 Principles
The Philippines is not a member of the OECD or the G20, and its
roadmap makes little direct mention of the OECD or G20 AI principles.
Despite having no explicit endorsement of these international principles,
comments from leaders and provisions from the roadmap embody a
commitment to OECD principles for human rights, inclusive growth and
transparency. In a keynote speech, DTI Secretary Ramon Lopez remarked
Smart Philippines through Data Analytics R&D and Adoption,
https://sparta.dap.edu.ph/
1271
Philippines’ Department of Trade and Industry, ARTIFICIAL INTELLIGENCE
ROADMAP, http://innovate.dti.gov.ph/resources/roadmaps/artificial-intelligence/
1272
Philippines’ Department of Trade and Industry, Artificial Intelligence: Opportunities
and Challenges for Philippine Industries (Oct. 22, 2021),
https://www.dti.gov.ph/advisories/artificial-intelligence-opportunities-challenges-phindustries/
1273
AskPriva, https://www.privacy.gov.ph/askpriva/
1270
326
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
how regulation “will protect human rights and put into place equitable AIactivating opportunities.”1274
In the same speech, Lopez noted that the roadmap itself is
“manifestation”1275 of the Philippine Innovation Act, which articulates and
mandates commitments to “innovation efforts to help the poor and the
marginalized, enable micro, small and medium enterprises (MSMEs) to be
a part of the domestic and global supply chain, and catalyze the growth of
Philippine industries and local economies.”1276 To this end, the roadmap
itself includes plans to:1277
• Build a national data center (NDC) with a reliable and robust data
infrastructure and data management system
• Encourage government agencies, research institutions, top
universities, and big state universities to main their own data
centers linked to the NDC
• Make internet accessible and affordable and improve its quality
• Promote data literacy for all
• Identify jobs that are vulnerable to automation and other Industry
4.0 technologies and map the skills that need upgrading or
retooling
Universal Declaration of Human Rights
Philippines has adopted the Universal Declaration of Human Rights,
having taken part in its formulation in 2008.1278 However, its human rights
implementation record is highly contested. The country is ranked “Partly
Free” on the Freedom House Global Freedom Scores Index, earning a
Department of Trade and Industry, Keynote Speech of Secretary Ramon M. Lopez,
National Artificial Intelligence Roadmap Launch, 5 May 2021,
https://www.dti.gov.ph/archives/national-artificial-intelligence-roadmap-launch/
1275
Ibid.
1276
National Economic and Development Authority, JOINT NEDA-DOST-DTI
ADMINISTRATIVE ORDER NO. 01 SERIES OF 2020, 19 February 2020,
https://neda.gov.ph/the-philippine-innovationact/#:~:text=11293%20otherwise%20known%20as%20the,signed%20by%20President%
20Rodrigo%20R.&text=The%20law%20mandates%20the%20creation,in%20the%20cou
ntry's%20innovation%20governance
1277
Philippines’ Department of Trade and Industry, ARTIFICIAL INTELLIGENCE
ROADMAP, http://innovate.dti.gov.ph/wp-content/uploads/2021/05/National-AIStrategy-infographic-2048x1105.jpg
1278
Secretary of Foreign Affairs Alberto G. Romulo, Behind a Common Cause:
Advancing with Resolve, Finding Strength in Synergy, March 2008,
https://www.ohchr.org/Documents/HRBodies/HRCouncil/RegularSession/Session7/HLS/
Philippines-E.pdf
1274
327
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
middling score of 56/100.1279 The Index explains that the “rule of law and
application of justice are haphazard and heavily favor political and
economic elites” in the country.1280 Moreover, a 2020 Amnesty
international report found instances of harassment, detention, unjust
charges, and extrajudicial executions of human rights defenders and
political activists, as well as other human rights violations, and “severe”
curtailing of media freedom.1281
Data Privacy Law
The Philippines the Data Privacy Act of 2012 (DPA) “to protect the
fundamental human right of privacy, of communication while ensuring free
flow of information to promote innovation and growth.” The Act also
ensures that “personal information in information and communications
systems in the government and in the private sector are secured and
protected.”1282
The Act mandated the creation of the National Privacy Commission,
an independent body, “To administer and implement the provisions of [the
Data Privacy Act of 2012], and to monitor and ensure compliance of the
country with international standards set for data protection.”1283 It was
established in March 2016, and is required to rule-make, provide advice and
guidance, and public education, in addition to its compliance, monitoring
and enforcement duties.1284 Its mission is to:
1) Be the authority on data privacy and protection, providing
knowledge, know-how, and relevant technology.
2) Establish a regulatory environment that ensures accountability in
the processing of personal data and promotes global standards for
data privacy and protection.
3) Build a culture of privacy, through people empowerment, that
enables and upholds the right to privacy and supports free flow of
information.
Freedom House, Philippines, 2021,
https://freedomhouse.org/country/philippines/freedom-world/2021
1280
Ibid.
1281
Amensty International, Philippines, https://www.amnesty.org/en/location/asia-andthe-pacific/south-east-asia-and-the-pacific/philippines/
1282
Republic Act 10173, Data Privacy Act of 2012, Section 1,
https://www.privacy.gov.ph/data-privacy-act/#1
1283
Republic Act 10173, Data Privacy Act of 2012, Section 7,
https://www.privacy.gov.ph/data-privacy-act/#7
1284
Republic Act 10173, Data Privacy Act of 2012, Section 1,
https://www.privacy.gov.ph/data-privacy-act/#1
1279
328
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Commission appears to be active. According to the Commission, by
2018, the Commission had “a total of 23,081 registered Data Protection
Officers (DPOs). The number of privacy-related cases the agency received
so far, [in 2018] has also increased by 145% from 2017. Of the 542 cases
in 2018, 35.52% involved unauthorized processing while 36.44% were on
data breaches.”1285
The National Privacy Commission, Philippines’ independent data
protection enforcement agency, is a cosponsor to the Global Privacy
Assembly (GPA)’s 2018 Resolution on AI and Ethics and its 2020 GPA
Resolution on AI and Accountability. Through its co-sponsorship, the
Commission endorses accountability, human rights, transparency,
auditability, security, and multi-stakeholder discussions — largely in
accordance with OECD AI principles. It is unclear how these principles
have been implemented.
Algorithmic Transparency
The Philippine’s national AI policies do not explicitly promulgate
any commitments to transparency, but its Data Privacy Act mandates
“adherence to the principles of transparency, legitimate purpose and
proportionality.”1286 It also includes provisions related to transparent
automated processing. Notably, Section 16 furnishes citizens have the right
to be informed whether their personal information pertaining is being or has
been processed, with the right to information regarding the “[m]ethods
utilized for automated access, if the same is allowed by the data subject, and
the extent to which such access is authorized,” and the right “to access,
correction, as well as the right to lodge a complaint before the
Commission.”1287 These are all elements of transparency, which the
National Privacy Commission has formally endorsed through its cosponsorship of the 2020 and 2018 Global Privacy Assembly resolutions
noted above.
Facial/Biometric Recognition
The government is in the process of registering 105 million citizens,
including infants, in its biometric national ID system, PhilSys, which is to
National Privacy Commission, NPC Eyes fully digital PH by 2040, (Sept. 19, 2018),
https://www.privacy.gov.ph/2018/09/npc-eyes-fully-digital-ph-by-2040/
1286
Republic Act 10173, Data Privacy Act of 2012, Section 11,
https://www.privacy.gov.ph/data-privacy-act/#7
1287
Republic Act 10173, Data Privacy Act of 2012, Section 16,
https://www.privacy.gov.ph/data-privacy-act/#7
1285
329
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
include facial, iris, and thumbprint biometrics.1288 Law enforcement in the
Philippines has also implemented and encouraged the use of real-time facial
recognition software to identify “persons with Warrants of Arrest, High
Value Targets and members of communist terror groups evading law
enforcers.”1289 According to Crizaldo Nieves, the regional director of the
Cagayan Valley Police, all police will eventually have smartphones that
enable real-time tracking and “appropriate awards [will be given] to the first
10 police stations that will effect an arrest through this technology.”1290
Evaluation
The Philippines continues to make strides in national AI policy,
being among the 193 countries to endorse UNESCO’s Recommendations
on Ethics of AI.1291 AI is in use across a number of initiatives in the country,
including law enforcement,1292 healthcare,1293 autonomous vehicles,
infrastructure, and data processing.1294 The active work of the Philippines
National Privacy Commission on both domestic issues and at the Global
Privacy Assembly contributes to the country’s overall favorable score.
Congress of the Philippines, Republic Act No. 11055, Section 5(c), (July 24, 2017),
https://psa.gov.ph/system/files/kmcd/RA11055_PhilSys.pdf
1289
Artemio Dumlao, Cagayan Valley cops roll out facial recognition system vs crime,
21 October 2020,
https://www.philstar.com/nation/2020/10/21/2051265/cagayan-valley-cops-roll-outfacial-recognition-system-vs-crime
1290
Ibid.
1291
Philippine News Agency, 193 countries adopt 1st global agreement on Ethics of AI,
(Nov, 26, 2021), https://www.pna.gov.ph/articles/1161070
1292
Artemio Dumlao, Cagayan Valley cops roll out facial recognition system vs crime,
(Oct. 21, 2020), https://www.philstar.com/nation/2020/10/21/2051265/cagayan-valleycops-roll-out-facial-recognition-system-vs-crime
1293
DOST-Science for Change Program, Launching of DOST AI Programs and
Technologies: AI for a Better Normal (June 23, 2021),
https://www.facebook.com/dost.s4cp/videos/3690780667689337/; Newsbytes, DOST
rolls out technologies for AI National Roadmap (June 24, 2021),
https://newsbytes.ph/2021/06/24/dost-rolls-out-technologies-for-ai-national-roadmap/
1294
The Philippine Council for Industry, Energy, and Emerging Technology Research
and Development, DOST-PCIEERD LAUNCH 9 NEW AI R&D PROJECTS (Apr. 8,
2021),
https://pcieerd.dost.gov.ph/news/latest-news/422-dost-pcieerd-launch-9-new-ai-r-dprojects
1288
330
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Poland
National AI Strategy
In September 2020, the Polish Council of Ministers Committee for
Digital Affairs1295 adopted the ‘Policy for the Development of Artificial
Intelligence in Poland.’1296 In December 2020, the Council of Ministers
adopted the Polish national AI strategy, entitled Policy for the development
of artificial intelligence in Poland from 2020 (Poland, 2020).1297 According
to a description of the Policy provided by the government website, it is
designed to support and complement the work of the EU and the OECD in
AI. The Policy establishes goals and actions for Poland in the field of AI in
the short-term (until 2023), medium-term (until 2027), and long-term (after
2027).1298 Six key categories are identified in the AI Policy:
• AI and society
• AI and innovative companies
• AI and science
• AI and education
• AI and international cooperation
• AI and the public sector
In particular, the Polish strategy is providing strategic guidance and
policy initiatives to develop a holistic AI ecosystem with the aim of meeting
the following objectives:
Reforming the educational system and providing lifelong
learning opportunities in AI-related fields;
Komitet Rady Ministrów do Spraw Cyfryzacjisss (KRMC). The KRMC is an
auxiliary body of the Council of Ministers and the Prime Minister. The Council of
Ministers serves as Poland’s Cabinet with the Prime Minister acting as the President of
the Council of Ministers. https://www.gov.pl/web/digitalization/council-of-ministerscommittee-for-digital-affairs.
1296
https://www.gov.pl/attachment/a8ea194c-d0ce-404e-a9ca-e007e9fbc93e. Developed
by the Ministry of Digital Affairs, Ministry of Development, Ministry of Science and
Education, Ministry of Funds and Regional Policy and Chancellery of the Prime Minister.
1297
OECD, STIP Compass, Poland's National AI Strategy (2020),
https://stip.oecd.org/stip/policy-initiatives/2019%2Fdata%2FpolicyInitiatives%2F24268.
1298
Government of Poland, The development of artificial intelligence in Poland - an
important decision (Sept. 14, 2020), https://www.gov.pl/web/cyfryzacja/rozwojsztucznej-inteligencji-w-polsce--wazna-decyzja; European Commission (AI Watch),
Poland AI Strategy Report, https://knowledge4policy.ec.europa.eu/ai-watch/poland-aistrategyreport_en#:~:text=In%20December%202020%2C%20the%20Council,2020%20(Poland
%2C%202020).&text=Reinforcing%20the%20digital%20infrastructure%2C%20regulato
ry,the%20development%20of%20AI%20innovati.
1295
331
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Encouraging growth and innovation of AI companies through
dedicated support in AI research, including the provision of
sufficient financial resources;
Increasing national and international partnerships in AI;
Creating a data ecosystem with trustworthy and high-quality
data and increased data exchange mechanisms;
The Polish Council of Ministers Committee for Digital Affairs will
steer the implementation of the strategy and evaluate its implementation on
a yearly basis.664
The Polish Government website sets out a roadmap towards
Poland’s AI strategy, with the first step towards the strategy being marked
in September 2016 and the most recent being marked in September 2020
(with the KRMC adoption of the 2020 Policy).1299 The 2020 Policy follows
on from a draft policy document that was released for consultation in
20191300 and a document released by the Minister for Digital Affairs in
2018, titled “Assumptions to AI Strategy in Poland.”1301 The Assumptions
document was the product of a group of independent experts who
volunteered their time to develop recommendations for the development of
AI in Poland under the guidance of the Ministry of Digital Affairs. The
Assumptions document identifies four key areas of importance:
•
•
•
•
data-driven economy
financing and development
education
law and ethics.
The Assumptions document states that Poland’s approach to ethical
and legal issues with AI should: be proactive in creating ethical standards
and legislation; be inclusive and cooperative; take into account the specific
circumstances in Poland; be flexible; instate consistent supportive
measures; engage in discussion and consultation; and be firm in response to
violations of ethical and legal standards. The primary goals are asserted to
be the development of transparent and effective mechanisms ensuring the
Government of Poland, Digitalization of the Chancellery of the Prime Minister [GT],
https://www.gov.pl/web/cyfryzacja/ai.
1300
https://www.gov.pl/attachment/a8ea194c-d0ce-404e-a9ca-e007e9fbc93e. The
document is only available in Polish and machine translation was used in combination
with other sources.
1301
Ministry for Digital Affairs, Assumptions for the AI Strategy in Poland (Nov, 9,
2018) [GT], https://www.gov.pl/attachment/1a3fba75-c9f9-4aff-96d8-aa65ce612eab
1299
332
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
protection of fundamental rights, gaining understanding of the social effects
of AI, the setting of ethical standards, and the creation of high-quality
legislation.
Fundamental rights and values identified in the Assumptions
document as being important to the development of a legal and ethical
approach in Poland include: dignity; freedom (described as including
freedom to understand processes with which individuals interact and the
making of free and independent decisions); privacy and data protection;
equality; and justice. The Assumptions document calls for the development
of an ethical impact assessment mechanism. Where AI projects are
supported by public funds, the Assumptions document calls for ethical
impact assessments at various stages of a project’s implementation, not just
at the application stage. As part of the objective of coordinating national
efforts in this space, the Assumptions document also proposes the
establishment of an entity designed to, amongst other things: monitor the
social impact of AI; recommend regulatory actions; participate in the
development of regulations and ethical standards; and perform ethical
impact assessments on publicly funded projects. It is envisioned that such
an entity would include representatives from science, government, business,
and NGOs. The legal analysis in the report was prepared by experts invited
by the Ministry of Digital Affairs to consider the legal aspects of AI as part
of the working group on the legal and ethical aspects of AI under the
guidance of the Ministry. The research was preliminary and based on the
presentation of selected legal issues by individual experts based on their
personal views. As a result, the recommendations primarily call for the
conducting of more research and discussion.
Subsequently, a Draft Policy for the Development of Artificial
Intelligence in Poland for the years 2019–2027 was released for
consultation in August 2019. The Draft Policy document was designed to
open a national debate from which a national strategy could be built.1302 The
Draft Policy states its goals to include the supporting of AI research and
development for the benefit of economic growth and innovation. In tandem
with this, the Policy describes the necessity to support citizens in the face
of transformations to the working environment and to protect human dignity
and fair competition.1303 The Draft Policy is designed to be coherent with
EU and OECD policies on AI. Strategic documents the Draft Policy takes
into account include: the EU Communication’s Coordinated Plan on
European Commission, Poland AI Strategy Report (Feb. 2020),
https://ec.europa.eu/knowledge4policy/ai-watch/poland-ai-strategy-report_en.
1303
https://www.gov.pl/attachment/a8ea194c-d0ce-404e-a9ca-e007e9fbc93e.
1302
333
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Artificial Intelligence;1304 the High-Level Expert Group on AI’s Ethics
Guidelines for Trustworthy Artificial Intelligence;1305 the High-Level
Expert Group on AI’s Policy and Investment Recommendations for
Trustworthy Artificial Intelligence;1306 and the OECD Council
Recommendation on Artificial Intelligence.1307
Supportive of human-centric AI, Poland rejects the idea of granting
legal personality to AI. The Draft Policy identifies the need to address
intellectual property issues with AI and to develop international consensus
around AI and liability. The Draft Policy states that Article 30 of the Polish
Constitution – which protects the inherent and inalienable dignity of the
person – forms the basis of its approach. While the Draft Policy recognizes
the importance of the Charter of Fundamental Rights and international
human rights treaties in providing a foundation for ethical principles, the
Policy suggests that Article 30 of the Polish Constitution provides broader
protection. In addition to supporting the OECD recommendations, the Draft
Policy asserts that the ethical development of AI should be based on the
European concept of Trustworthy AI.
The Virtual Chair of Ethics and Law criticized the Draft Policy
document and recommended significant changes.1308 In addition to
criticizing the structure of the Draft Policy, the Virtual Chair of Ethics and
Law called for increased detail regarding strategic goals and objectives and
the legal acts required to implement policy.
In response to the EU Commission’s White Paper on Artificial
Intelligence, Poland professed to sharing the Commission's view on the
need to define a clear European regulatory framework that would contribute
to building confidence in the AI among consumers and businesses, thereby
accelerating the spread of this technology, while ensuring socially,
European Commission, Coordinated Plan on Artificial Intelligence (Dec. 7, 2018),
https://ec.europa.eu/transparency/regdoc/rep/1/2018/EN/COM-2018-795-F1-EN-MAINPART-1.PDF.
1305
European Commission, Ethics Guidelines for Trustworthy AI (Apr. 8, 2019),
https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai.
1306
European Commission, Policy and investment recommendations for trustworthy
Artificial Intelligence (June 26, 2019), https://ec.europa.eu/digital-singlemarket/en/news/policy-and-investment-recommendations-trustworthy-artificialintelligence
1307
OECD Legal Instruments, Recommendation of the Council on Artificial Intelligence
(May 21, 2019), https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0449.
1308
The Council of the Virtual Chair of Ethics and Law, Comments on the Policy for the
Development of Artificial Intelligence in Poland for the years 2019 - 2027 (Nov. 11,
2019), http://cpptint.wpia.uni.opole.pl/rada-wirtualnej-katedry-etyki-i-prawa-zabieraglos-w-sprawie-ai/.
1304
334
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
environmentally, and economically optimal results and compliance with
EU’s laws, principles, and values. However, Poland suggests limiting
regulatory action “only to the areas of necessary intervention that promote
legal certainty and relations, ensure coordination within the EU, and limit
the negative social impact.”1309 Poland states that the regulations “should be
sufficiently effective to achieve their objectives, but should not be overly
prescriptive, as this could lead to disproportionate burdens, especially for
SMEs and MSMEs.” Poland endorsed incentives for voluntary ex ante
control rather than mandatory. In particular, Poland cautioned against the
imposition of mandatory certification regimes.
Ultima Ratio
Poland has begun an online arbitration court which incorporates AI
techniques. Ultimately, according to Polish Science, “artificial intelligence
will automatically prepare a ready draft judgment together with
justification, processing for this purpose the data and positions of the parties
collected in the course of the proceedings.”1310 The first modules began in
2020. Ultimately, artificial intelligence will automatically prepare a ready
draft judgment together with justification, processing for this purpose the
data and positions of the parties collected in the course of the proceedings.
The Ultima Ratio judgment has the same legal force as a decision before a
common court.1311 The main legal issue raised by the use of Ultima Ratio is
whether it is compatible with Article 47 of the EU Charter on Fundamental
Rights which guarantees the right to a fair trial before an independent and
impartial court.
Poland’s Position on AI and Fundamental Rights
Earlier this year, the Presidency of the Council of the EU failed to
secure unanimous support from the Member States for its conclusions on
the application of the Charter of Fundamental Rights in the AI context.1312
Government of Poland, Poland’s position in the consultations on the White Paper on
Artificial Intelligence - a European approach to excellence and trust (June 12, 2020),
https://www.gov.pl/attachment/583eb32c-7344-4317-b607-fee0532c3eeb.
1310
Polish Science, Ultima Ratio- the first online court of arbitration in Poland to
implement Artificial Intelligence (Jan. 2, 2020), http://polishscience.pl/en/ultima-ratiothe-first-online-court-of-arbitration-in-poland-to-implement-artificial-intelligence/
1311
Warsaw Business Journal, Online arbitration as remedy for closed common courts
and pandemics (March 16, 2020), https://wbj.pl/online-arbitration-as-remedy-for-closedcommon-courts-and-pandemics/post/126416
1312
European Council, Artificial intelligence: Presidency issues conclusions on ensuring
respect for fundamental rights (Oct. 21, 2020),
https://www.consilium.europa.eu/en/press/press-releases/2020/10/21/artificial1309
335
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Poland objected to the inclusion of “gender equality.” Poland was the only
member state in the European Union to oppose the resolution on AI and
fundamental rights.1313 Although Ambassador Andrzej Saros said that
Poland will work to support the conclusions in the future, he also stated that:
“The Treaties refer to equality between women and men, similar to the
Charter of Fundamental Rights. The meaning of ‘gender’ is unclear; the lack
of definition and unambiguous understanding for all member states may
cause semantic problems. Neither the Treaties nor the Charter of
Fundamental Rights use the term ‘gender’.” 1314
The position occurs in the context of the Polish government
opposition to the Convention on Preventing and Combating Violence
against Women and Domestic Violence.1315 While consensus was not
reached as regards the Presidency’s conclusions, the Presidency stressed
that the core elements of the conclusions, anchoring the Union’s
fundamental rights and values in the age of digitalization, fostering the EU’s
digital sovereignty and actively participating in the global debate on the use
of artificial intelligence with a view to shaping the international framework,
were shared by all delegations.1316
OECD/G20 AI Principles
As a member of the OECD, Poland is committed to the OECD
Principles on Artificial Intelligence and references the OECD principles in
its Draft Policy document. The OECD did not identify any instances of
implementation of the AI Principles in Poland in the 2020 survey.1317
The Global Partnership on AI (GPAI) is an international and multistakeholder initiative, conceived by Canada and France during their
respective 2018 and 2019 G7 presidencies, to undertake cutting-edge
intelligence-presidency-issues-conclusions-on-ensuring-respect-for-fundamental-rights/;
https://www.consilium.europa.eu/media/46496/st11481-en20.pdf.
1313
Warsaw Business Journal, Poland rejects artificial intelligence because of gender
(Oct. 29, 2020), https://wbj.pl/poland-rejects-artificial-intelligence-because-ofgender/post/128788
1314
Samuel Stolton, Poland rejects Presidency conclusions on Artificial Intelligence,
rights, Euroactiv, Oct. 26, 2020, https://www.euractiv.com/section/digital/news/polandrejects-presidency-conclusions-on-artificial-intelligence-rights/.
1315
Eline Schaart, Poland to withdraw from treaty on violence against women, Politico
(July 25, 2020), https://www.politico.eu/article/poland-to-withdraw-from-istanbulconvention-treaty-on-violence-against-women/
1316
Council of the European Union, Presidency conclusions - The Charter of
Fundamental Rights in the context of Artificial Intelligence and Digital Change (Oct. 21,
2020), https://www.consilium.europa.eu/media/46496/st11481-en20.pdf.
1317
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
336
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
research and pilot projects on AI priorities to advance the responsible
development and use of AI. 1318 In December 2020, Poland joined GPAI.1319
Human Rights
Poland is a member of the European Union and Council of Europe
and is, accordingly, committed to the upholding of the Charter of
Fundamental Rights and the European Convention on Human Rights. In
addition, Poland has acceded to international human rights treaties and has
signed the Universal Declaration of Human Rights. The Polish Constitution
also grants basic rights to citizens and includes prohibitions against
discrimination.1320 While Poland ranks highly in the Freedom House 2021
Country Report (82/100), down slightly from 2020, concerns were raised
about the fairness of parliamentary elections, media freedom, judicial
reforms, and LGBT+ rights.1321 Since taking power in late 2015, the
populist, socially conservative Law and Justice (PiS) party has enacted
numerous measures that increase political influence over state institutions
and threaten to reverse Poland’s democratic progress. Poland's ranking on
the World Press Freedom Index has dropped from 18th to 62nd place since
2015. 686
Algorithmic Transparency
As a member of the European Union, Poland is committed to the
protection of personal data as required by Article 8 of the Charter of
Fundamental Rights and the data protection laws of the EU. The Personal
Data Protection Act 2018 and the GDPR Implementation Act 2019 were
enacted in order to adapt the GDPR and to implement the Law Enforcement
Directive into domestic law. The Polish supervisory authority is the
President of the Office of Personal Data Protection.1322 Poland supports the
Ethics Guidelines for Trustworthy AI, including the requirements of human
OECD (2021), State of implementation of the OECD AI Principles: Insights
from national AI policies (Jun 18, 2021), https://doi.org/10.1787/1cd40c44-en.
1319
Digital Affairs – Chancellery of the Minister, Poland joins the Global Partnership on
Artificial Intelligence (Dec. 8, 2020), https://www.gov.pl/web/digitalization/poland-joinsthe-global-partnership-on-artificial-intelligence
1320
The Constitution of the Republic of Poland, Article 32,
https://www.sejm.gov.pl/prawo/konst/angielski/kon1.htm
1321
Freedom House, Freedom in the World 2021 – Poland (2021),
https://freedomhouse.org/country/poland/freedom-world/2021. See also Case C-791/19 R
Commission v Poland and Case C-619/18 Commission v Poland.
1322
President of the Office of Personal Date Protection, https://uodo.gov.pl/en.
1318
337
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
agency and oversight; privacy and data governance; transparency; and
diversity, non-discrimination and fairness; and accountability.1323
In the Polish response to the EU Commission’s White Paper on
Artificial Intelligence, human oversight of AI systems is identified as the
key guiding principle for all AI applications, not just for high-risk
examples.1324 Poland supports the introduction of an ex-ante conformity
assessment procedure for certain high-risk AI applications. The 2019 Draft
Policy recognizes the importance of transparent, accountable, and impartial
AI and endorses the use of voluntary standards systems for the certification
of AI. The Draft Policy recommends the establishment of regulatory
sandboxes to enable the early testing of AI systems before they meet
compliance or certification standards and supports the mutual recognition
of interoperability standards.1325
In 2014, a profiling system was introduced in order to divide
unemployed people into three categories based on their responses to a series
of questions asked during a computer-based interview.1326 In a report on the
issue, the NGO, Panoptykon, described the process as the computer system
calculating the ‘employment potential’ of a given person on the basis of the
provided answers.1327 The amount of assistance the individual received was
determined by their categorization. The Polish data protection supervisory
authority expressed reservations regarding the use of profiling in this
Digitalization of the Chancellery of the Prime Minister, Public consultations on the
project Policy for the Development of Artificial Intelligence in Poland for 2019-2027
(Aug. 21, 2019), https://www.gov.pl/web/cyfryzacja/konsultacje-spoleczne-projektupolityki-rozwoju-sztucznej-inteligencji-w-polsce-na-lata-2019--2027; European
Commission, Ethics guidelines for trustworthy AI (Apr. 8, 2019),
https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai.
1324
Government of Poland, Poland’s position in the consultations on the White Paper on
Artificial Intelligence - a European approach to excellence and trust (June 12, 2020),
https://www.gov.pl/attachment/583eb32c-7344-4317-b607-fee0532c3eeb.
1325
Government of Poland, Konsultacje społeczne projektu „Polityki Rozwoju Sztucznej
Inteligencji w Polsce na lata 2019 – 2027,” (Aug. 21, 2019),
https://www.gov.pl/web/cyfryzacja/konsultacje-spoleczne-projektu-polityki-rozwojusztucznej-inteligencji-w-polsce-na-lata-2019--2027.
1326
Amendment to the Act on the Promotion of Employment and Labor Market
Institutions and Ordinance on the Profiling of Assistance for the Unemployed;
https://panoptykon.org/sites/default/files/leadimagebiblioteka/panoptykon_profiling_report_final.pdf.
1327
Amendment to the Act on the Promotion of Employment and Labor Market
Institutions and ordinance on the profiling of assistance for the unemployed;
https://panoptykon.org/sites/default/files/leadimagebiblioteka/panoptykon_profiling_report_final.pdf.
1323
338
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
context.1328 In particular, concerns were expressed regarding the protection
of personal data and the absence of a transparent procedure to facilitate
appeals.1329 In addition, concerns were raised by the Polish Supreme Audit
Office (Najwyższa Izba Kontroli) and the Human Rights Commissioner.
Eventually, Poland’s Constitutional Court found the system to be a breach
of the Polish Constitution.1330 The system was abolished by December
2019.
In taking steps to implement the GDPR in 2019, Poland provided all
banking customers with the right to an explanation regarding their credit
assessment when applying for a loan.1331 STIR – System Teleinformatyczny
Izby Rozliczeniowej – is a government tool that analyses information
collected by financial institutions to detect illegal activity. If suspicion
arises, the financial institution can block a flagged account for 72 hours at
the request of the tax authorities.1332 The algorithms behind the system are
not publicly available and a criminal offense – with a maximum prison
sentence of up to five years – exists prohibiting the disclosure of relevant
information.
Data Protection
Regarding data retention, the UODO took the position that a
controller should delete the data right after a relationship with a data subject
ends. As a result, the controller cannot claim it has a legitimate interest to
keep the data later on for the purposes of exercising or defending legal
claims. The DPA stated if the controller cannot prove it has good reasons to
See, for example,
https://archiwum.mpips.gov.pl/download/gfx/mpips/pl/defaultopisy/8216/1/1/Uwagi%20
GIODO-IV.pdf; https://panoptykon.org/sites/default/files/leadimagebiblioteka/panoptykon_profiling_report_final.pdf.
1329
Fundacja Panoptykon, Profiling the Unemployed in Poland: Social and Political
Implications of Algorithmic Decision Making (2015),
https://panoptykon.org/sites/default/files/leadimagebiblioteka/panoptykon_profiling_report_final.pdf. See also
https://legislacja.rcl.gov.pl/docs//2/171820/171829/171833/dokument89898.pdf.
1330
AlgorithmWatch, Poland: Government to scrap controversial unemployment scoring
system (Apr. 16, 2019). https://algorithmwatch.org/en/story/poland-government-to-scrapcontroversial-unemployment-scoring-system/.
1331
Article 46 of the GDPR,
http://orka.sejm.gov.pl/opinie8.nsf/nazwa/3050_u/%24file/3050_u.pdf; Panoptykon
Foundation, The right to explanation of creditworthiness assessment – first such law in
Europe (June 12, 2019), https://en.panoptykon.org/right-to-explanation.
1332
Algorithm Watch, Automating Society 2020 (October 2020),
https://automatingsociety.algorithmwatch.org/report2020/poland/; Government of
Poland, Sukces analityki STIR (Jan. 15, 2020), https://www.gov.pl/web/kas/sukcesanalityki-stir.
1328
339
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
believe such claims will be raised, the controller should not store the data
for the purpose of potential future claims. Polish courts overturned such
decisions, stating a controller cannot predict if and when a data subject may
raise claims. But this does not exclude such claims being raised, and the
controller has a right to store data to defend itself or exercise its claims.
Regarding notification of data breaches, the UODO issued decisions
where data controllers were found responsible for mail lost by professional
mail deliverers (such as Polish post or private couriers). It will be interesting
to observe how the situation develops and whether controllers will be
obliged to control more entities that act on their own.1333
Public Participation
Documents relating to Poland’s development of its AI policy are
accessible on the internet. The process that led to the ‘Assumptions to AI
Strategy in Poland’ document involved the participation of a broad range of
representatives of science, business, social organizations and public
administration.1334 The subsequently published Draft Policy document was
released for public consultation in August 2019 (the consultation period
closed in September 2019).1335 According to the government website, 46
entities took part in the consultation.
Evaluation
Poland has yet to release its official national policy for the
development of Artificial Intelligence. The precursor documents, including
the draft policy, address the legal and ethical implications of AI, but it is
difficult to predict what form the final policy will take. As a member of the
EU, the Council of Europe, and the OECD, Poland has made commitments
to upholding human rights and ethics in and endorsed the OECD AI
Principles. Despite these commitments, Poland opposed the Council of
Europe’s Resolution on AI and fundamental rights. Also of concern is the
prospect of the administration of justice by opaque AI techniques.
The International Association of Privacy Professionals (IAPP), 2022 Global
Legislative Predictions,
https://iapp.org/media/pdf/resource_center/2022_iapp_global_legislative_predictions.pdf.
1334
Digitization of the Chancery of the Prime Minister, Artificial Intelligence - Poland
2118 (Nov. 9, 2018) [GT], https://www.gov.pl/web/cyfryzacja/sztuczna-inteligencjapolska-2118.
1335
Digitization of the Chancery of the Prime Minister, Public consultations on the
project "Policy for the Development of Artificial Intelligence in Poland for 2019-2027"
(Aug. 21, 2019), [GT], https://www.gov.pl/web/cyfryzacja/konsultacje-spoleczneprojektu-polityki-rozwoju-sztucznej-inteligencji-w-polsce-na-lata-2019--2027.
1333
340
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Russia
National AI Strategy
Russian president Vladimir Putin famously said, in a 2017 address
to students in Moscow, “Artificial intelligence is the future not only of
Russia but of all of mankind. There are huge opportunities, but also threats
that are difficult to foresee today. Whoever becomes the leader in this sphere
will become the ruler of the world.”1336 Putin then stated that it is better to
avoid a monopoly on the sector and promised that if Russia became the
leader in developing AI, then Russia will share their technology with the
rest of the world, just as they share their atomic and nuclear technology
today.
Russia's national strategy for Artificial Intelligence (AI) was
announced in October 2019.1337 This strategy defines the goals and primary
objectives of the development of artificial intelligence in the Russian
Federation, as well as the measures aimed at its use for the purpose of
protecting national interests and implementing strategic national priorities,
including those in the field of scientific and technological development.
The goals of the development of AI in the Russian Federation
include the improvement of the well-being and quality of life of its
population, national security and rule of law, and sustainable
competitiveness of the Russian economy, including leading positions the
world over in the field of AI. The primary objectives of the Russian
development of AI are to support scientific AI research, engineering AI
software development, data quality, hardware availability, qualified
personnel and integrated system to extend Russian artificial intelligence
technology market.
In the strategy, the basic principles of the development and use of
artificial intelligence technologies include the protection of human rights
and liberties, security, transparency, technological sovereignty, innovation
cycle integrity, reasonable thrift, and support for competition in the field of
artificial intelligence.
The use of AI technologies in sectors of the economy supports the
efficiency of planning, forecasting, and management decision-making
processes; the automation of routine production operations; the use of selfCNN, Who Vladimir Putin thinks will rule the world (Sept. 2, 2017), Who Vladimir
Putin thinks will rule the world
1337
Decree of the President of the Russian Federation on the Development of Artificial
Intelligence in the Russian Federation (Oct. 10, 2019), https://cset.georgetown.edu/wpcontent/uploads/Decree-of-the-President-of-the-Russian-Federation-on-the-Developmentof-Artificial-Intelligence-in-the-Russian-Federation-.pdf;
http://www.kremlin.ru/acts/bank/44731 (in Russian)
1336
341
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
contained intelligent equipment, robotic systems, and intelligent logistic
management systems; the improvement of employee safety during the
performance of business processes; an increase in the loyalty and
satisfaction of customers, and; the optimization of the personnel selection
and training processes.
The use of AI technologies in the social sphere facilitates the
creation of conditions that favor the improvement of the standard of living
of the population including an increase in the quality of healthcare services;
the improvement of the quality of education services, and; the improvement
of the quality of the provision of public and municipal services, as well as
the reduction of the cost of their provision.
Digital Economy in Russia
The Russian government has put a high priority and already
achieved some remarkable accomplishments in the Digital Transformation.
A key strategic objective formulated by its leadership in the May 2018
Presidential Decree (The Decree on the National Goals and Strategic
National Development Tasks of the Russian Federation until 2024)1338 is
that policymakers must build on the country's traditional industrial
strengths, develop new technology processes for fast implementation in all
the main competitive domains and continuously tackle any obstacles.1339
AI Strategy for Russian start-up
Russia aims to increase the start-up ecosystem and many companies
have been helped by the traditional hard science education in the
country.1340 This report breaks down the importance of Artificial
Intelligence in Russian startups, with a large number of startups active in
AI as a logical result of Russia's big AI talent pool, taking over 16% of the
market. Moreover, the report provides case studies of the top AI industries
in Russia, along with the application of AI in Russia and how Russia has
moved up to the ladder, aiming to be a leading global startup venue by 2030.
The President signed Executive Order on National Goals and Strategic Objectives
of the Russian Federation through to 2024,
http://en.kremlin.ru/events/president/news/57425
1339
World Bank Group, Competing in Digital Age: Policy Implications for the Russian
Federation (Sept. 2018),
http://documents1.worldbank.org/curated/en/860291539115402187/pdf/Competing-inthe-Digital-Age-Policy-Implications-for-the-Russian-Federation-Russia-DigitalEconomy-Report.pdf
1340
GMIS, Artificial Intelligence: A Strategy for Russian start-up (June 11, 2019),
https://gmisummit.com/wp-content/uploads/2019/06/Ai-A-startegy-for-russianstartups.pdf
1338
342
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Russia has enormous potential for start-up development. The most
appropriate and easy way to implement new start-ups in the Russian
Federation is to become an individual entrepreneur. Citizens may
participate in entrepreneurship without having to set up a legal entity once
they are registered as an individual contractor. The simple method for
registering and controlling the activity of individual contractors is the
perfect start-up for fresh participants in the market.1341
According to A.T. Kearney, there are approximately 1,000 digital
start-ups in Russia every year. Market specialists estimate that
approximately half of these are promising projects, about 20% of which are
safe in Russia from risk capitalists, company angels, friends and family.
Another 1 to 2% go abroad for funding. Of the approximately 100 Russianfunded start-ups that are still in the country, around 50% are viable and 10
have become extremely successful, prosperous and increasing firms.
AI Policy in Russia
In January 2019, Russian President Vladimir Putin had approved
a list of instructions1342 following the meeting of the supervisory board of
the Agency for Strategic Initiatives, which included the instruction to the
Russian government to create a national AI strategy. A draft version of a
national AI strategy, developed by the country’s largest bank – Sberbank,
was announced September 2019.1343
According to the Future of Life Institute,1344 several projects helped
pave the way for a domestic approach throughout 2018. In March 2018, for
example, a conference was organised by the Russian Defense Ministry,
Education and Science Ministry and the Academy for Science on AI issues
and alternatives and a 10-point AI development plan in Russia was
subsequently published. The plan involves the establishment of an AI and
Big Data Consortium among academic and industrial organisations; The
development of a fund to assist provide knowledge on automated systems;
Increased state aid to AI education and training; The establishment of an AI
Academy of Strategic Management Journal, The Development if Innovative Startups
in Russia: The Regional Aspect (2017), https://www.abacademies.org/articles/thedevelopment-of-innovative-startups-in-russia-the-regional-aspect-1939-6104-16-SI-1117.pdf
1342
List of instructions following the meeting of the Supervisory Board of the Agency for
Strategic Initiatives [GT], http://kremlin.ru/acts/assignments/orders/59758
1341
First draft of Russian AI strategy,
https://www.defenseone.com/technology/2019/09/whats-russias-national-aistrategy/159740/
1344
Future of Life, AI Policy – Russia (Feb. 2020), https://futureoflife.org/ai-policyrussia/
1343
343
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
laboratory at the leading technological university; The establishment of a
national AI R&D center.
Digital Rights Law and AI Regulation
According to the OECD, Russia’s Digital Rights Law, which came
into force in October 2019 and introduced several new legal concepts,
including digital rights, e-transactions, smart contracts, and big data.1345 The
law aimed to enable the development of an efficient legal framework of
digital economy in Russia, reflecting current digital technologies and
challenges including big data and AI.
Russia’s draft legal framework AI Technologies and Robotics aims
to establish a legal framework for the development of AI technologies and
robotics in Russia and eliminate excessive legal barriers. The initiative aims
to give guidance for regulators and is under the responsibility of the
Ministry of Economic Development.1346
Facial Recognition Controversy
Russia is moving rapidly to deploy AI-based face surveillance
across the country, often with government funding that goes to business
associates of President Putin. According to the Moscow Times, more than
43,000 Russian schools will be equipped with facial recognition cameras
ominously named “Orwell.”1347 The system will be integrated with face
recognition developed by NTechLab, a subsidiary of Russian President
Vladimir Putin’s associate Sergei Chemezov’s Rostec conglomerate.
NTechLab has already deployed facial recognition technology in Moscow
to identify criminal suspects across a network of almost 200,000
surveillance cameras. “Critics have accused the technology of violating
citizens' privacy and have staged protests against the system by painting
their faces,” reported Moscow Times.
In September 2020, Kommersant daily reported that CCTV cameras
with facial recognition software, already used in Moscow, will be installed
Government of Russia, Official Internet Portal for Legal Information, Federal Law of
18.03.2019 No. 34-FZ "On Amendments to Parts One, Two and Article 1124 of Part
Three of the Civil Code of the Russian Federation [GT],
http://publication.pravo.gov.ru/Document/View/0001201903180027
1345
1346
http://sk.ru/foundation/legal/p/11.aspx;https://economy.gov.ru/material/directions/gosuda
rstvennoe_upravlenie/normativnoe_regulirovanie_cifrovoy_sredy/regulirovanie_primene
niya_tehnologiy_iskusstvennogo_intellekta/
1347
Moscow Times, Russia to Install ‘Orwell’ Facial Recognition Tech in Every School –
Vedomosti (June 16, 2020), https://www.themoscowtimes.com/2020/06/16/russia-toinstall-orwell-facial-recognition-tech-in-every-school-vedomosti-a70585
344
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
by the regional authorities in public spaces and at the entryway of apartment
buildings in 10 pilot cities across Russia with the purported aim of
protecting public safety.1348 Moscow authorities are also planning to expand
the use of this technology, installing CCTV cameras with facial recognition
software in trams and underground trains.1349
Human Rights Watch said “The authorities’ intention to expand the
use of invasive technology across the country causes serious concern over
the potential threat to privacy. Russia’s track record of rights violations
means that the authorities should be prepared to answer tough questions to
prove they are not are undermining people’s rights by pretending to protect
public safety.”1350 HRW also stated that Russian national security laws and
surveillance practices enable law enforcement agencies to access practically
any data in the name of protecting public safety.
Earlier in the year, Amnesty International criticized Russia’s plans
to broaden the use of widespread facial-recognition systems, saying their
expected deployment during public gatherings will “inevitably have a
chilling effect” on protesters.1351
Data Protection
There are many laws in Russia that regulate the processing of
personal data, including the Constitution of the Russian Federation, The
Council of Europe Convention 108, and federal law.1352 The Law on
Personal Data of 2006 is the most comprehensive federal law and sets out
broad rights and responsibilities associated with the collection and use of
Kommersant, Regions will recognize by sight: Moscow video surveillance system will
be launched in ten more cities (Sept. 25, 2020) [GT],
https://www.kommersant.ru/doc/4503379
1349
Government of Russia, Unified information system in the field of procurement,
Implementation of work on equipping the rolling stock of the State Unitary Enterprise
"Moscow Metro" with video surveillance equipment (Aug. 3, 2020),
https://zakupki.gov.ru/epz/order/notice/ea44/view/documents.html?regNumber=0173200
001420000752&backUrl=89687dbf-73a1-4346-a608-3634c2a98681
1350
Human Rights Watch, Russia Expands Facial Recognition Despite Privacy Concerns
- Lack of Accountability, Oversight, Data Protection (Oct. 2, 2020),
https://www.hrw.org/news/2020/10/02/russia-expands-facial-recognition-despite-privacyconcerns
1351
Radio Free Europe, Watchdog Warns About 'Chilling Effect' Of Russia’s Use Of
Facial-Recognition Technology (Jan. 31, 2020), https://www.rferl.org/a/watchdog-warnsabout-chilling-effect-of-russia-s-use-of-facial-recognition-technology/30410014.html
1352
Constitution of the Russian Federation (Articles 23 and 24),
http://archive.government.ru/eng/gov/base/54.html
1348
345
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
personal data.1353 The Roskomnadzor, Russia’s data protection agency,
interprets the federal law and brings enforcement actions.1354
Russia is also moving to update and expand its national data
protection law.1355 A draft law on the Protection of Consumer Rights would
limit the ability of companies to collect personal data from consumers,
unless there is a legal basis or the data is necessary for the transaction. A
proposed law in the Duma would expand penalties for breach of personal
data confidentiality and infringement of personal data anonymization rules.
And the Supreme Court of Russia ruled this summer that the personal data
of a Russian citizen posted by the Whois Privacy Corporation, based in the
Bahamas, without consent is subject to legal action under the Russian Civil
Procedure Code.
Algorithmic Transparency
Russia is a member of the Council of Europe and ratified
Convention 108 regarding the automated process of personal data in
2013.1356 Russia has not yet ratified the modernized Privacy Convention,
which includes a broad provision regarding algorithm transparency.1357
Russian data protection law does broadly provide rights of access and
transparency to the data subject.1358
OECD/G20 AI Principles
Russia, a member of the G20, endorsed the G20 AI Principles at the
G20 Ministerial in 2019.1359 According to the OECD, many of the G20 AI
Principles are addressed in the Russia AI Strategy.
Human Rights
Russia is a signatory to the Universal Declaration of Human Rights.
Russia is a member of the Council of Europe and ratified the original
Roskomnadzor, Federal Law of 27 July 2006 N 152-FZ on Personal Data,
https://pd.rkn.gov.ru/authority/p146/p164/
1354
Roskomnadzor, About the Competent Authority, http://eng.pd.rkn.gov.ru
1355
Olga Novinskaya, Recent changes in personal data regulation in Russia,
International Lawyers Network (Nov. 12, 2020),
https://www.jdsupra.com/legalnews/recent-changes-in-personal-data-58095/
1356
Council of Europe, Chart of signatures and ratifications of Treaty 108: Convention
for the Protection of Individuals with regard to Automatic Processing of Personal Data
(Dec. 11, 2020), https://www.coe.int/en/web/conventions/full-list//conventions/treaty/108/signatures
1357
Article 9(1)(c).
1358
OneTrust, Russia – Data Protection Overview (Nov. 2020),
https://www.dataguidance.com/notes/russia-data-protection-overview
1359
http://www.oecd.org/digital/g20-digital-economy-ministers-meeting-july-2020.htm
1353
346
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Convention 108. However, Russia’s derogation from the Council of Europe
Convention on Human Rights remains controversial.1360 And a recent
decision from the European Court of Human Rights, Zakharov v. Russia,
found that Russia’s legislation on surveillance “does not provide for
adequate and effective guarantees against arbitrariness and the risk of
abuse.”1361
Freedom House gives Russia low marks for political rights and civil
liberties.1362 According to Freedom House, “Power in Russia’s authoritarian
political system is concentrated in the hands of President Vladimir Putin.
With loyalist security forces, a subservient judiciary, a controlled media
environment, and a legislature consisting of a ruling party and pliable
opposition factions, the Kremlin is able to manipulate elections and
suppress genuine dissent. Rampant corruption facilitates shifting links
among bureaucrats and organized crime groups.”
Evaluation
Russia’s development of a National AI Strategy, endorsement of the
G20 AI Principles, its efforts to develop laws for digital rights and
regulation for AI, as well as initiatives to involve the public in the
development of AI policy count favorably. But beyond data protection
legislation, the absence of robust measures to limit surveillance and protect
human rights, coupled with the rapid adoption of facial recognition in public
places raise concerns about the future of Russia’s AI program.
See generally Council of Europe, Derogation in Time of Emergency (Sept. 2020),
https://www.echr.coe.int/documents/fs_derogation_eng.pdf. See also Marc Rotenberg
and Eleni Kyriakides, Preserving Article 8 in Times of Crisis, in Francesca Bignami, i
(2020)
1361
European Court of Human Rights, Roman Zakharov v. Russia, No. 47143/06 (Dec. 4,
2015), https://hudoc.echr.coe.int/fre#{"itemid":["001-159324"]}; Paul De Hert and Pedro
Cristobal Bocos, Case of Roman Zakharov v. Russia: The Strasbourg follow up to the
Luxembourg Court’s Schrems judgment, Strasbourg Observers (Dec. 23, 2015),
https://strasbourgobservers.com/2015/12/23/case-of-roman-zakharov-v-russia-thestrasbourg-follow-up-to-the-luxembourg-courts-schrems-judgment/
1362
Freedom House, Freedom in the World 2020 – Russia (2020),
https://freedomhouse.org/country/russia/freedom-world/2020
1360
347
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Rwanda
National AI Strategy
Rwanda, with Vision 2050, aspires to increase the quality of life and
develop modern infrastructure 1363 by strengthening capacity, service
delivery and accountability of public institutions; increasing citizens’
participation and engagement in development; and strengthening justice and
rule of law. National Strategy for Transformation (NST1) is the vehicle for
achieving Vision 2050.1364 The government pledges to establish legal
frameworks that spur economic development and instill fairness,
transparency and accountability across institutions.1365 The Emerging
Technologies Strategy and Action Plan aims to position Rwanda as an
emerging technology testbed, solution and export hub; propel the social and
economic application of new technologies; prepare the foundations for new
technologies and protect citizens and institutions from the negative
consequences.1366
The Government of Rwanda is developing a national AI strategy to
equip government agencies and other stakeholders in the country to
empower AI developers, citizens and users, and support the beneficial and
ethical adoption of AI.1367 The government has engaged The Future Society
to support the development of Rwanda’s national artificial intelligence
strategy, along with AI ethical guidelines, and a practical implementation
Republic of Rwanda, Vision 2050 (2015)
http://www.minecofin.gov.rw/fileadmin/templates/documents/NDPR/Vision_2050/Visio
n_2050_-Full_Document.pdf
1364
Republic of Rwanda, 7 Years Government Programme: National Strategy for
Transformation (NST1)
http://www.minecofin.gov.rw/fileadmin/user_upload/MINECOFIN_Documents/NST_A5
_booklet_final_2.04.19_WEB.pdfhttp://www.minecofin.gov.rw/fileadmin/user_upload/M
INECOFIN_Documents/NST_A5_booklet_final_2.04.19_WEB.pdf
1365 Republic of Rwanda, 7 Years Government Programme: National Strategy for
Transformation (NST1)
1366
Lasry, F. Transforming Rwanda into a living Laboratory of Emerging Technologies:
MINICT and DigiCenter develop National Emerging Technology Strategy and Action
Plan. Digital Transformation Center Kigali (June 15, 2020)
https://digicenter.rw/transforming-rwanda-into-a-living-laboratory-of-emergingtechnologies/
1367
UNICEF, Policy Guidance on AI for Children: Pilot testing and case studies (02
November 2020) https://www.unicef.org/globalinsight/stories/policy-guidance-aichildren-pilot-testing-and-case-studies
1363
348
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
strategy fit for the local context.1368 For implementation, GIZ FAIR Forward
will be utilized. “FAIR Forward – Artificial Intelligence for All” is a global
initiative of German Development Cooperation, working together with
Rwanda and four other countries to lay the foundations for developing local
AI, to strengthen local skills and knowledge in AI; remove barriers of entry
to developing AI and to develop AI policy frameworks on ethical AI, data
protection and privacy.1369 FAIR Forward advocates for ethical AI that is
rooted in human rights, international norms such as accountability,
transparency of decision-making and privacy, and draws on European
experiences such as the EU General Data Protection Regulation (GDPR).
The stakeholders defined six priority areas for effective AI policy in
Rwanda. 1) 21st century skills and high AI literacy, 2) Reliable
infrastructure and compute capacity, 3) Robust Data Strategy, 4)
Trustworthy AI adoption in the public sector, 5) Widely beneficial AI
adoption in the private sector, and 6) Practical Ethical Guidelines. The result
of this partnership is The National AI Policy which is currently being
validated by the Government of Rwanda. At the same time Rwanda Utilities
Regulatory Authority (RURA) has been developing ethical guidelines for
the use of AI that are supposed to guide AI developers in Rwanda on how
to mitigate the risks and harms.1370
AI System for Identity Management
Rwanda used biometric identification for its census in 2007 to unify
all identity information under a single authority, the National Identification
Agency (NIDA), and a unique National Identity Number (NIN). This
number is now used for health, education, telecom, banking, electoral lists,
social protection programs and border crossings.1371 It also proposed to
create a country-wide DNA database to crack down on crime, raising
concerns that the data could be misused by the government and violate
The Future Society, The Development of Rwanda’s National Artificial Intelligence
Policy, (Aug. 31, 2020) https://thefuturesociety.org/2020/08/31/development-of-rwandasnational-artificial-intelligence-policy/
1369
, FAIR Forward – Artificial Intelligence for All, https://toolkitdigitalisierung.de/en/fair-forward/
1370
https://digicenter.rw/how-rwandas-ai-policy-helps-to-shape-the-evolving-aiecosystem/
1371
ID4Africa, Rwanda National ID Strategy,
https://www.id4africa.com/2019_event/presentations/PS1/5-Josephine-Mukesha-NIDARwanda.pdf
1368
349
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
international human rights laws.1372 In a country that has bitter memories of
genocide along ethnic identity lines, the right to data privacy becomes a
crucial issue. National social protection program, Ubudehe, database, was
created in 20011373 to classify wealth and identify the poorest households
using community assessments. The database is interlinked with the national
ID number. The lack of transparency on who makes the wealth
determination and how this impacts an individual’s access to opportunities
and resources and the extent of stigmatization remains questionable.
Smart Cities
Established in 2000 and revised in 2012, the aim of Rwanda Vision
2020 was to “transform Rwanda from an agrarian economy to a knowledgebased society by 2020.” Under this vision, Smart Rwanda Master Plan had
three goals: economic transformation, job creation and accountable
governance. In 2015, Rwanda adopted a National Urbanization Policy to
demonstrate how urban development can drive economic transformation.
One of the focus is to “promote quality of life, mitigation of disaster risks,
social inclusion and cultural preservation” through “digital service points
for rural settlements, smart urban agriculture projects, sensor-based
environmental data, smart and green building labs.1374 The policy requires
public engagement and open data as building blocks. The Rwanda smart
city model is centered around 3 main pillars, 9 strategic building blocks and
27 action initiatives. The three pillars are smart governance and planning;
smart and efficient services and utilities; and localized innovation for social
and eco-nomic development.
Kigali Innovation City (KIC) is the government’s flagship program
to create a hi-tech ecosystem, modelling itself on the southeast Asian citystate of Singapore. The City of Innovation is to be built as part of Africa50.
It is a 62-hectare development located in Kigali’s special economic zone
(SEZ). Main goal is to create an innovative business hub in the heart of
Africa that will include four first-rate universities, innovative agriculture,
International Association of Privacy Professionals, Rwanda announces plans for
countrywide DNA database (Mar. 26, 2019) https://iapp.org/news/a/rwanda-announcedplans-for-country-wide-dna-database/
1373
Republic of Rwanda Ministry of Health & World Health Organization, Rwanda's
Performance in Addressing Social Determinants of Health and Intersectoral Action
(2018) https://www.afro.who.int/sites/default/files/201803/Rwanda_s_Performance_in_Addressing_Social_Determinants_of_Health__and%20in
tersectoral%20action%20final%20Report.pdf
1374
UN Habitat, Smart City Rwanda Master Plan,
https://unhabitat.org/sites/default/files/documents/2019-05/rwanda_smart_citymaster_plan.pdf
1372
350
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
healthcare, technology, financial services, biotech firms and both
commercial and residential space.1375
The Smart City Masterplan was developed in 2017, with
participation of various stakeholders in Rwanda, including regulatory
bodies, local authorities, academia, civil society and the private sector. The
development was aligned with the Smart Africa Alliance Smart Sustainable
Cities Blueprint for Africa.1376
Drone Regulation
Following a successful partnership with a startup, Zipline, to deliver
blood, vaccines and other medical supplies to rural hospitals in Rwanda, the
country decided to regulate the use of drones. It entered into partnership
with World Economic Forum to draft a framework for governing drones at
scale and foster an ecosystem of unmanned aircraft systems. The
government created a performance-based regulation focused on safety
standards and is the first country in the world to implement it for all
drones.1377 This partnership also resulted in The Advanced Drone
Operations Toolkit which provides a modular approach for governments to
enable societally important and safe drone projects.1378 The country is now
looking to use these technologies to promote agricultural resilience and food
security in Rwanda. It has not called for a prohibition on fully autonomous
weapons.1379
Rwanda is working on another proof-of-concept with the World
Economic Forum to apply a framework of ten principles selected from AI
ethics and healthcare ethics and interpreted within the context of the use of
chatbots in healthcare.1380
Thelwell, K. Big Plans for Rwandan Infrastructure, The Borgen Project (Oct. 6,
2019) https://borgenproject.org/tag/kigali-innovation-city/
1376
UN Habitat, Smart City Rwanda Master Plan
1377 Russo A., Wolf H., What the world can learn from Rwanda’s approach to drones,
World Economic Forum (Jan.16, 2019) https://www.weforum.org/agenda/2019/01/whatthe-world-can-learn-from-rwandas-approach-to-drones/
1378
World Economic Forum, Advanced Drone Operations Toolkit: Accelerating the
Drone Revolution (Feb. 26, 2019) https://www.weforum.org/reports/advanced-droneoperations-toolkit-accelerating-the-drone-revolution
1379
The Campaign to Stop Killer Robots, Country Views on Killer Robots (Nov. 13,
2018) https://www.stopkillerrobots.org/wpcontent/uploads/2018/11/KRC_CountryViews13Nov2018.pdf
1380
How Rwanda Is Shaping The Global Rollout Of Everything From AI Health Bots To
Drones (February 8,2021) https://medaditus.org/news-articles/how-rwanda-is-shapingthe-global-rollout-of-everything-from-ai-health-bots-to-drones/
1375
351
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Public Participation
Ongoing policy development to cater for AI is being supervised by
the Ministry of ICT and Innovation, in partnership with Rwanda Utilities
and Regulatory Authority, the Rwanda Information Society Authority, and
all relevant stakeholders from the public and private sectors and civil
society.1381 The partnership strategy with The Future Society for
development of national AI strategy will shape how much public and civic
participation will be incorporated to the process.
OECD AI Principles
Rwanda is not a signatory to OECD AI Principles. While the nation
does not have an established AI strategy yet, the engagement of The Future
Society and GIZ FAIR Forward to help develop it, as well as its close
alignment with GDPR is a positive sign for future direction.
Fundamental Rights
After the genocide of 1994, Rwanda had to rebuild its infrastructure
and relations from ground up. A unity and reconciliation process was
followed by a combination of traditional systems of justice and international
tribunals.1382 The National Commission for Human Rights was created in
1999 as an independent institution responsible for the promotion and
protection of human rights in Rwanda.1383 In 2017 Rwanda withdrew from
Protocol to the African Charter on Human and Peoples’ Rights which
allows individuals and NGOs to bring cases directly to the African Court on
Human and Peoples’ Rights came into effect.1384 Freedom House scores
Rwanda as “Not Free.”1385 The Ibrahim Index of African Governance scores
1381 Habumuremyi, E. AI eyed to transform health care in Rwanda, Global Information
Society Watch, https://www.giswatch.org/node/6186
1382 Nkusi, A. The Rwandan Miracle, UNICEF, https://en.unesco.org/courier/20192/rwandan-miracle
1383 Republic of Rwanda Ministry of Justice, The National Human Rights Action Plan of
Rwanda 2017-2020
https://minijust.gov.rw/fileadmin/Documents/MoJ_Document/NHRAP_FINAL__version
_for_cabinet-1.pdf
1384 Amnesty International, Rwanda: More progress needed on human rights
commitments. Amnesty International submission for the UN Universal Periodic Review –
37th Session of the UPR Working Group, January-February 2021 (Aug. 2020),
https://www.amnesty.org/download/Documents/AFR4728582020ENGLISH.PDF
1385
Freedom House, Global Freedom Scores: Rwanda.
https://freedomhouse.org/countries/freedom-world/scores
352
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Rwanda’s 2019 record on Rights at 29.1 out of 100, and ranks Rwanda at
60.5 for overall governance.1386
The European Union concluded in its 2018 human rights report that
“area with the most significant restrictions of human rights were the
politically related rights and freedoms such as the freedom of
expression/freedom of media, freedom of association and freedom of
assembly.”1387 Ahead of the 2021 UN Universal Periodic Review of
Rwanda, the government announced National Action Plan for Human
Rights (NHRAP) 2017-2020, formulated through a participatory process.
This the first of its kind in Rwanda and builds on extensive work by
Rwandan government to create an inclusive society where all are valued
and have equal opportunity. The government commits that Universal
Declaration of Human Rights should guide all future programs and policies
in all sectors and in all phases of the programming process including
monitoring and evaluation.
The Rwandan judiciary lacks independence from the executive.
Top judicial officials are appointed by the president and confirmed by the
Senate1388 dominated by governing party. In addition, the 2013 law allows
for security organs to record or listen to communications both offline and
online, and without necessarily facilitating through a service provider, if it
is done in the interest of national security.1389 The regulations require
mandatory SIM card registration and a limit of three cards per national ID
per operators. Service providers are required to maintain databases and
share information with law enforcement if necessary.
Data Protection
In 2019 Rwanda adopted Child Online Protection Policy (“the COP
Policy”) designed to mitigate against those risks and harms, and to deliver
a framework that meets children’s needs and fulfils their rights. It makes it
Ibrahim Index of African Governance, Rwanda, 2019, Comparison of 54 locations by
measure “Participation, Rights, & Inclusion”. https://iiag.online
1386
1387 European Union, Annual Report on Human Rights and Democracy in the World
2018 – Rwanda (May 21, 2019) https://eeas.europa.eu/delegations/rwanda/62839/euannual-report-human-rights-and-democracy-world-2018-rwanda_en
1388 Freedom House, Freedom in the World 2020,
https://freedomhouse.org/country/rwanda/freedom-world/2020
1389 Republic of Rwanda, N° 60/2013 of 22/08/2013 Law regulating the interception of
communications. Official Gazette nº 41 of 14/10/2013
https://rema.gov.rw/rema_doc/Laws/Itegeko%20rishya%20rya%20REMA.pdf
353
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
one of the first countries to adopt such protections.1390 In 2019 Rwanda
ratified the African Union Convention on Cyber Security and Personal Data
Protection.1391 Law No. 058/2021 Relating to the Protection of Personal
Data and Privacy was published on 15 October 2021. It establishes
principles of lawfulness, fairness and transparency, purpose limitation and
accuracy. It also requires personal data protection impact assessments. The
legislation does not establish an independent data protection agency but
states ‘supervisory authority is 'a public authority in charge of cyber
security.”1392 In introduction of the new legislation, National Cyber Security
Authority defined personal data as a “fundamental right”.1393 Following the
GDPR, the Privacy Law seeks to safeguard fundamental rights to privacy
by regulating the processing of data and providing the individual with rights
over their data.1394 The law establishes systems of accountability and clear
obligations for those who control the processing of the personal data.
According to One Trust, “The bill is relatively comprehensive and would
introduce obligations related to data subject rights, data processing
notifications, pseudonymisation, sensitive data, data transfers, and data
breach notifications.”1395
Algorithmic Transparency
Government provides open datasets1396 and government services.1397
Under NST1, government targets to ensure 100% Government services are
1390
https://rura.rw/fileadmin/Documents/ICT/Laws/Rwanda_Child_Online_Protection_Polic
y.pdf
1391 Daniel Sabiiti, Rwanda Ratifies Malabo Convention On Personal Data Protection,
KT Press (July 25, 2019), https://www.ktpress.rw/2019/07/rwanda-ratifies-malaboconvention-on-personal-data-protection/
1392
Law No. 058/2021 of 13 October 2021 Relating to the Protection of Personal Data
and Privacy
https://www.minijust.gov.rw/fileadmin/user_upload/Minijust/Publications/Official_Gazet
te/2021_Official_Gazettes/October/OG_Special_of_15.10.2021_Amakuru_bwite.pdf
1393
https://cyber.gov.rw/updates/article/rwanda-passes-new-law-protecting-personaldata-1/
1394
Julius Bizimungu, Rwanda moves to tighten data protection, privacy, the New Times
(Nov. 6, 2020), https://www.newtimes.co.rw/news/rwanda-moves-tighten-dataprotection-privacy
1395
OneTrust DataGuidance, Rwanda (Oct. 29, 2020),
https://www.dataguidance.com/jurisdiction/rwanda
1396 Rwanda Data Portal dhttps://rwanda.opendataforafrica.org/
1397 Irembo.gov https://irembo.gov.rw/home/citizen/all_services
354
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
delivered online by 2024.1398 However the ability in practice to obtain
information about state operations remains questionable.
Evaluation
Rwanda's national strategy, including Vision 2050 and FAIR
Forward, aligns with the OECD/G20 AI Principles and encourages public
participation in future AI decisions. Rwanda has moved toward stronger
standards for data protection but has not yet established an independent
agency to ensure data protection or to oversee AI deployment. The Smart
City initiative in Kigali offers enormous promise, but also must be carefully
monitored to ensure that a system of mass surveillance does not take place.
The pioneering work on drone regulation needs to be coupled with a stand
against lethal autonomous weapons.
1398 Republic of Rwanda, 7 Years Government Programme: National Strategy for
Transformation (NST1)
355
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Saudi Arabia
National AI Strategy
The Saudi AI initiatives are led by the Saudi Data and AI Authority
(SDAIA), which reports directly to the Prime Minister and consists of
members chosen by the Prime Minister.1399 The Saudi Data and AI
Authority’s website provides basic information about the Kingdoms goals
for AI.1400 In the September 2020 SDAIA and Riyadh signed a cooperative
agreement for an AI Oasis.
In August 2020, SDAIA published a National Strategy for Data and
1401
AI. The AI Strategy is to advance the KSA Vision 2030.14021403 The AI
Strategy states that this Vision will be achieved “through a multi-phased
approach focused on addressing the national priorities by 2025, building
foundations for competitive advantage in key niche areas by 2030, and
becoming one of the leading economies utilizing and exporting Data & AI
after 2030.”1404 The National Strategy states 6 objectives:
1. “Ambition: Position KSA as the global hub where the best of
Data & AI is made reality
2. Skills: Transform KSA’s workforce with a steady local supply
of Data & AI-empowered talents
3. Policies & Regulations: Enact the most welcoming legislation
for Data & AI businesses and talents
4. Investment: Attract efficient, stable funding for qualified Data
& AI investment opportunities
5. Research & Innovation: Empower top Data & AI institutions to
spearhead innovation and impact creation
6. Ecosystem: Stimulate Data & AI adoption with the most
collaborative, and forward-thinking ecosystem”
Most of the goals focus on fostering an enabling business and regulatory
environment. This includes education schemes that promote the
development of a workforce that fits the industry’s new needs. The
government is undertaking major educational reform to foster the
Saudi Gazette, King Salman issues royal decrees, including setting up of industry and
resources ministry (Aug. 2019), https://saudigazette.com.sa/article/575953
1400
Saudi Data and AI Authority, Home, https://sdaia.gov.sa/
1401
Carrington Malin, Saudi National Strategy for Data and AI (Aug. 2020),
https://www.sme10x.com/technology/saudi-national-strategy-for-data-and-ai-approved
1402
Government of Saudi Arabia, Vision 2030 (2020), https://vision2030.gov.sa/en
1403
Catherine Jewell, Saudi Arabia embraces AI-driven innovation (Sept. 2018),
https://www.wipo.int/wipo_magazine/en/2018/05/article_0002.html
1404
SDAIA, National Strategy for Data & AI: Realizing our best tomorrow (Oct. 2020),
https://ai.sa/Brochure_NSDAI_Summit%20version_EN.pdf
1399
356
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
development of digital skills for jobs in emerging technology including
AI.1405
The National Strategy states, “As part of providing an attractive
regulatory framework for Data & AI investments and businesses, we aim at
having a strong framework to promote and support ethical development of
Data & AI research and solutions. This framework will provide guidelines
for the development of our data protection and privacy standards.”1406 On
policies and regulations, “NDMO is developing a number of regulatory
frameworks including topics such as data privacy and freedom of
information. In particular, Open Data has been identified as a priority topic,
and the government has already made investments in open data platforms.”
Late in October 2020, the SDAIA published National Data
Governance Interim Regulations to govern the collection and use of
personal data and the management of data by government entities.1407 These
regulations appear to be broadly influenced by the GDPR and include
extensive rights for data subjects and obligation for data controllers. For
example, individuals will have the “right to be informed of the legal basis
and purpose for the collection and processing of their personal data.
Personal data cannot be collected or processed without the Data Subject’s
express consent.” Data subjects will also have the “right to access personal
data in possession of the Data Controller, including the right to correct,
delete, or update personal data, destroy unnecessary data, and obtain a copy
of the data in a clear format.”
The SDAIA was established in 2019 with the aim of fostering “the
digital ecosystem while also supporting the suite of values-based G20 AI
Principles.”1408 “Data is the single most important driver of our growth and
reform and we have a clear roadmap for transforming Saudi Arabia into a
leading AI and data-driven economy,” said Dr Abdullah bin Sharaf Al
Catherine Early, Saudi Arabia signs off on Artificial Intelligence policy (Aug. 2020),
https://www.globalgovernmentforum.com/saudi-arabia-signs-off-on-artificialintelligence-policy/
1406
SDAIA, National Strategy for Data & AI: Realizing our best tomorrow (Oct. 2020),
https://ai.sa/Brochure_NSDAI_Summit%20version_EN.pdf
1407
Albright Stonebridge Group, ASG Analysis: Saudi Arabia Publishes National Data
Governance Interim Regulations (Oct. 21, 2020),
https://www.albrightstonebridge.com/news/asg-analysis-saudi-arabia-publishes-nationaldata-governance-interim-regulations
1408
OECD G20 Digital Economy Task Force, Examples of AI National Policies 35, 41,
42 (2020), https://www.mcit.gov.sa/sites/default/files/examples-of-ai-nationalpolicies.pdf
1405
357
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Ghamdi, president of SDAIA.1409 SDAIA oversees three organizations: the
National Data Management Office (NDMO), the National Information
Center (NIC) and the National Center for AI (NCAI). The NDMO is
responsible for the regulation of data which includes standardization and
regulation of artificial intelligence as well as ensuring compliance. The NIC
oversees the operation of government data infrastructure and government
analytics. The implementation of the national AI strategy is the main
responsibility of the NCAI. This includes facilitating capacity-building, AI
innovation and raising awareness of AI as well as expanding education on
AI.14101411
Saudi Arabia’s Deputy Minister Dr. Ahmed AL Theneyan
emphasized the importance of regulation in interviews about the Kingdom’s
AI Strategy. This includes education plans that promote the development of
a workforce that fits the industry’s new needs. The government is
undertaking major educational reform to foster the development of digital
skills for jobs in emerging technology including AI.1412 ”The country is
establishing a national data bank to consolidate more than 80 government
datasets, the equivalent to 30 per cent of the government’s digital assets. It
is also planning to build one of the largest clouds in the region by merging
83 data centres owned by more than 40 government bodies.”1413
During the fight against COVID-19 the SDAIA launched two apps:
the Tawakkalna app to manage movement permits for government and
private sector employees and Tabaud to notify citizens when they have
come in contact with someone who was infected with the virus.14141415 MIT
Gulf News, Saudi Arabia approves policy on Artificial Intelligence, expects SR500b
windfall by 2030, (Aug. 10, 2020), https://gulfnews.com/business/saudi-arabia-approvespolicy-on-artificial-intelligence-expects-sr500b-windfall-by-2030-1.1597032000775
1410
Future of Life, AI-Policy Saudi Arabia, https://futureoflife.org/ai-policy-saudi-arabia/
1411
OECD G20 Digital Economy Task Force, Examples of AI National Policies 35, 41,
42 (2020), https://www.mcit.gov.sa/sites/default/files/examples-of-ai-nationalpolicies.pdf
1412
Early, Catherine, Saudi Arabia signs off on Artificial Intelligence policy (Aug. 2020),
https://www.globalgovernmentforum.com/saudi-arabia-signs-off-on-artificialintelligence-policy/
1413
Vishal Chawla, How Saudi Arabia Is Looking To Develop & Integrate Artificial
Intelligence In Its Economy, Analytics India Magazine (Aug. 23, 2020),
https://analyticsindiamag.com/how-saudi-arabia-is-looking-to-develop-integrateartificial-intelligence-in-its-economy/
1414
SDAIA, Tawakkalna, https://ta.sdaia.gov.sa/En/
1415
SDAIA, Tabaud, https://tabaud.sdaia.gov.sa/indexEn
1409
358
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Technology Review reported that Tabaud is transparent, voluntary, and
minimizes data collection.1416
In 2017 Saudi Arabia granted the robot, Sophia, citizenship. This is
a first worldwide and was met with mixed reactions.1417 CNBC said,
“Sophia been touted as the future of AI, but it may be more of a social
experiment masquerading as a PR stunt.”1418 Bloomberg noted that
“Migrant laborers can’t become citizens; android Sophia can.” 1419
Global AI Summit
The Global AI Summit, held in October 2020, is described as the
“world's premier platform for dialogue that brings together stakeholders
from public sector, academia and private sector, including technology
companies, investors, entrepreneurs and startups to shape the future of
Artificial Intelligence (AI).”1420 Speakers from across sectors explored the
theme “AI for the Good of Humanity.” Notably, several of the sessions
focused on ethics and making use of AI for social causes with titles like:
“AI for the good of humanity”, “AI for good, AI for all: Collective thinking
on how access to AI can be democratized to shape and deliver positive
societal impact” and “Misuse vs. Missed Use: The Ethics Question: Ethics
and ethical frameworks as a tool in unleashing AI innovation”.1421 In
opening remarks, the President of the SDAIA stated “during the two days
we will also be announcing several major partnerships and initiatives with
our international partners to accelerate AI for sustainable development in
low and middle income countries and to enable the sharing of AI best
practices globally to ensure a more inclusive future powered by AI where
no one is left behind.” He emphasized the importance of working together
internationally to ensure the sustainable development of AI.1422
MIT Technology Review, COVID Tracing Tracker (May 7, 2020),
https://www.technologyreview.com/2020/05/07/1000961/launching-mittr-covid-tracingtracker/,
1417
Future of Life, AI Policy-Saudi Arabia, https://futureoflife.org/ai-policy-saudi-arabia/
1418
Jaden Urbi and Sigalos MacKenzie, The Complicated Truth about Sophia the Robotan almost human robot or a PR stunt, CNBC (June 2018),
https://www.cnbc.com/2018/06/05/hanson-robotics-sophia-the-robot-pr-stunt-artificialintelligence.html
1419
Tracy Alloway, Saudi Arabia Gives Citizenship to a Robot, Bloomberg (Oct. 2017),
https://www.bloomberg.com/news/articles/2017-10-26/saudi-arabia-gives-citizenship-toa-robot-claims-global-first
1420
Global AI Summit, About Us, https://www.theglobalaisummit.com/#about-us
1421
Global AI Summit, Program, https://www.theglobalaisummit.com/#program
1422
Global AI Summit, AI for the Good of Humanity (Oct. 21, 2020) (livestream),
https://www.youtube.com/watch?v=uOGYQlhmb_8
1416
359
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
During the Global AI Summit, several international organizations
announced new initiatives. The World Bank Group and SDAIA set out a
new partnership to “help finance, stimulate, and accelerate the development
and adoption of artificial intelligence technologies to serve people and
development initially in Africa and globally at a later stage” and “to
strengthen Saudi Arabia's role as a key contributor in supporting developing
countries.”1423 The International Telecommunications Union signed a
Memorandum of Understanding with the Kingdom to “support global
cooperation in the field of artificial intelligence.” The ITU will also “see the
development of an internationally-recognized system for countries to
mobilize resources, providing assistance for official agencies that want to
adopt AI technologies, and accreditation to meet economic requirements.”
Dr. Abdullah bin Sharaf Alghamdi, President of the SDAIA, stated: “The
International Telecommunication Union will share the best practices in the
field of artificial intelligence with the Kingdom. This will help in shedding
light on how to sponsor and support emerging companies and new
incubators in the national space, especially as there is no official framework
that currently exists to support the AI readiness of countries and
international cooperation."1424
Neom
The Kingdom has also initiated a smart city project called Neom.
Neom “is an international project that will be led, populated and funded by
people from all over the world.” According to the FAQ, Neom will be a
“semi-autonomous region with its own government and laws” in northwest
Saudi Arabia on the Red Sea and home to one Million people by 2030.1425
Neom is envisioned to become a city that “will introduce a new model for
urbanization and sustainability,” built on five principles: sustainability,
community, technology, nature, livability.1426
Public Participation
The Saudi AI initiatives are led by the Saudi Data and AI Authority,
which reports directly to the Prime Minister and consists of members
Global AI Summit, Press Releases, Keen to harness benefits of artificial intelligence
for all Saudi Arabia establishes new partnership with World Bank Group (Oct. 2020),
https://theglobalaisummit.com/news4.html
1424
Global AI Summit, Press Releases, SDAIA and International Telecommunication
Union sign MoU to Develop International Artificial Intelligence Framework (Oct. 2020),
https://theglobalaisummit.com/news10.html
1425
NEOM, FAQ, https://www.neom.com/en-us/static/pdf/en/NEOM_FAQ_EN.pdf
1426
NEOM, Brochure, https://www.neom.com/enus/static/pdf/en/NEOM_BROCHURE_EN.pdf
1423
360
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
chosen by the Prime Minister.1427 As far as can be gathered from the website,
the Vision 2030 Strategy was developed by the Council of Ministers and
the Council of Economic Affairs.1428 The SDAIA Website lists its current
activities, achievements and general information about the authority and its
work.1429 Further, the SDAIA provides a digitized version of the strategy
report.1430 The Vision 2030 website sets out broad policy objectives.1431
There are indicators and targets for every Theme as well as information on
the respective initiatives. The Vision 2030, however, encompasses many
different objectives, AI being only one of many.
The Global AI Summit 2019 provided important opportunities for
individuals and organizations to express their views on AI policy. However,
the conference was not held in 2021. It is anticipated that the conference
will resume in 2022.
In 2021, the KSA Communications and Information Technology
Commission (CITC) sought public comment on the Digital Content
Platform Regulations Document.” According to the Commission, "The
initiative is aimed to regulate, govern, activate, and motivate digital content
platforms to expand and grow. In addition to engage the private sector,
empower entrepreneurs as well as attract investments and protect users of
digital content platforms.” They stated: "The commission calls on interested
parties from the Kingdom and abroad as well as the public to submit their
views on the consultations document before November 30, 2021. The
Digital Content Council and CITC value the importance of engaging
interested parties, investors and entrepreneurs in regulations drafting
process."1432
G-20 Meetings
Saudi Arabia hosted the G-20 Digital Economy Ministers Meeting
in June 2020. AI policy was a focal point of the discussions. The Digital
Economy Task Force released a new report on the implementation of the
Saudi Gazette, King Salman issues royal decrees, including setting up of industry and
resources ministry (Aug. 2019), https://saudigazette.com.sa/article/575953
1428
Vision 2030, Governance, https://vision2030.gov.sa/en/governance
1429
SDAIA (SDAIA), https://sdaia.gov.sa/?Lang=en
1430
Kingdom of Saudi Arabia, National Strategy for Data and AI, https://ai.sa/indexen.html
1431
Kingdom of Saudi Arabia, Vision 2030, Programs,
https://vision2030.gov.sa/en/programs#
1432
KSA Communications and Information Technology Commission, CITC Publishes a
Public Consultation on Digital Content Platform Regulations Document (2021)
https://www.citc.gov.sa/en/new/publicConsultation/Pages/144304.aspx).
1427
361
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD AI Principles.1433 The key agenda items selected by the Saudi
government were: “Empowering People, by creating the conditions in
which all people – especially women and youth – can live, work and thrive”;
“Safeguarding the Planet, by fostering collective efforts to protect our
global commons”; and “Shaping New Frontiers, by adopting long-term and
bold strategies to share benefits of innovation and technological
advancement.”14341435
In November 2020, Saudi Arabia hosted the G20 Ministerial
meeting in Riyadh. There was controversy surrounding the event as many
human rights organizations protested the decision to allow the Kingdom to
host the G20 meeting. Still, there was progress on fundamental rights
associated with AI and digital technologies. The G20 Leaders in Riyadh
stated, “We will continue to promote multi-stakeholder discussions to
advance innovation and a human-centered approach to Artificial
Intelligence (AI), taking note of the Examples of National Policies to
Advance the G20 AI Principles. We welcome both the G20 Smart Mobility
Practices, as a contribution to the well-being and resilience of smart cities
and communities, and the G20 Roadmap toward a Common Framework for
Measuring the Digital Economy.”1436
On Digital Economy, the G20 said “We acknowledge that universal,
secure, and affordable connectivity, is a fundamental enabler for the digital
economy as well as a catalyst for inclusive growth, innovation and
sustainable development. We acknowledge the importance of data free flow
with trust and cross-border data flows.” The G20 Declaration further said,
“We support fostering an open, fair, and non-discriminatory environment,
and protecting and empowering consumers, while addressing the challenges
related to privacy, data protection, intellectual property rights, and
security.”
1433
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1434
SDG Knowledge Hub, G20 Digital Economy Ministers Meeting, July 2020,
https://sdg.iisd.org/events/g20-digital-economy-ministers-meeting/
1435
G20 Saudi Arabia, https://g20.org/en/Pages/home.aspx. [Editorial note: At the time of
publication we found that the materials from the G20 summit that were available shortly
after the Summit concluded, including the Leaders Declaration, were no longer available
at the G20 website. Fortunately, copies of these documents are archived and available at
the Internet Archive, https://web.archive.org]
1436
G20 Riyadh Summit, Leaders Declaration (Nov. 21-22, 2020),
https://g20.org/en/media/Documents/G20%20Riyadh%20Summit%20Leaders%20Declar
ation_EN.pdf
362
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI Oversight
The Saudi Data and Artificial Intelligence Authority (SDAIA) was
established by a royal decree in 2019.1437 The SDAIA is directly linked to
the Prime Minister and will be governed by a board of directors chaired by
the Deputy Prime Minister.
The KSA Human Rights Commission was founded in 2005 and "has
full independence in the exercise of its tasks for which it was established
and stipulated in its organization." The Commissions states that it “aims to
protect and promote human rights in accordance with standards
International human rights in all fields, raising awareness of them and
contributing to ensuring that this is implemented in light of the provisions
of Islamic Sharia."1438
In 2021, the SDAIA initiated the Open Data Strategy 2022-2024
with the aim to "Provide high-value and re-usable Open Data for the nation
to increase transparency and foster innovation through collaboration,
enabling a data driven economy.” With this initiative the SDAIA also hopes
to empower governance and regulatory clarity, enable economic growth,
prioritise and publish accessible, quality and demanded data sets and create
impact through awareness, innovation and international and local
partnerships.1439
Data Protection
In 2021 the first data protection law in Saudi Arabia was introduced.
It is expected to take effect in March 2022 and there will be a transition
period of 18 months. The data protection law, modelled after the GDPR,
aims to prevent the misuse of personal data and specifically implements
principles such as “purpose limitation and data minimization, controller
obligations, including registration and maintenance of data processing
records, data subject rights, and penalties for breach of provisions.”1440 The
law also aims to align the Kingdom with other countries in the region and
Arab News, King Salman issues royal decrees, including creation of industry and
resources ministry (Aug. 30, 2019), https://www.arabnews.com/node/1547546/saudiarabia
1438
Kingdom of Saudi Arabia, Human Rights Commission, https://hrc.gov.sa/enus/aboutHRC/AboutHRC/Pages/HRCvision.aspx.
1439
SDAIA, Open Data Strategy 2022-2024 (2021),
https://sdaia.gov.sa/files/KSAOpenData%20StrategyExecutiveSummary.pdf ).
1440
Data Guidance, Saudi Arabia (2021),
https://www.dataguidance.com/jurisdiction/saudi-arabia
1437
363
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
with international standards.1441 In the meantime, the National Data
Governance Interim Regulations will remain in place.
As noted above, the National AI Strategy proposed strong rules for
data protection. “This initiative requires a strong regulatory framework to
provide high standards in terms of data protection and privacy, in line with
our ethical approach to developing our data sector. Doing this would define
the framework through which government and private organizations will be
able to benefit from the opportunities provided by data. The regulatory
framework will include specifications on data collection, classification,
sharing, open data policy and freedom of information.”1442 The Strategy also
notes that “NDMO is developing a number of regulatory frameworks
including topics such as data privacy and freedom of information.”
According to PWC, the E-Commerce Law of 2019 “focuses on
regulating e-commerce business practices requiring increased transparency
and consumer protection, with the goal of enhancing trust in online
transactions. The law also contains provisions aimed at protecting the
personal data of e-commerce customers. Specifically, the law specifies that
service providers will be responsible for protecting the personal data of
customers in their possession or ‘under their control.’ ‘Control’ in a data
protection context exists where an organisation can make decisions
concerning that personal data, such as why to collect it in the first place,
what to do with it, how long to keep it, and who to share it with. A service
provider may still have ‘control’ of personal data where it passes the data
on to a third party as part of an outsourcing or other arrangement.”1443 The
KSA E-Commerce Law also prohibits service providers from using
customers’ personal data for ‘unlicensed or unauthorised’ purposes, and
from disclosing personal data to third parties without the customer’s
consent.
Data Governance
Further, in October 2020, the SDAIA, published National Data
Governance Interim Regulations. The regulations cover five topics: “data
classification by public entities, protection of personal data, data sharing
Bureau of Experts at the Council
Ministers, https://laws.boe.gov.sa/BoeLaws/Laws/LawDetails/b7cfae89-828e-4994b167-adaa00e37188/1).
1442
SDAIA, National Strategy for Data & AI: Realizing our best tomorrow (Oct. 2020),
https://ai.sa/Brochure_NSDAI_Summit%20version_EN.pdf
1443
PWC, Saudi Arabia Data Privacy Landscape (Nov 2019),
https://www.pwc.com/m1/en/services/tax/me-tax-legal-news/2019/saudi-arabia-dataprivacy-landscape-ksa.html
1441
364
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
between public entities, freedom of information requests, and open data.
Much of the document, including the regulation on the protection of
personal data, draws significantly from international regulations such as the
EU’s General Data Protection Regulation (GDPR).”1444
OECD/G20 AI Principles
The Kingdom has endorsed the G20 AI Principles. Regarding
implementation of the AI Principles, the OECD notes their National Center
for AI (NCAI), the Saudi Data and AI Authority and highlights their work
towards trustworthy AI in health.1445
Human Rights
Freedom House gives Saudi Arabia very low marks for to civil
liberties and political rights (7/100).1446 Freedom Hosue reports, “Saudi
Arabia’s absolute monarchy restricts almost all political rights and civil
liberties. No officials at the national level are elected. The regime relies on
pervasive surveillance, the criminalization of dissent, appeals to
sectarianism and ethnicity, and public spending supported by oil revenues
to maintain power.” However, in 2021 Saudi Arabia was one of 193
countries that endorsed the UNESCO Recommendation on AI Ethics.
Due to its membership in the United Nations the Saudi Arabian
Government (UDHR) has inherently committed to upholding human rights
standards which include those laid out in the (UDHR).1447 However, Saudi
Arabia was the sole abstainer on the Declaration among Muslim nations,
stating that it violated Sharia law.1448
Upon pressure from the American foreign policy advocacy group,
Freedom Forward, the mayors of New York, London, Paris and Los
Angeles chose to boycott the 2019 G20 meeting in Riyadh due to the human
rights violations committed by the Saudi government.1449
Albright Stonebridge Group, ASG Analysis: Saudi Arabia Publishes National Data
Governance Interim Regulations (Oct. 21, 2020),
https://www.albrightstonebridge.com/news/asg-analysis-saudi-arabia-publishes-nationaldata-governance-interim-regulations
1444
G20 Digital Economy Task Force, Examples of National AI Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1446
Freedom House, Freedom in the World 2021: Saudi Arabia (2021),
https://freedomhouse.org/country/saudi-arabia/freedom-world/2021
1447
Human Right Watch, International Human Rights Standards,
https://www.hrw.org/reports/1997/saudi/Saudi-07.htm
1448
Human Rights Watch, Saudi Arabia: Human Rights Developments,
https://www.hrw.org/reports/1992/WR92/MEW2-02.htm
1449
Natasha Turak, Saudi Arabia loses vote to stay on UN Human Rights Council; China,
Russia and Cuba win seats, CNBC (Oct. 14, 2020),
1445
365
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Evaluation
Saudi Arabia has emerged as an influential leader among the G20
countries and a powerful AI economic force globally. Although the country
has a poor track record on human rights, Saudi Arabia’s engagement with
global AI policy is having a liberalizing influence. Not only has Saudi
Arabia hosted important meetings of the G-20, the Kingdom also organized
a successful global summit on AI that brought together government
representatives, industry leaders, and academics and civil society. Saudi
Arabia has endorsed the G20 AI Principles, but steps still should be taken
to strengthen human rights, to promote public participation in AI
policymaking within country, and to create mechanisms, including a data
protection authority, to provide independent oversight of AI deployment.
https://www.cnbc.com/2020/10/14/saudi-arabia-loses-vote-for-un-human-rights-councilseat-china-russia-win.html; Freedom Forward, Boycotting the Saudi G20: Our Successes
(Nov. 16, 2020), https://freedomforward.org/2020/11/16/boycotting-the-saudi-g20-oursuccesses/
366
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Singapore
National AI Strategy
Singapore’s national AI strategy1450 is part of its Smart Nation1451
policy and nation-building exercise. Smart Nation agenda aims to digitalize
health, transport, urban solutions, finance, and education domains and make
use of AI to the maximum. The vision is to make Singapore “global hub for
developing, test-bedding, deploying, and scaling AI solutions.”1452
Government supports a “dare to try” mind-set, experimentation with new
ideas and manageable risks.1453 The strategy also envisions “human-centric
approach towards AI governance that builds and sustains public trust.”1454
An industry-led initiative, Advisory Council on the Ethical Use of
AI and Data, has been established to assess the ethical and legal use of AI
and data as well as recommend policies and governance to encourage
industry to develop and adopt AI technologies in an accountable and
responsible manner. Advisory Council is also tasked to assist the
Government develop voluntary codes of practice to guide corporate
decision makers, monitor consumers’ acceptance of such data use, and
make recommendations on ethical and legal issues that may require policy
or regulatory changes.1455 The scope of work addresses all five principles of
the G20 AI Principles.
In 2019, Singapore published Asia’s first Model AI Governance
Framework1456 (updated in 2020) that provides implementable guidance to
private sector to address key ethical and governance issues when deploying
AI solutions. Feedback from participation in European Commission’s HighSingapore, National Artificial Intelligence Strategy (2019):
https://www.smartnation.gov.sg/why-Smart-Nation/NationalAIStrategy
1451
Singapore, Smart Nation: The Way Forward Executive Summary (2018):
https://www.smartnation.gov.sg/docs/default-source/default-document-library/smartnation-strategy_nov2018.pdf?sfvrsn=3f5c2af8_2
1452
Singapore, National Artificial Intelligence Strategy (2019):
https://www.smartnation.gov.sg/why-Smart-Nation/NationalAIStrategy
1453
Smart Nation: The Way Forward Executive Summary (2018)
1454
Infocomm Media Development Authority (IMDA) and Personal Data Protection
Commission (PDPC), Model AI Governance Framework, 2nd Edition (2020)
https://www.pdpc.gov.sg/-/media/files/pdpc/pdf-files/resource-fororganisation/ai/sgmodelaigovframework2.pdf
1455
IMDA, The full composition of Singapore’s Advisory Council on the Ethical Use of
AI and Data (Advisory Council) was announced by Minister for Communications and
Information Mr S Iswaran at AI Singapore’s first year anniversary (Aug. 30, 2018)
https://www.imda.gov.sg/news-and-events/Media-Room/MediaReleases/2018/composition-of-the-advisory-council-on-the-ethical-use-of-ai-and-data
1450
1456
IMDA and PDPC, Model AI Governance Framework, 2nd Edition (2020)
367
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Level Expert Group and the OECD Expert Group on AI is reflected in the
framework. The framework is accompanied by an Implementation and SelfAssessment Guide for Organizations. The AI Governance Framework is
intended to help organizations “demonstrate reasonable efforts to align
internal policies, structures and processes with relevant accountabilitybased practices (e.g., the Personal Data Protection Act 2012 (PDPA) and
the OECD Privacy Principles)” and hence build stakeholder confidence in
AI. These documents focus on implementation instead of high-level
discussions which is reflective of Singapore’s mindset of test and deploy.
All of Singapore’s strategy and action towards AI is based on
voluntary governance, requiring organizations using AI in decision-making
ensure that process is explainable, transparent, fair with clear roles and
responsibilities. Solutions are also expected to have protection of interest of
human beings as primary consideration, including their well-being and
safety.
The Centre for AI & Data Governance (CAIDG), funded by
government, is established to develop international thought leadership and
advance scholarship and discourse in legal, ethical, regulatory and policy
issues arising from the use of AI and data and inform implementation of
G20 AI Principles.
In October 2020, Singapore Computer Society (SCS), supported by
the regulator Infocomm Media Development Authority (IMDA), launched
the AI Ethics and Governance Body of Knowledge (BoK). BoK is expected
to “guide the development of curricula on AI ethics and governance and
form the basis of future training and certification for professionals.” The
document underlines that “accountability, transparency, explainability, and
auditability must become the hallmark of all AI solutions” and that “ethical
guidelines should not be an afterthought but integrated as part of standards
and expectations from the onset of any AI-related effort.”1457
AI System for Online Surveillance
Maintaining racial and religious harmony has been the
Government’s stated top priority.1458 Right to privacy is not a right protected
by the Singapore constitution.1459 Protection from Online Falsehoods and
The Singapore Computer Society, Artificial Intelligence Ethics & Governance Body
of Knowledge (2020) https://ai-ethics-bok.scs.org.sg/document/15
1458
UN Human Rights Council, Universal Periodic Review – Singapore, National
Report, Second Cycle (2015)
https://www.ohchr.org/EN/HRBodies/UPR/Pages/SGIndex.aspx
1459
Privacy International, Universal Periodic Review, Stakeholder Report: 24th Session,
Singapore, The Right to Privacy in Singapore (2015)
1457
368
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Manipulation Act 20191460 was introduced to regulate “fake news” by
malicious actors. However, the law does not clearly define what is meant
by falsehood and gives power to any government minister to declare that
information posted online is “false” and instruct the correction or removal
of such content if he/she thinks it is in the public interest to remove.11 The
law applies to digital content that is accessible in Singapore, whether it is
an online post, text or chat message by a person or a bot. A person found
guilty of the offense can be fined monetarily or be imprisoned.
Public Order Act’s definition of assembly and its requirements for
permit for such assembly has recently extended to online conferences. This
is compounded by the fact that Singapore has not ratified the International
Covenant on Civil and Political Rights1461 which protects against arbitrary
or unlawful interference with privacy, family, home or correspondence. No
court warrant is required to monitor personal phone, messaging or other
electronic communication.1462 Government’s use of online surveillance tools
and power to act without need for legal authorization is concerning on the
citizen’s ability to exercise their rights of freedom of speech, expression and
assembly.
Singapore also utilizes ABBSS (Automated Biometrics &
Behavioral Screening Suite) at immigration and border checkpoints. The
system is a network of cameras with facial recognition capabilities that can
also be deployed as a body-worn-camera for officers. It is used both to build
a biometrics database of travelers and detect travelers wanted for various
offenses.1463
In October 2021, Singapore Parliament passed The Foreign
Interference (Countermeasures) Bill (FICA) to deal with foreign
interference after 10-hour debate, despite protests over lack of public
consultation. Under the new law, the Minister for Home Affairs is granted
powers to issue directions to Internet, social media service providers and
https://privacyinternational.org/sites/default/files/201712/Singapore_UPR_PI_submission_FINAL.pdf
1460
Singapore Statutes Online, Protection from Online Falsehoods and Manipulation Act
2019 https://sso.agc.gov.sg/Acts-Supp/182019/Published/20190625?DocDate=20190625
1461
UN Human Rights Council, Universal Periodic Review – Singapore, Outcome of the
Review, Second Cycle (2015)
https://www.ohchr.org/EN/HRBodies/UPR/Pages/SGIndex.aspx
1462
Privacy International, Universal Periodic Review, Stakeholder Report: 24th Session,
Singapore
1463
Wong, K. Facial recognition, biometrics tech at more checkpoints: ICA. The Strait
Times (Nov. 13, 2018) https://www.straitstimes.com/singapore/facial-recognitionbiometrics-tech-at-more-checkpoints-ica
369
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
website operators to provide user information, block content and remove
applications. Authorities can also require politically significant persons to
declare foreign affiliations.1464 The bill provides the Ministry overreach of
power with little oversight, opening concerns on impact on freedom of
speech and assembly.
Smart Cities
Singapore is leading the efforts to develop an ASEAN Framework
on Digital Data Governance to facilitate harmonization of data
regulations.1465 It is also one of the 26 within the ASEAN Smart Cities
Network (ASCN) named by member states to pilot smart city project.
Digital Identification and Surveillance
SingPass, the National Digital Identity (NDI) initiative, 1466 is a
personal authentication system that allows users to access various
Government services. It is a move to digitalize all transactions on public
and private space and share data. The app does provide the users with option
to use a 6-digit passcode if users do not want to utilize biometrics such as
fingerprint or face recognition. By 2025, Singapore plans to establish a fully
automated immigration clearance system for all travelers, including firsttime social visitors.1467 This includes using AI for retinal and facerecognition procedures which could potentially remove the need for
passports.
In 2020, Singapore had introduced TraceTogether, Covid-19 contact
tracing application, advising that data would "never be accessed unless the
user tests positive" for the virus.1468 In January 2021, Minister of State for
Home Affairs stated that under the Criminal Procedure Code, the Singapore
Police Force can obtain any data and TraceTogether and SafeEntry app data
CNA, Parliament passes Bill to deal with foreign interference after 10-hour debate
(Oct. 5, 2021), <https://www.channelnewsasia.com/singapore/fica-parliament-singaporeforeign-interference-countermeasures-bill-2221236
1465
Smart Nation Singapore: The Way Forward (June 2, 2020) https://smartnationstrategy.opendoc.sg/08-strengthen-collaboration.html
1466
SmartNation, National Digital Identity,
https://www.smartnation.gov.sg/initiatives/strategic-national-projects/national-digitalidentity
1467
The Strait Times, All immigration checkpoints to have fingerprint and face scans by
2025 as part of Singapore's AI push (Nov. 15, 2019),
https://www.straitstimes.com/singapore/fingerprint-and-face-scans-at-all-immigrationcheckpoints-by-2025-as-part-of-singapores-ai
1468
TraceTogether, What data is collected? Are you able to see my personal data?,
https://support.tracetogether.gov.sg/hc/en-sg/articles/360043735693-What-data-iscollected-Are-you-able-to-see-my-personal-data1464
370
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
are used for criminal probes.1469 In February 2021, The Covid-19
(Temporary Measures) (Amendment) Bill, restricted the use of personal
contact tracing data in criminal investigations to only serious crimes, such
as murder and terrorism, was passed in Parliament.1470 In August, Minister
of State for Home Affairs advised that Singapore aims to have more than
200,000 police cameras by at least 2030, more than doubling its current
use.1471 In September 2021, Singapore police started trialing patrol robots
for surveillance, to detect "undesirable social behaviors" and displaying
messages to educate the public on proper behavior.1472
Singapore also utilizes ABBSS (Automated Biometrics &
Behavioral Screening Suite) at immigration and border checkpoints. The
system is a network of cameras with facial recognition capabilities that can
also be deployed as a body-worn-camera for officers. It is used both to build
a biometrics database of travelers and detect travelers wanted for various
offenses.1473
Public Participation
A National AI Office is created under the Smart Nation and Digital
Government Office. Ministry of Communications and Information provides
public consultation access to legislation under its control.1474 The Smart
Nation and Digital Government Office (SNDGO), under the Prime
Minister’s Office (PMO), provides publicizes information about key Smart
Nation projects government digital transformation on its website.1475
Parliament Sitting 4 January 2021: Minute 33:40.
https://www.youtube.com/watch?v=LjhIegyyLHk
1470
The Straits Times, Bill limiting police use of TraceTogether data to serious crimes
passed (Feb. 2, 2021), https://www.straitstimes.com/singapore/politics/bill-limiting-useof-tracetogether-for-serious-crimes-passed-with-govt-assurances
1471
Reuters, Singapore to double police cameras to more than 200,000 over next decade
(Aug. 4, 2021), https://www.reuters.com/world/asia-pacific/singapore-double-policecameras-more-than-200000-over-next-decade-2021-08-04/
1472
Reuters, Singapore trials patrol robots to deter bad social behaviour
(Sept. 6, 2021), https://www.reuters.com/technology/singapore-trials-patrol-robots-deterbad-social-behaviour-2021-09-06/
1473
Wong, K. Facial recognition, biometrics tech at more checkpoints: ICA. The Strait
Times (Nov. 13, 2018) https://www.straitstimes.com/singapore/facial-recognitionbiometrics-tech-at-more-checkpoints-ica
1474
Ministry of Communications and Information, Public Consultations
https://www.mci.gov.sg/public-consultations/archived?pagesize=24
1475
Smart Nation Singapore: https://www.smartnation.gov.sg/about-smart-nation/sndgg
1469
371
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD/G20 AI Principles
Singapore is not a member of the OECD or the G20. However, the
country is well aware of the OECD/G20 AI Principles. The OECD noted
several significant examples of positive AI practices in Singapore.1476 There
is, for example, the Advisory Council on the Ethical Use of AI and Data,
described above. The OECD also notes that the AI Governance Framework
incorporates all of the OECD AI Principles. Singapore is a founding
member of The Global Partnership on AI (GPAI), built around a shared
commitment to the OECD AI Principles.1477
Data Protection and Algorithmic Transparency
The Personal Data Protection Act (PDPA) was enacted in 2012 and
came into effect in 2014. It is the baseline law for personal data protection.
1478
Singapore also passed Personal Data Protection (Enforcement)
Regulations 202. Unlike the GDPR, the PDPA does not expressly provide
for Data Protection Impact Assessments ('DPIA') to be carried out, does not
provide data subjects with the right to erasure, or requirement to inform data
subjects of the existence of automated decision-making, including profiling.
It does not explicitly address the right not to be subject to discrimination.
The PDPC is part of the Info-communications Media Development
Authority ('IMDA').1479
The PDP Commission expects AI systems to be human-centric, and
decisions made by or with the assistance of AI to be explainable, transparent
and fair.1480 PDPA, however, does not provide protection against police or
any public agency use of personal data. Public sector agencies are covered
by Government Instruction Manuals and the Public Sector (Governance)
Act (PSGA).Monetary Authority of Singapore (MAS) and financial
industry co-created a set of principles in 2018 to guide the responsible use
of AI, focusing on Fairness, Ethics, Accountability and Transparency
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1477
Government of France, Launch of the Global Partnership on Artificial Intelligence
(June 17, 2020), https://www.gouvernement.fr/en/launch-of-the-global-partnership-onartificial-intelligence
1478
The Personal Data Protection Act (PDPA): https://www.pdpc.gov.sg/Overview-ofPDPA/The-Legislation/Personal-Data-Protection-Act
1479
OneTrust DataGuidance. Comparing privacy laws: GDPR v. Singapore's PDPA
https://www.dataguidance.com/sites/default/files/gdpr_v_singapore_final.pdf
1480
PDPC Singapore, Discussion Paper on Artificial Intelligence (AI) and Personal
Data— Fostering Responsible Development and Adoption of AI (June 5, 2018)
https://www.pdpc.gov.sg/-/media/Files/PDPC/PDF-Files/Resource-forOrganisation/AI/Discussion-Paper-on-AI-and-PD---050618.pdf
1476
372
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
(FEAT)1481. The principles have established a standard across the financial
sector in Singapore. The regulator is now working to create a standardized
modular implementation framework of the FEAT principles, called Veritas
which will provide tools for institutions to validate their models against the
FEAT principles.
Singapore shares publicly available datasets1482 from 70 public
agencies, API library and resources for application developers using these
data sets. Public Sector (Governance) Act 20181483 provides a governance
framework for data sharing among government agencies. It is a step in the
right direction for data quality and improved services. On the commercial
side, the regulator IMDA, introduced a “Trusted Data Sharing Framework”
1484
as a guide to establish safeguards and baseline “common data sharing
language” and systematic approach to understanding the broad
considerations for establishing trusted data sharing partnerships. A Data
Regulatory Sandbox1485 is also offered to businesses to pilot innovative use
of data in a safe “environment”, in consultation with IMDA and PDPC. The
regulator also provides a Data Protection Trustmark (DPTM), a voluntary
enterprise-wide certification for organizations to demonstrate accountable
data protection practices.1486
Singapore has not openly stated its position on a ban of fully
autonomous weapons yet.1487 However, Road Traffic (Autonomous Motor
Vehicles) Rules 2017 regulates liability trials and use of autonomous motor
vehicles1488 Singapore has not endorsed the Universal Guidelines for AI,1489
Monetary Authority of Singapore, Principles to Promote Fairness, Ethics,
Accountability and Transparency (FEAT) in the Use of Artificial Intelligence and Data
Analytics in Singapore’s Financial Sector (2018)
https://www.mas.gov.sg/~/media/MAS/News%20and%20Publications/Monographs%20a
nd%20Information%20Papers/FEAT%20Principles%20Final.pdf
1482
Smart Nation Singapore: Open Data Resources
https://www.smartnation.gov.sg/resources/open-data-resources
1483
Public Sector (Governance) Act 2018: https://sso.agc.gov.sg/Act/PSGA2018
1484
IMDA and PDPC, Trusted Data Sharing Framework (2019)
https://www.imda.gov.sg/-/media/Imda/Files/Programme/AI-Data-Innovation/TrustedData-Sharing-Framework.pdf
1485
IMDA, Data Collaboratives Programme (DCP),
https://www.imda.gov.sg/programme-listing/data-collaborative-programme
1486
IMDA, Data Protection Trustmark Certification,
https://www.imda.gov.sg/programme-listing/data-protection-trustmark-certification
1487
https://www.stopkillerrobots.org/wpcontent/uploads/2020/03/KRC_CountryViews_11Mar2020.pdf
1488
Campaign to Stop Killer Robots, Country Views on Killer Robots (March 11, 2020)
https://sso.agc.gov.sg/Act/RTA1961
1489
The Public Voice, Universal Guidelines for AI (2018), https://thepublicvoice.org/AIuniversal-guidelines/
1481
373
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
or GPA Resolution on AI Accountability.1490 However, Singapore’s second
edition Model AI Governance Framework provides clear practical guidance
that essentially aligns with the GPA Resolution on AI Accountability.1491
Human Rights
Singapore has endorsed the Universal Declaration of Human Rights.
However, it has not adopted several international human rights conventions,
reasoning that it is not in a position to fully implement the obligations
contained in an international treaty before ratifying it. Singapore has an
Inter-Ministerial Committee on Human Rights but no national human rights
institution. Sexual relations between two male persons remains a criminal
offense. There are no legal protections against discrimination based on
sexual orientation or gender identity.1492
Freedom House rates Singapore as partly free.1493 According to
Freedom House, “Singapore’s parliamentary political system has been
dominated by the ruling People’s Action Party (PAP) and the family of
current prime minister Lee Hsien Loong since 1959. The electoral and legal
framework that the PAP has constructed allows for some political pluralism,
but it constrains the growth of credible opposition parties and limits
freedoms of expression, assembly, and association.” On transparency,
Freedom House notes, “The government provides limited transparency on
its operations. The Singapore Public Sector Outcome Review is published
every two years and includes metrics on the functioning of the bureaucracy;
regular audits of public-sector financial processes are also made accessible
to the public.”
Evaluation
Singapore is one of the leaders in providing guidance for ethical
development of AI, providing regulatory sandboxes for testing of
responsible practices and developing risk-based governance frameworks. It
is focused on voluntary adoption of these methods for both public and
Global Privacy Assembly, Adopted Resolution on Accountability in the Development
and Use of Artificial Intelligence (October 2020) https://globalprivacyassembly.org/wpcontent/uploads/2020/10/FINAL-GPA-Resolution-on-Accountability-in-theDevelopment-and-Use-of-AI-EN-1.pdf
1491
IMDA, PDPC, Model: Artificial Intelligence Governance Framework, Second
Edition, https://www.pdpc.gov.sg/-/media/files/pdpc/pdf-files/resource-fororganisation/ai/sgmodelaigovframework2.pdf
1492
Human Rights Watch, World Report 2020 – Singapore (2020)
https://www.hrw.org/world-report/2020/country-chapters/singapore
1493
Freedom House, Freedom in the World 2020 – Singapore (2020),
https://freedomhouse.org/country/singapore/freedom-world/2020
1490
374
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
private use cases. Singapore’s privacy agency has significant
responsibilities for data protection and a growing role in AI policy.
Questions do remain about independent oversight of government AI
systems.
375
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Slovenia
National AI Strategy
On May 27, 2021, the Slovenian Government adopted the National
programme promoting the deployment of AI in the Republic of Slovenia by
2025 (NpUI).1494 The NpUI lays out a detailed workplan for social and
economic development across the government, which include specific
indicators, guidance on measuring progress, and instruments for
implementation and financing.1495 Led by an inter-ministerial working
group, the national program was a result of a series of multi-disciplinary
consultations with national experts and industrial representatives through
the ICT Association of Slovenia1496 of the Chamber of Commerce and
Industry of Slovenia,1497 researchers and practitioners in the field of AI
through the Slovenian Artificial Intelligence Society,1498 and stakeholders
of the Strategic Research and Innovation Partnerships on Smart Cities,1499
Factories of the Future,1500 and civil society through the Slovenian Digital
Coalition.1501 The Ministry of Public Administration prepared the first draft,
which was released soliciting public comments in August 2020.1502
The NpAI is a part of the Development Strategy of Slovenia
1503
2030,
which sets out a plan for digital transformation towards the
“Fourth Industrial Revolution, [which is marked by] the digital economy
Government of the Republic of Slovenia, National Program for the Promotion of the
Development and Use of Artificial Intelligence in the Republic of Slovenia until 2025.
(27 May 2021) (hereinafter “NpAI”) 2-3. Original Slovenian version at
https://www.gov.si/assets/ministrstva/MJU/DID/NpUI-SI-2025.docx
Document translated into English by automated process by onlinedoctortranslator.com on
file with author.
1495
Republic of Slovenia, Digitalisation of Society,
https://www.gov.si/en/topics/digitalisation-of-society/
1496
ICT Association of Slovenia (ZIT), https://www.gzs.si/en_zit.
1497
The Chamber of Commerce and Industry of Slovenia (CCIS).
https://eng.gzs.si/vsebina/About-Us.
1498
Slovenian Artificial Intelligence Society (SAIS), https://slais.ijs.si/.
1499
Strategic Research and Innovation Partnerships on Smart Cities (CRIP SC&C)
1494
http://pmis.ijs.si/wp-content/uploads/2019/02/SRIP_SC_C-1.pdf.
Strategic Research and Innovation Partnerships for Factories of the Future (SRIP
FoF) https://www.effra.eu/jozef-stefan-institute-srip-strategic-research-innovationpartnership.
1501
Slovenian Digital Coalition (SDC) https://www.digitalna.si/en.
1502
European Commission, Knowledge for Policy: AI Watch, Slovenia AI Strategy
Report. https://knowledge4policy.ec.europa.eu/ai-watch/slovenia-ai-strategy-report_en
1503
Government of the Republic of Slovenia, Slovenian Development Strategy 2030 (SrS
2030), Dec. 7, 2017. https://www.gov.si/assets/vladne-sluzbe/SVRK/Strategija-razvojaSlovenije-2030/Slovenian-Development-Strategy-2030.pdf
1500
376
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and the development of sensors, robotics and AI, and establishing new
business, work and job models, and skills and adjustments in many areas of
economic, social and environmental development.”1504 The national AI
strategy is aligned with the EU framework programmes for research,
innovation and deployment such as Horizon Europe and Digital Europe.
The NpAI articulates specific support for Slovenian firms and institutions
that develop standards in the field of AI as well as promoting collaborations
with national, EU and international standardisation organisations.1505
European Union
Slovenia's efforts track with the European Union's plans, having
signed the EU Declaration on Cooperation on Artificial Intelligence.1506 The
National AI Program is also consistent with the European Coordinated Plan
on Artificial Intelligence 2021,1507 which operationalized the Declaration,
and is also going through the process of implementing the proposals of the
Regulation of the European Parliament and of the Council Establishing a
Digital Europe Program for the Period 2021-2027,1508 which proposes to
focus on AI, amongst five priority areas. The latter provides support for the
development and strengthening of basic AI capacities, such as data
resources and libraries of AI algorithms and their accessibility for all
companies, public administration and the wider public sector, and
strengthening and promoting integration between existing R&D capacities
in Member States.1509
Slovenia's efforts are also guided by and in line with the OECD
Principles on Artificial Intelligence, which promote artificial intelligence
that is innovative and trustworthy and respects human rights and
democratic values.1510
SrS 2030, 10.
Ibid.
1506
EU Declaration on Cooperation on Artificial Intelligence,
https://ec.europa.eu/jrc/communities/en/community/digitranscope/document/eudeclaration-cooperation-artificial-intelligence
1507
European Commission, Coordinated Plan on Artificial Intelligence 2021,
https://digital-strategy.ec.europa.eu/en/policies/plan-ai.
1508
Proposal for a Regulation Establishing the Digital Europe Programme 2021-2027
(June 14, 2018), https://www.europarl.europa.eu/legislative-train/theme-a-europe-fit-forthe-digital-age/file-mff-digital-europe-programme
1509
NpAI, 9.
1510
NpAI, 9.
1504
1505
377
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
European Council Presidency
2021 has been an important year for Slovenia because of its
leadership of the Council of the EU Presidency in the latter half of 2021.1511
The Slovenian Digital Minister Boštjan Koritnik articulated that their
priority for AI regulation “is to ensure flow of data within the EU and across
sectors in line with clear and fair rules for access storage and reuse, … we
have to increase the control ... and of course, trust of citizens and companies
regarding their data.”1512
After the first debate on the European Parliament’s proposed AI Act,
Koritnik announced that the EU's AI act should “serve as a model across
the globe, in the same vein as the general data protection regulation, GDPR,
in the area of protection of personal data.”1513 He continued to further
explain that “we want to make sure that the Artificial Intelligence Act will
achieve its twin aims of ensuring safety and respect for fundamental rights
and stimulating the development and uptake of AI-based technology in all
sectors."1514
International Research Center on Artificial Intelligence (IRCAI)
On March 29, 2021, Borut Pahor, President of the Republic of Slovenia,
and Audrey Azoulay, Director-General of UNESCO, inaugurated the
International Research Centre on Artificial Intelligence (IRCAI), as a
Category 2 centre under the auspices of UNESCO in Ljubljana,
Slovenia.1515 The IRCAI is designed to be a communication platform for
the collection and dissemination of good practices and case studies on the
use and deployment of AI in society.1516 The IRCAI will focus on advancing
1511
Slovenian Presidency of the Council of the European Union, 1 July–31 December
2021, https://www.gov.si/assets/vlada/Projekti/PSEU2021/The-programme-of-theSlovenian-Presidency-of-the-Council-of-the-European-Union.pdf
1512
Clothilde Goujard and Leonie Cater, Politico. Slovenia eyes quick wins on AI during
EU presidency, May 21, 2021. https://www.politico.eu/article/slovenia-ai-digital-eupresidency-bostjan-koritnik/
1513
Euronews, EU’s artificial intellgience law should serve as ‘model across the globe’
https://www.euronews.com/2021/10/14/eu-s-artificial-intelligence-law-should-serve-asmodel-across-the-globe
1514
Ibid.
1515
UNESCO, UNESCO Director-General and President of Slovenia inaugurate first
research centre on artificial intelligence (Apr. 6, 2021)
https://en.unesco.org/news/unesco-director-general-and-president-slovenia-inauguratefirst-research-centre-artificial
1516
Van Roy, V., Rossetti, F., Perset, K. and Galindo-Romero, L., AI Watch - National
strategies on Artificial Intelligence: A European perspective, 2021 edition, EUR 30745
EN, Publications Office of the European Union, Luxembourg, 2021,
https://publications.jrc.ec.europa.eu/repository/handle/JRC122684.
378
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
research on the use of AI in order to help achieve the Sustainable
Development Goals (SDGs).1517 The core research functions of the Centre
will be guided by its four international scientific committees on:
● AI and Climate that will address the issue of water quality
measurement;
● AI and Education that will focus on AI algorithms that can make
Open Educational Resources more accessible and easier to use;
● AI and Assistive Technologies that will highlight the potential of
using AI technologies to assist persons with disabilities; and
● AI and Healthcare that will focus on the use of AI in vaccine
development processes.1518
Privacy and Data Protection
The Slovenian Constitution of 1991 guarantees the protection of
personal data at the constitutional level and within the framework of
guaranteed human rights. This right is explicitly enshrined in Article 38.1519
In Slovenia the GDPR is not implemented by the national law. Slovenia
remains the only EU country without the implementing act. As a
consequence, the Information Commissioner is without the power to impose
fines and therefore has not been able to impose a single administrative fine
since the adoption of the GDPR. However, a recent report indicates that
Slovenia may soon enact a data protection law.1520 Accordimng to another
report, the current draft follows the GDPR and only amends a few
aspects.1521
Algorithmic Transparency
Slovenia is a member of the Council of Europe but has not ratified
the modernized Privacy Convention.1522
International Research Center on Artificial Intelligence (IRCAI) Launch Report,
April 2021 https://ircai.org/wp-content/uploads/2021/05/IRCAI_Launch2021_Report.pdf
1518
Id.
1519
GDPRhub, Data Protection in Slovenia,
https://gdprhub.eu/Data_Protection_in_Slovenia
1520
Lexology, Slovenia finally one step closer to the new Data protection act (Jan. 22,
2022) https://www.lexology.com/library/detail.aspx?g=a3406dc8-6ad4-4d5f-b7276a84253938a5
1521
DLA Piper, Data Protection Laws of the World – Slovenia (Jan. 17, 2022),
https://www.dlapiperdataprotection.com/index.html?t=law&c=SI
1522
Council of Europe, Treaty Office: Chart of signatures and ratifications of Treaty 223
(Feb. 2, 2022), https://www.coe.int/en/web/conventions/full-list?module=signatures-bytreaty&treatynum=223
1517
379
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD/G20 AI Principles
Slovenia is a member of the OECD and has endorsed the OECD AI
Principles.1523 The OECD noted Slovenia’s Digital Coalitona and
AI4Slovenia as an example of national AI polices that are shaping an
enabling environment for AI.1524 Slovenia is also a founding member of the
Global Partnership on AI. 1525
Human Rights
According to Freedom House, Slovenia receives high scores for
Political Rights and Civil Liberties and a slight uptick in 2021 for a
combined score of 95/100.1526 The report notes that the government
“generally operates with openness and transparency.” Slovenia is a
signatory to major human rights treaties.1527 However, a report by the
Greens/EFA of the European Parliament,1528 revealed that Slovenia, as well
as in 10 other EU member states, use facial recognition technologies for
‘ex-post identification’ in their criminal investigations.1529 In Slovenia, the
use of face recognition technology by the police was legalized five years
after they first had already started using it.1530 The Slovenian Presidency of
the Council has also accelerated negotiations on a vast expansion of the
Eurodac database, which will hold sensitive data on millions of asylum
seekers and migrants in an irregular situation, by ‘delinking’ the proposed
OECD, Forty-two countries adopt new OECD Principles on Artificial Intelligence
(May 22, 2019), https://www.oecd.org/science/forty-two-countries-adopt-new-oecdprinciples-on-artificial-intelligence.htm
1524
OECD, State of Implementation of the OECD AI Principles: Insights from National
AI Policies (June 2021), https://www.caidp.org/app/download/8328376363/OECDAIpolicies-2021.pdf
1525
GPAI: Community, https://gpai.ai/community/
1526
Freedom House, Freedom in the World 2021 – Slovenia,
https://freedomhouse.org/country/slovenia/freedom-world/202,
1527
Office of the United Nations High Commissioner for Human Rights, Treaty
Ratification Status for Slovenia,
https://tbinternet.ohchr.org/_layouts/15/TreatyBodyExternal/Treaty.aspx?CountryID=159
&Lang=EN
1528
Ragazzi, F., Kuskonmaz, E., PlIájás, I., Van de Ven, R., Wagner, B., report for the
Greens/EFA in the European Parliament, Behavioural Mass Surveillance in EU Member
States, October 2021, http://extranet.greens-efa.eu/public/media/file/1/7297
1529
Luca Bertuzzi, Euractiv, Facial recognition technologies already used in 11 EU
countries and counting, report says, Oct. 26, 2021.
https://www.euractiv.com/section/data-protection/news/facial-recognition-technologiesalready-used-in-11-eu-countries-and-counting-report-says/
1530
Lenart J. Kučić, Algorithm Watch, Automating Society Report 2020: Slovenia,
October, 2020. https://automatingsociety.algorithmwatch.org/report2020/slovenia/
1523
380
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
rules from other EU asylum and migration laws under discussion.1531 The
database uses VeriLook and Face Trace software from the Lithuanian
company Neurotechnology and is managed by the Ministry of Interior
(Slovenia).1532
Evaluation
For a small and relatively young country, Slovenia has had an
outsized influence over the development of policy in the past year due to its
leadership of the presidency of the European Council for the second half of
2021, as well as its AI collaboration with UNESCO, and its work in support
of the Council of Europe AI expert group, the CAHAI. While the national
AI strategy and other technology-related workplans pledge to implement a
legislative and regulatory regime that also protects fundamental freedoms,
there are necessary reforms pending of its data protection law and a need
for better alignment of its laws with surveillance practices, especially by
law enforcement.
Statewatch, in European Digital Rights Initiative (EDRi), Eurodac: Council seeks
swift agreement on expanded migrant biometric database, Sept. 22, 2021.
https://edri.org/our-work/eurodac-council-seeks-swift-agreement-on-expanded-migrantbiometric-database/
1532
Behavioural Mass Surveillance in EU Member States, October 2021,
http://extranet.greens-efa.eu/public/media/file/1/7297
1531
381
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
South Africa
National AI Strategy
There is no specific national AI strategy in South Africa but there
are piecemeal AI strategies entrenched within key national plans and
policies. The Presidential Commission on the Fourth Industrial Revolution
(PC4IR) Summary Report and Recommendations (the PC4IR Report)
released in January 2020 which provides guidance on the fourth industrial
revolution, the impact on South Africa and recommends actions for the
future includes an AI strategy.1533 The recommendations provide an
industrial development strategy that prescribes among other things focus on
regulation, ethics, and cultural aspects of the Internet not only for an
enabling policy environment but to also ensure ethical and transparent use
of new technologies.1534 The PC4IR report also focuses on technological
developments which are part of the fourth industrial revolution. The
developments move towards computer use of sense-making which is
facilitated by large amounts of data leading to reliance on algorithms.
Furthermore, technological developments are in clusters that include
artificial intelligence and robotics.1535
The National Planning Commission published the National
Development Plan: Vision for 2030 (NDP) which aims to achieve improved
information and communication technologies (ICTs) by 2030.1536 Part of
the strategy is to enhance ICTs through a national e-strategy that places
South Africa at an international stage with international governance
agencies such as the Information Telecommunications Union (ITU) and
World Trade Organisation (WTO).1537 The aim for an improved ICT sector
is to bridge the digital divide, boost economic activity, and improve the
1533
Presidential Commission on the 4th Industrial Revolution (PC4IR) ‘Summary Report
and Recommendations’ 114-24 (2020)
https://www.gov.za/sites/default/files/gcis_document/202010/43834gen591.pdf. The 4th
Industrial Revolution South Africa partnership (4IRSA) - an alliance between partners
from the public and private sectors, academia and civil society launched by President
Ramaphosa in 2019, reaffirms a national push towards promoting the digital economy for
growth.
1534
Id.
1535
Id. at 28.
1536
National Planning Commission (NPC) ‘National Development Plan’ (2012). The
NPC was established in May 2010 to develop a long-term vision and strategic plan for
South Africa. Its main objective is to gather the nation around a common set of objectives
and priorities to drive development over the longer term. The Commission advises the
government on cross- cutting issues that influence the long-term development of South
Africa.
1537
Id. at 195-96.
382
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
education, health, transport and labour sectors. The National Development
Plan enunciates this vision by laying out the need to bring in growth in
digital technologies as part of the growth in ICTs.1538
Strides in Innovation
Based on results of the Business Innovation Survey (BIS) 20142016 that were released by the Minister of Higher Education, Science and
Innovation, Dr Blade Nzimande, two-thirds (69,9%) of South African
businesses are innovation-active.1539 Financial barriers were cited as
affecting innovation among other challenges market-related.1540 An
innovation-active business sector reflects a growing interest in emerging
technologies exploring the use of artificial intelligence. The Department of
Trade and Industry in South Africa in partnership with the Production
Technologies Association of South Africa formed a multi-stakeholder
initiative, Intsimbi Future Production Technologies Initiative which
implements a ‘turnaround strategy for South Africa’s distressed tooling
industry.’1541 The initiative is a technological solution which helps
overcome financial barriers through employing technology in business.1542
The financial technology intergovernmental stakeholders in South Africa
have a vision of leading in promoting financial inclusion while spurring
competition, digital skills, and economic growth through innovation.1543
Cities in South Africa are embracing technology and datafication, a
new mode of informing decision making in the quest to develop smart cities.
The City of Cape Town launched an Open Data Portal in 2015 which
facilitates open access to data for public use.1544 Similarly, eThekwini
Municipality’s Economic Development and Growth (EDGE)is an open data
portal with 14 datasets which focus on economic aspects of the city like
labour data, property data, electricity data, business licensing data, and
Id. at 3.
South African Government ‘Minister Blade Nzimande releases results of Business
Innovation Survey’ (2020) https://www.gov.za/speeches/minister-blade-nzimanderesults-business-innovation-survey-9-jul-2020-0000.
1540
Id.
1541
About Insimbi http://www.intsimbi.co.za/about.html.
1542
Id.
1543
Intergovernmental Fintech Working Group ‘South Africa Fintech Vision.’ The
Intergovernmental Fintech Working Group (IFWG), formed in 2016, comprises
representatives from the National Treasury (NT), South African Reserve Bank (SARB),
Financial Sector Conduct Authority (FSCA), National Credit Regulator (NCR), Financial
Intelligence Centre (FIC), and South African Revenue Service (SARS).
1544
South African Cities Network Smart cities paper series: Smart governance in South
African cities 18 https://www.sacities.net/wpcontent/uploads/2020/10/Smart_Cities_Papers_Volume_1_Final-Draft.pdf
1538
1539
383
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
educational data.1545 There is also great reliance on data collected for
addressing healthcare and allocation of other services.1546 While the benefits
are to increase development and service delivery in cities, reliance on
datasets and algorithms in smart city initiatives must be applied in a way
that is transparent to avoid exclusion of other members of South African
society.1547
Data Protection
In 2020, Experian, a South African company experienced a data
breach that affected over 24 million people.1548 Data breaches have an effect
of making the public lose confidence in the companies safeguarding
personal information belonging to data subjects. This may be reflected in
negative public opinions. South Africa’s data protection law, the Protection
of Personal Information Act (POPIA), 2013 came into effect on 1 June 2020
and has been fully operational since 1 July 2021 and all relevant entities are
expected to comply with the provisions.1549 POPIA protects personal
information and is meant to give effect to the right to privacy guaranteed
under section 14 of the Constitution of South Africa, 1996.
Section 71(1) of POPIA stipulates that a data subject may not be
subject to a decision which results in legal consequences which affect the
data subject based entirely on automated decision-making.1550 This
protection ensures that profiling based on factors including performance,
credit worthiness, reliability, location, health, personal preferences or
conduct is not used in isolation of any other considerations. Section 71(2)
provides exceptions where the sole reliance on automated decision-making
is governed by a contract, law or code of conduct. Furthermore, section
71(3) provides that in dealing with the exceptions, a responsible party must
provide a data subject with sufficient information about the underlying logic
of the automated processing of the information the data subject must have
an opportunity to make representations about a decision.
Id. at 18.
Policy Action Network, AI and data in South Africa’s cities and towns: Centering the
citizen 11 (2020),
https://policyaction.org.za/sites/default/files/PAN_TopicalGuide_AIData4_CitiesTowns_
Elec.pdf.
1547
Id. at 10.
1548
Experian, Experian South Africa open letter from Ferdie Pieterse (2020)
https://www.experian.co.za/fraudulent-data-incident/open-letter-from-ferdie-pieterse .
1549
Protection of Personal Information Act (POPIA) No.4 of 2013 (POPIA).
1550
Id. at sec 71(1).
1545
1546
384
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
POPIA also ensures that before a responsible party1551 decides to
process any unique identifiers of data subjects for any other purpose other
than that which is agreed upon by data subjects and with the purpose of
linking up the information gathered with information collected by other
responsible parties, the responsible parties must obtain prior authorisation
from the Regulator. 1552For companies intending to rely on machine
learning, there is a need to first get authozisation in line with section 57(1)
of POPIA. The safeguard built in POPIA is meant to ensure responsible use
of personal information in automated decision-making. It is hoped that
beyond 1 July 2021, companies will ensure a more human-centered
approach to protecting the privacy of data subjects thereby increasing public
trust of technology. This has an impact on how artificial intelligence and
other emerging technologies are perceived.
COVID-19 Surveillance
The University of Witwatersrand recently designed an AI-based
algorithm in partnership with iThemba Labs, the Provincial Government of
Gauteng and York University in Canada to show the risk incidence of the
third wave of COVID-19 in South Africa. The AI driven early detection
predicts the future daily confirmed cases based on the past infection history
including elements such as ‘mobility indices, stringency indices and
epidemiological parameters.’ In this case, AI is used as an early warning
mechanism. AI has to be applied well with human reasoning, noting that the
system is predictive and not conclusive.
Research and Development on AI
The Department of Science and Innovation funds the Centre for
Artificial Intelligence Research (CAIR) which is a research network of nine
groups across six main universities in South Africa focusing on AI.1553 This
initiative is critical to the development of emerging technologies as research
promotes AI and enhances socio-economic developments.1554 The CAIR is
a great platform for conducting such research as it comprises experts in the
field of AI. The CAIR harnesses expertise from leading researchers from
the University of Cape Town, the University of KwaZulu-Natal, NorthId. at sec 1. A responsible party is defined in section 1 of POPIA as a public or
private body or any other person which, alone or in conjunction with others, determines
the purpose of and means for processing personal information.
1552
n 15 above, sec 57(1).
1553
Centre for Artificial Intelligence https://www.cair.org.za/about.
1554
Centre for The Fourth Industrial Revolution South Africa, About the WEF Affiliate
Centre for 4IR South Africa (C4IR SA) 6 (2020), https://www.c4ir-sa.co.za/wpcontent/uploads/2020/12/C4IR-SA_-Introduction_-November-2020-4.pdf.
1551
385
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
West University, the University of Pretoria, Stellenbosch University and the
University of the Western Cape. CAIR conducts ‘foundational, directed and
applied research’ on AI through nine research groups: Adaptive and
Cognitive Systems, AI and Cybersecurity, AI for Development,
Applications of Machine Learning, Computational Logic, Ethics of AI,
Foundations of Machine Learning, Knowledge Representation and
Reasoning, and Probabilistic Modelling.1555
Another organization doing research on AI is the Council for
Scientific and Industrial Research (CSIR) which is a leading scientific and
technology research organization that researches, develops, localizes and
diffuses technologies to accelerate socioeconomic prosperity in South
Africa.1556 CSIR supports the public and private sector through specific
research on technology. One of its current projects is the autonomous
exploration mapping project which developed an experimental system for
autonomous exploration and mapping in which a robot autonomously
explores an unknown environment and produces a map.1557
The African Commission on Human and Peoples’ Rights (ACHPR)
adopted Resolution 473 having recognized that emerging technologies such
as AI have a bearing on the enjoyment of human rights under the African
Charter on Human and Peoples’ Rights (the African Charter).1558 This
adoption was cognizant of the challenges posed by autonomous systems that
are not under meaningful human control and the use of algorithms in
Google, Amazon, Facebook and Android and as such, sought to promote
multi-disciplinary research on AI and other emerging technologies. The
ACHPR called on state parties to the African Charter, South Africa
included, to among other things, do the following:
● ensure that the development and use of AI, robotics and other new
and emerging technologies is compatible with the rights and duties
in the African Charter and other regional and international human
rights instruments, in order to uphold human dignity, privacy,
equality, non-discrimination, inclusion, diversity, safety, fairness,
transparency, accountability and economic development as
n 18 above.
CSIR, The CSIR in Brief, https://www.csir.co.za/csir-brief.
1557
CSIR, Autonomous Exploration and Mapping, https://www.csir.co.za/autonomousexploration-and-mapping.
1558
African Commission on Human and Peoples’ Rights Resolution 473 on the need to
undertake a Study on human and peoples’ rights and artificial intelligence (AI), robotics
and other new and emerging technologies in Africa (Feb. 25, 2021),
https://www.achpr.org/sessions/resolutions?id=504
1555
1556
386
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
underlying principles that guide the development and use of AI,
robotics and other new and emerging technologies.
● to ensure transparency in the use of AI technologies, robotics and
other new and emerging technologies and that decisions made in
the use of AI technologies, robotics and other new and emerging
technologies are easily understandable to those affected by such
decisions.
● to work towards a comprehensive legal and ethical governance
framework for AI technologies, robotics and other new and
emerging technologies so as to ensure compliance with the African
Charter and other regional treaties.
Public Participation
South Africa's Department of Cooperative Governance and
Traditional Affairs (CoGTA) partnered with a private company to launch
GovChat, an online citizen engagement application in 2018 which promotes
accountable local governance and allows citizens to engage with their local
councilors.1559 GovChat integrates AI technologies in its design and is
aimed at enhancing local governance in South Africa.1560 In April 2019, the
Department of Science and Innovation signed an agreement with the World
Economic Forum (WEF) to set up the Centre for the Fourth Industrial
Revolution Network (C4IR Network) affiliate Centre in South Africa
hosted by the Council for Scientific and Industrial Research (CSIR).1561
C4IR South Africa works with core members, project partners and
knowledge partners who include academia, civil society groups, small
businesses and government. Knowledge partners are invited to participate
in working groups or discussions. C4IR South Africa mainly focuses on
policy developments in the following areas:1562
▪ Internet of Things (IoT), Robotics and Smart Cities.
▪ Blockchain and Distributed Ledger Technology.
▪ Artificial Intelligence and Machine Learning.
▪ Data Policy.
Human Sciences Research Council, AI technologies for responsive local government
in South Africa (2019), http://www.hsrc.ac.za/en/research-outputs/view/10337.
1560
Republic of South Africa, Government Communications, GovChat,
https://www.salga.org.za/SALGA%20National%20Communicators%20Forum%20Web/
Documents/GovChat%20Presentation.pdf.
1561
Centre for The Fourth Industrial Revolution South Africa, The Network for Global
Technology Governance https://www.c4ir-sa.co.za/.
1562
Centre for The Fourth Industrial Revolution South Africa, About the WEF Affiliate
Centre for 4IR South Africa (C4IR SA) (2020) https://www.c4ir-sa.co.za/wpcontent/uploads/2020/12/C4IR-SA_-Introduction_-November-2020-4.pdf.
1559
387
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Council for Scientific and Industrial Research (CSIR) and Deloitte
South Africa signed a memorandum of understanding to support the World
Economic Forum Affiliate Centre for the Fourth Industrial Revolution
South Africa (C4IR South Africa) in ensuring the application of emerging
technologies.1563 This support should enhance policy developments on AI.
Public Opinion
In a survey conducted by SIENA researchers in 2019 on South
African’s awareness of robots, 36 % of the respondents hardly knew about
robots and 23% had never heard of AI.1564 The lack of awareness of AI
results in a lack of trust in AI. There is a need for more awareness raising,
highlighting benefits where there is responsible use of AI and emphasizing
the rights-based approach. The South African Social Attitudes Survey
revealed that out of 2763 participants, 57% were concerned that the
government did not have effective strategies to ensure that jobs are not lost
as a result of the fourth industrial revolution.1565 Less than a fifth of
respondents were entertaining the idea of a robot administering a medical
procedure on them or driving them showing a low acceptance rate of AI.
OECD/G20 Principles
South Africa forms part of the G20 which adopted the G20 humancentered AI principles (AI Principles) which highlight the need to foster
public trust and confidence in AI technologies.1566 Furthermore, the AI
principles highlight the challenges in the use of AI including privacy,
security, ethical issues, new digital divides and the need for AI capacity
building. These principles are drawn from the Organisation for Economic
Co-operation and Development (OECD). This adoption when viewed
against the backdrop of occurring breaches emphasizes the need for
implementation of policies that protect human rights.
1563
Creamer Media’s Engineering News, CSIR, Deloitte ink MoU for 4IR centre
collaboration (2021), https://www.engineeringnews.co.za/article/csir-deloitte-ink-moufor-4ir-centre-collaboration-2021-06-09
1564
SIENNA, D4:5 Public views on artificial intelligence and robots across 11 EU and
non-EU countries (2020), https://zenodo.org/record/4068220#.YIGoDegzbIV. The
SIENA Project looks at stakeholder-informed ethics for new technologies with high
socio-economic and human rights impact.
1565
4IRSA, Digital revolution has South Africans worried: survey
https://www.4irsa.org/south-africa-4-0/digital-revolution-has-south-africans-worriedsurvey/
1566
G20, Ministerial statement on trade and digital economy 3 (2019),
https://trade.ec.europa.eu/doclib/docs/2019/june/tradoc_157920.pdf (accessed 28 April
2021).
388
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Human Rights
The African Charter on Human and Peoples’ Rights (the Charter) to
which South Africa is a state party, mandates the African Commission on
Human and Peoples’ Rights (ACHPR) with the role of interpreting the
Charter.1567 On February 25, 2021, the ACHPR passed a resolution on the
need to undertake a study on human and peoples’ rights and AI, robotics
and other new and emerging technologies in Africa.1568 The resolution
acknowledges the need for comprehensive and multidisciplinary research
on the legal, ethical, safety and security opportunities and challenges raised
by AI technologies, robotics and other new and emerging technologies in
Africa. It also acknowledges the likely impact of AI on the right to privacy,
and socio-economic rights in particular, as provided for in the Charter such
as the right to work, education, health, social security and access to social
services.
Evaluation
Artificial intelligence (AI) is a growing phenomenon in Africa.
Based on slow development in information and communication technology,
growth in AI policy is happening at a slower rate than in other continents.
In South Africa, some notable policy developments are in place which show
a growth in use of emerging technologies. Many factors affect this growth
such as lack of awareness on AI, lack of relevant skills, limited resources to
stimulate development in AI and fear of the adverse effects of technology
such as concerns of human jobs being replaced by AI and automation. South
Africa human rights violations in the form of data breaches are a cause for
concern in South Africa.1569 Data breaches can negatively affect how
emerging technologies are viewed and interfere with public perceptions.
Nevertheless, AI has the potential to transform service delivery in some
notable instances such as the roll-out of COVID-19 health surveillance.1570
The National Development Plan lays down principles for upholding
human rights in embracing artificial intelligence in South Africa. It is
recommended that South Africa develops a specific national AI strategy that
best suits the needs of the country. It also focuses specifically on measures
African Charter on Human and Peoples’ Rights, articles 30 and 45(3).
n 20 above.
1569
ESI Africa, Data breaches are becoming a common phenomenon in South Africa
(2020) https://www.esi-africa.com/industry-sectors/smart-technologies/data-breachesare-becoming-a-common-phenomenon-in-south-africa/
1570
University of Witwatersrand, AI-powered Algorithm released to detect the third wave
in South Africa (2021) https://www.wits.ac.za/news/latest-news/researchnews/2021/2021-03/ai-powered-algorithm-released-to-detect-the-third-wave-in-southafrica.html.
1567
1568
389
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
that are aimed at enhancing the use and development of AI. In the meantime,
through piecemeal AI strategies found in the PC4IR Report and NDP,
economic growth and development of ICTS will be fostered. Hopefully,
more specific policies which address emerging technologies such as the use
of autonomous weapons and reliance on machine learning will also be
developed. There is a need for more awareness-raising in South Africa to
improve public understanding of AI.
390
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Spain
National AI Strategy
Spain has identified AI as “one of the disciplines most likely to
influence the rapid transition to a new society and economy.”1571 On
December 2, 2020. 1572 Spain unveiled the National Strategy for Artificial
Intelligence. The Spanish government stated it will allocate €600 million
for the implementation of the Strategy from 2021 to 2023. 1573 The objective
of the National Strategy for AI is to generate trust in the development of
inclusive and sustainable AI which focuses on the needs of the citizens. The
AI Strategy’s focuses on six goals:
1) invigorating scientific research, technical development, and
innovation of AI;
2) promoting the development of digital capabilities, encouraging
national talent and attracting global talent in AI;
3) developing data platforms and infrastructure technology to
provide support to AI;
4) integrating IA in value chains to transform the economy;
5) encouraging the use of AI in public administration and in
strategic national missions;
6) and establishing an ethical and normative framework to
strengthen the protection of individual and collective rights and
to guarantee inclusion and social wellbeing.
Spain’s National AI Strategy sets out five measures for the
successful ethical framework:
1) the development of a national stamp of quality for AI,
2) the creation of observatories for ethical and juridical evaluation
of AI systems,
3) the development and launch of Digital Rights Charter,
4) the implementation of a national governance model for ethics in
AI though the AI Advisory Council,
Government of Spain & Ministry of Science, Innovation, and Universities, Spanish
RDI Strategy in Artificial Intelligence (2019),
https://www.ciencia.gob.es/stfls/MICINN/Ciencia/Ficheros/Estrategia_Inteligencia_Artif
icial_EN.PDF
1572
Government of Spain, Estrategia Nacional de Inteligencia Artificial (Dec. 2, 2020),
https://www.lamoncloa.gob.es/presidente/actividades/Documents/2020/021220ENIA.pdf
1573
Government of Spain, Pedro Sánchez presenta la Estrategia Nacional de Inteligencia
Artificial con una inversión pública de 600 millones en el periodo 2021-2023 (Dec. 2,
2020), https://www.lamoncloa.gob.es/presidente/actividades/Paginas/2020/021220sanchezenia.aspx
1571
391
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
5) and the promotion of multisectoral national and international
fora for dialogue, awareness, and participation.
The National AI Strategy is part of several overlapping initiatives
launched by the Spanish government. The Spanish government has
published its 2025 Digital Agenda, which enumerates 50 measures for the
Spanish government to take between 2020 and 2025 across 10 thematic axes
in order to propel the country’s digital transformation. The ninth axis is
“Data Economy and Artificial Intelligence” and one of the measures
contained therein is the accomplishment of the goals set out by the Strategy.
The Strategy is also part of the Plan for Recovery, Transformation,
and Resilience for the Spanish Economy launched in April 2021. The
Strategy is part of the Plan’s measures on science and innovation,
particularly as they relate to reinforcing the National Health System’s
capacities. The Plan also describes Spain’s ambitions to be a global leader
in the integration of artificial intelligence into a digitalized economy and its
focus on developing this technology in a human-centric way. The Plan titles
this a “humanist digitalization”, which it seeks to accomplish through the
Charter of Digital Rights.
The Strategy intersects with the Spanish government’s 2015 Plan
for the Advancement of Language Technologies. Within the framework of
the Plan, the Spanish government, in collaboration with the Barcelona
Supercomputing Center, created the world’s first massive Spanish-language
AI system (MarIA) in 2021, which can generate and analyze texts in
Spanish. This is part of Spain’s strategic objectives to promote the
development of AI systems in Spanish instead of relying on the more
ubiquitous English-language systems.
The National Strategy for AI follows from the Spanish Research,
Development and Innovation (RDI) Strategy for Artificial Intelligence,
published in 2019.1574 The Spanish RDI Strategy for AI sets out Priorities
and Recommendations “to be developed in initiatives and activities defined
and financed through the Science, Technology and Innovation Stares Plans
(PECTI), mobilizing the synergies between the different levels of public
administration and through the co-development of the public and private
sectors.” The Spanish RDI Strategy included plans to create a National AI
Strategy, a Spanish AI Observatory, and a strategic framework for the
development of AI in compliance with the “ethical, legal, and social
Government of Spain & Ministry of Science, Innovation, and Universities, Spanish
RDI Strategy in Artificial Intelligence (2019),
https://www.ciencia.gob.es/stfls/MICINN/Ciencia/Ficheros/Estrategia_Inteligencia_Artif
icial_EN.PDF
1574
392
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
commitments” of Spain and the European environment. Priority six of the
RDI Strategy is to “analyze the ethics of AI from the perspective of RDI.”
Spain also endorses the ethical standards proposed by the European
Union. Spain particularly support the ethical guidelines put forward by the
EU’s High-Level Expert Group (HELG) on Artificial Intelligence in 2018.
The EU HLEG AI Guidelines encourage the development of “Trustworthy
AI” defined as AI which “(1) should respect fundamental rights, applicable
regulation and core principles and values, ensuring an “ethical purpose” and
(2) should be technically robust and reliable since, even with good
intentions, a lack of technological mastery can cause unintentional
harm.”1575
To track the development of AI and ensure the carrying out of the
RDI goals, Spain has created a dynamic Map of AI Technology Capabilities
which “tracks and displays essential information on the entities which
develop, investigate, use, or lend services with AI technologies, on a
national level as well as at the level of autonomous communities and
provinces.”1576 The information on the map details Spanish technical AI
capacity and informs on areas of priority in AI to meet RDI objectives. The
Map of AI Technology Capabilities detail the use of AI in public agencies,
private businesses, institutions of higher learning, and private non-profit
institutions. The Map notes that AI in Spain is currently most used for big
data and data analysis as well as machine learning. This information is
publicly available, and additions can be submitted by the public.
In October 2020, the Spanish government, along with thirteen other
countries, published a position paper on innovative and trustworthy AI.1577
This paper delineates a two-fold vision of the EU’s AI development seeking
to promote innovation while managing risks through a clear framework and
establish trustworthy AI as a competitive advantage. They state that “The
main aim must be to create a common framework where trustworthy and
human-centric AI goes hand in hand with innovation, economic growth and
competitiveness in order to protect our society, maintain our high-quality
The European Commission’s High Level Expert Group on Artificial Intelligence,
Draft: Ethics Guidelines for AI (2018), https://ec.europa.eu/futurium/en/ai-allianceconsultation/guidelines#Top
1576
Government of Spain, Mapa de capacidades de tecnologías de IA,
https://mapa.estrategiaia.es/
1577
Position Paper on Behalf of Denmark, Belgium, the Czech Republic, Finland, France
Estonia, Ireland, Latvia, Luxembourg, the Netherlands, Poland, Portugal, Spain and
Sweden, Innovative and Trustworthy AI: Two Sides to the Same Coin (Oct.
2020), https://www.permanentrepresentations.nl/binaries/nlatio/documents/publications/2
020/10/8/non-paper---innovative-and-trustworthy-ai/Non-paper++Innovative+and+trustworthy+AI+-+Two+side+of+the+same+coin.pdf)
1575
393
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
public service and benefit our citizens and businesses. This can help the EU
to protect and empower their citizens, underpin innovation and progress in
society and ensure that their values are protected.”
According a report of the European consumer organization BEUC,
83% of those in Spain think that consumers should be well informed when
they deal with an automatic decision system and 80% believe they should
have the right to say “no” to automated decision-making.1578 The BEUC
report also found high levels of concern in Spain about the potential failure
of AI machines, the privacy of voice assistants, manipulation of consumer
decisions, and unfair discrimination. More than half of those from Spain
surveyed in the BEUC poll “disagree or strongly disagree that current
regulation is adequate to efficiently regulate AI.”
Artificial Intelligence Advisory Council
In July of 2020, the Ministry of Economic Affairs and Digital
Transformation established the Artificial Intelligence Advisory Council.1579
The Council is composed of Spanish experts in science, economics,
education, and other relevant fields. The AI Advisory Council was created
to analyze, assess, and support the government on matters of AI. According
to the Ministry, the Advisory Council is intended to provide
recommendations to the government on measures for the safe and ethical
use of AI. The Council members will analyze the implications of AI in
different areas, such as industry, the future of work, protection of
fundamental rights, data management, the fight against discrimination, and
the elimination of social disparities.
Artificial Intelligence Supervision Agency
In 2021, the Spanish government agreed to the creation of an
Artificial Intelligence Supervision Agency. The agency’s objective is to
minimize “the significant risks to the health and safety of people, as well as
to their fundamental rights, which could be derived from the use of AI
BEUC, Artificial Intelligence What Consumers Say: Findings and Policy
Recommendations of a Multi-Country Survey of AI (Sept. 2020),
https://www.beuc.eu/publications/beuc-x-2020078_artificial_intelligence_what_consumers_say_report.pdf
1579
Government of Spain, El Gobierno constituye el Consejo Asesor de Inteligencia
Artificial (July 20, 2020),
https://www.mineco.gob.es/portal/site/mineco/menuitem.ac30f9268750bd56a0b0240e02
6041a0/?vgnextoid=51884ba89bc63710VgnVCM1000001d04140aRCRD
1578
394
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
systems.”1580 The Spanish government has allotted for its creation in its
proposed budget for 2022.1581
Charter on Digital Rights
In November 2020, the Spanish Government proposed a Charter
on Digital Rights aimed at “recogniz[ing] the challenges posed by the
adaptation of existing rights to the virtual environment, and propos[ing] a
frame of reference for their protection in that context.”1582 The proposal
was published for public comment. The 12-page document was opened for
public comment.1583 Following the public participation process, the final
Digital Rights Charter was unveiled in July 2021.1584
Spanish secretary of State for Digitalization and Artificial
Intelligence, Carme Artigas says the Charter for Digital Rights works as a
“prescriptive document, not regulatory, proposes a framework for the public
authorities' action in a way that allows navigating in the current digital
environment, taking advantage of all its potential and opportunities.”1585
The Charter includes six main categories of rights, covering all areas
of uncertainty and risk: rights of freedom; rights of equality; rights of
participation and shaping the public space; rights of the working and
business environment; digital rights in specific environments; and rights of
guarantees and efficiencies.
Xataka.com, Los algoritmos de las redes sociales serán controlados por una agencia
pública: el Gobierno vigilará cómo se aplica la IA (2021),
https://www.xataka.com/legislacion-y-derechos/algoritmos-redes-sociales-serancontrolados-agencia-publica-gobierno-quiere-vigilar-como-se-aplica-ia
1581
Ibid
1582
Documento para Consulta Pública: Carta de Derechos Digitales (Nov. 18, 2020),
https://portal.mineco.gob.es/RecursosArticulo/mineco/ministerio/participacion_publica/a
udiencia/ficheros/SEDIACartaDerechosDigitales.pdf; Government of Spain, Ministry of
economic affairs and Digital Transformation, The Government promotes the
development of the letter of Digital Rights (Nov. 18, 2020),
https://administracionelectronica.gob.es/pae_Home/pae_Actualidad/pae_Noticias/Anio20
20/Noviembre/Noticia-2020-11-18--Gobierno-impulsa-CartaDerechosDigitales.html?idioma=en
1583
Explica, Government of Spain passes the letter of digital rights to public consultation
(Nov. 19, 2020), https://www.explica.co/government-of-spain-passes-the-letter-ofdigitalrights-to-public-consultation/
1584
Government of Spain (La Moncloa), Sánchez presents the Digital Rights Charter with
which "Spain is at the international forefront in protecting citizens' rights" (2021)
https://www.lamoncloa.gob.es/lang/en/presidente/news/Paginas/2021/20210714_digitalrights.aspx
1585
IT Europa, Spain offers digital rights charter as model (Nov. 18, 2020),
https://www.iteuropa.com/news/spain-offers-digital-rights-charter-model
1580
395
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Algorithmic Transparency
Spain is subject to the General Data Protection Regulation which
established rights to “meaningful information about the logic involved” as
well as about “the significance and the envisaged consequences.”1586 The
Spanish data protection agency (AEPD) has published a detailed guide on
processing with AI.1587 The AEPD report explains that “complying with this
obligation by making a technical reference to the algorithm implementation
may be obscure, confuse or excessive and leading to information fatigue.
However, sufficient information must be provided to understand the
behaviour of the relevant processing.” The AEPD provides many examples
that “must be provided” to understand the behavior of the relevant
processing, such as the relative importance or weight of each data category
in the decision making, the quality of training data and the type of patterns
used, and any reference to audits, “especially on the possible deviation of
inference results, as well as certification or certifications performed on the
AI system.” As indicated above, the Charter of Digital Rights strengthens
principles of Algorithmic Transparency.
The Charter on Digital Rights, described above, guarantees the right
that no citizen is discriminated against for decisions based on algorithms
and maintains that “transparency, auditability, explicability and
traceability” of the same will be ensured. It also adds that people have the
right not to be the subject of a decision based solely on automated decision
processes, thus recognizing the right to “request human supervision and
intervention and challenge automated or algorithmic decisions.” The text
recognizes that citizens must be explicitly informed when they are talking
to an artificial intelligence system and that assistance by a human being
must be guaranteed if the person concerned requests it.”1588
Labor Minister Yolanda Díaz said at a May 2021 press conference,
“Now algorithms are going to be put at the service of the majority in society.
All commercial companies today use artificial intelligence in one way or
another and it is key for us to be courageous and fearlessly govern the
technological transition.”1589
[GDPR Art. 22, Art. 13.2.f]
AEPD, RGPD compliance of processings that embed Artificial Intelligence An
introduction (Feb. 2020), https://www.aepd.es/sites/default/files/2020-02/adecuacionrgpd-ia-en_0.pdf
1588
Government of Spain (La Moncloa), Carta de Derechos Digitales (2021),
https://www.lamoncloa.gob.es/presidente/actividades/Documents/2021/140721Carta_Derechos_Digitales_RedEs.pdf
1589
El Pais, Spain approves landmark law recognizing food-delivery riders as employees:
The legislation also makes it mandatory for digital platforms to be transparent about how
1586
1587
396
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Lethal Autonomous Weapons
Spain has responded to the threat of Lethal Autonomous Weapons
Systems (LAWS) by affirming that the Spanish military does not have and
will not develop such technology and emphasizing the need for
meaningful human control for weapons systems to be compliant with
international humanitarian law. Additionally, Spain has consented to the
11 Principles on LAWS launched by France in 2019.1590 The Spanish
government, however, does not endorse the creation of a preemptive treaty
prohibiting LAWS.1591
OECD/G20 AI Principles
Spain has endorsed the OECD AI Principles and incorporated many
of them into its National AI Strategy and Digital Rights Charter.1592 In June
2020, the OECD reported that Spain is in the “final stages” of developing
its National Strategy on AI, with “a suite of objectives from promotion of
research to the prevention of discrimination and respect for human
rights.”1593 Spain published its national AI Strategy in December 2020.1594
Human Rights
Spain ranks highly for political rights and civil liberties, through
there was a slight dip in the 2021 score (90/100) as compared with 2020
(92/100). According to Freedom House, “Spain’s parliamentary system
features competitive multiparty elections and peaceful transfers of power
their algorithms affect working conditions (May 12, 2021),
https://english.elpais.com/economy_and_business/2021-05-12/spain-approves-landmarklaw-recognizing-food-delivery-riders-as-employees.html
1590
French Ministry for Europe and Foreign Affairs, 11 Principles on Lethal Autonomous
Weapons Systems (LAWS) (September 2019), https://www.diplomatie.gouv.fr/en/frenchforeign-policy/united-nations/multilateralism-a-principle-of-action-for-france/alliancefor-multilateralism/article/11-principles-on-lethal-autonomous-weapons-systems-laws
1591
Human Rights Watch, Stopping Killer Robots- Country Positions on Banning Fully
Autonomous Weapons Systems and Retaining Human Control (Aug. 20, 2020),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and#
1592
OECD.AI Policy Obvservatory – Spain (2021),
https://oecd.ai/en/dashboards/countries/Spain
1593
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1594
https://www.lamoncloa.gob.es/lang/en/presidente/news/Paginas/2020/20201202_enia.asp
x
397
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
between rival parties. The rule of law prevails, and civil liberties are
generally respected.”1595
Evaluation
Spain has developed a strong ethics-focused national AI strategy,
created an independent AI advisory council, enacted a Charter for Digital
Rights, and strongly endorsed the EU’s policies and statements on ethical
AI as well as the OECD AI Principles. The Map of AI Technology
Capabilities is the most authoritative source on AI developments in Spain,
but it only includes technological capabilities without informing on
adherence to ethical principles. On issues of data privacy, Spain has
expanded upon the EU’s General Data Protection Regulation (GDPR) and
issued a more robust human rights centered law.1596 This precedent shows
that Spain is capable of undertaking initiative beyond what is required by
the EU. Spain has not explicitly endorsed the Universal Guidelines for AI,
but the national AI strategy reflects elements of the UGAI. The decision in
2021 to establish a legal right for algorithmic transparency is an AI policy
milestone.
Freedom House, Freedom in the World 2021 – Spain (2021),
https://freedomhouse.org/country/spain/freedom-world/2021; Freedom House, Freedom
in the World 2020 – Spain (2020), https://freedomhouse.org/country/spain/freedomworld/2020
1596
Gobierno de España- Boletín Oficial del Estado, Ley Orgánica 3/2018, de 5 de
diciembre, de Protección de Datos Personales y garantía de los derechos digitales
(2018), https://www.boe.es/eli/es/lo/2018/12/05/3/con
1595
398
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Sweden
National AI Strategy
In May 2018, Swedish announced the National Approach to
Artificial Intelligence. The national approach to AI reflects the
government’s goal “to make Sweden a leader in harnessing the
opportunities that the use of AI can offer, with the aim of strengthening
Sweden’s welfare and competitiveness.”1597
Sweden emphasized that a “cross-cutting theme should be
sustainable AI, meaning that AI applications should be ethical, safe, secure,
reliable and transparent” more specifically regarding “the use of AI
algorithms.” When referring to the potential threats, the Government refers
to “challenges related to rule of law procedures and the automation of
agency decisions,” “the risks to both society and individuals,” “fundamental
needs for privacy”, “discrimination, loss of trust,” and the consequences for
the functioning of democracy.
Accountability is not mentioned. The emphasis is on responsible
design and use of AI. For example, “it is important that AI systems are
carefully designed to prevent them from doing harm. It is therefore
important that companies and public institutions collaborate with relevant
academics, for example through joint projects or staff exchanges.”
The Swedish National Approach to AI should also be read against
the background of a 2020 joint response to the European Commission’s
White Paper on AI.1598 In this “non-paper,” Sweden and 13 other EU
member states, describe human-centric and trustworthy AI “as a
competitive advantage.” According to the non-paper, Sweden supports the
use of hard law tools for “creating a genuinely single market for AI.”
Sweden also favors the use of “soft law solutions such as self-regulation,
voluntary labelling and other voluntary practices as well as robust
standardisation process.” As for risks to individuals or to society stemming
from the use of AI, Sweden advocates for an evidence-based and “wellcalibrated and proportionate approach.”
Government Offices of Sweden, National Approach to Artificial Intelligence, (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf
1598
Non-paper - Innovative and trustworthy AI: two sides of the same coin (Aug. 10,
2020),
https://www.permanentrepresentations.nl/binaries/nlatio/documents/publications/2020/10
/8/non-paper---innovative-and-trustworthy-ai/Non-paper++Innovative+and+trustworthy+AI+-+Two+side+of+the+same+coin.pdf.
1597
399
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The original National Approach to AI states “Sweden must create
the enabling conditions.” The Government defines the “key conditions for
realising the potential of AI as 1) Education and training, (2) Research, (3)
Innovation, and (4) Framework and infrastructure. Three Ministries –
Ministry of Infrastructure; Ministry of Enterprise and Innovation; and
Ministry of Education and Research – are responsible for AI policies and
independent agencies under these Ministries implement these policies. In
relation to education and training, “the Government’s assessment is that
• Swedish higher education institutions need to provide a sufficient
number of people with AI education and training, particularly in
continuing and further education for professionals with a university
degree or equivalent.
• Sweden needs a strong AI component in non-technical programmes
to create the conditions for broad and responsible application of the
technology
• Sweden needs a strong link between research, higher education and
innovation in the field of AI.”1599
In the “Sweden AI Strategy Report,”1600 the European Commission
noted that “Swedish universities have started proposing bachelor’s and
master’s programmes in AI fields,” with some courses tackling ethical
aspects of AI.1601 The same goes for continuing and further education for
professionals and citizens “rolled out in an effective policy in the form of a
course on the Elements of AI”, financed by Vinnova, Sweden’s innovation
agency.
In relation to research, the Government’s assessment1602 is that
Government Offices of Sweden, National Approach to Artificial Intelligence, (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf6.
1600
European Commission, Sweden AI Strategy Report, (Feb. 2020),
https://knowledge4policy.ec.europa.eu/ai-watch/sweden-ai-strategy-report_en
1601
See Uppsala University, Master's Programme in Data Science, Syllabus for Data,
Ethics and Law, Syllabus for artificial intelligence; Stockholm University, Master’s
Programme in Artificial Intelligence, Course on Open and big data management tackling
Ethical aspects of big data and open data; Chalmers University, Data Science and AI,
MSc, the Wallenberg Artificial Intelligence, Autonomous Systems and Software Program
(WASP) Graduate School offers an AI-track mandatory course tackling ethical aspects.
1602
Government Offices of Sweden, National Approach to Artificial Intelligence, (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf6.
1599
400
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Sweden needs both strong basic research and strong applied
research in AI to ensure knowledge and skills supply in the field.
• Sweden needs strong relations with leading international AI
research environments.
• Sweden needs to exploit the synergies between civil research and
defense research from a total defense perspective.
In relation to innovation and use, the Government’s assessment is that
• Sweden needs pilot projects, testbeds and environments for
development of AI applications in the public and private sectors,
that can contribute to the use of AI evolving in a safe, secure and
responsible manner.
• Sweden needs to develop partnerships and collaborations on the use
of AI applications with other countries, especially within the EU.
Apart from EU funding programs, Vinnova, Sweden’s innovation
agency is one of the main sources of funding which has fostered the
development of AI applications through AI Innovation of Sweden,1603
organized as a national center for applied AI research and innovation with
almost 70 partners from the industrial and public sectors, research
institutions, and the academic world. It also funds AI-related innovation
projects promoting citizen science which “is considered an important
instrument for maintaining confidence in science and society.”1604 In April
2021, a total of 256 ongoing projects matched the keywords “artificial
intelligence” and “AI” in Vinnova’s project database.1605 It also funds the
development of innovative services that help citizens and journalists to
review the public sector1606 and SMEs and public organisations when they
start the first innovation project in AI.
In relation to framework and infrastructure, the Government’s
assessment is that
• Sweden needs to develop rules, standards, norms and ethical
principles to guide ethical and sustainable AI and the use of AI.
•
AI Innovation of Sweden, https://www.ai.se/en. Nodes and co-location areas are
created or planned across Sweden to develop partnerships for AI innovation: the
Gothenburg node, the Greater Stockholm node, the Southern Sweden node, the Northern
Sweden node, the Örebro node.
1604
Vinnova, https://www.vinnova.se/en/. See also Vinnova, Civic tech: Digital services
for strengthening trust between citizens and the public sector.
1605
Sweden AI Strategy Report https://knowledge4policy.ec.europa.eu/ai-watch/swedenai-strategy-report_en
1606
“This is done through interactive vizualisation of climate data at the local, level,
increased transparency in procurement data and simulation of how the citizen’s own data
can be used in a secure way”, Vinnova, News, (Oct. 26, 2020),
https://www.vinnova.se/en/.
1603
401
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Sweden needs to push for Swedish and international standards and
regulations that promote the use of AI and prevent risks.
• Sweden needs to continuously review the need for digital
infrastructure to harness the opportunities that AI can provide.
• Sweden needs to continue to work on making data available to serve
as infrastructure for AI use in areas where it adds value.
• Sweden needs to continue to play an active role in the EU’s efforts
to promote digitization and reap the benefits that the use of AI can
bring.1607
The National Approach to Artificial Intelligence states, “The goal is
closely linked to the digital transformation goal adopted by the Riksdag [the
Swedish Parliament] and complements the Government’s Digital Strategy.”
In August 2018, the Swedish government created a Committee for
Technological Innovation and Ethics (KOMET). AlgorithmWatch noted
that the head of the Committee is “a former entrepreneur and CEO, whose
prior work for the government included being head of the section for
innovation within the Ministry of Enterprise and Innovation.”1608
AlgorithmWatch also complained that: “The committee is tasked with
producing analyses of barriers for the adoption of ‘the fourth industrial
revolution’, such as existing regulatory frameworks, to map the need for
adjusting existing regulatory frameworks, to continuously come up with
suggestions for the government regarding policy developments, promote a
dialogue between relevant governmental agencies and regional actors,
educational institutions, the non-governmental sector, and business for
efficient collaboration concerning policy-developments. It is, however, not
stated how, and more precisely which of these actors will be involved.”
The Swedish Government acknowledged the General Data
Protection Regulation (GDPR) as “an important part of the AI
framework.”1609
•
Government Offices of Sweden, National Approach to Artificial Intelligence, (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf6.
1608
Algorithmwatch, Sweden: New governmental committee for coordinated and
accelerated development of policies related to the fourth industrial revolution (20192020), https://algorithmwatch.org/en/automating-society/sweden/.
1609
Government Offices of Sweden, National Approach to Artificial Intelligence (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf6.
1607
402
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In September 2020, the International Bar Association released a
report which stated, “there is currently no AI laws in Sweden. Historically,
the legislative approach in Sweden has been to pass technology-agnostic
legislation that does not need to be changed with every advance in
technology.”1610 Thus, “it is of central priority for the Swedish legislator to
assess current legislation from an AI perspective and implement necessary
changes. Furthermore, support in the interpretation of legislation is required
from courts and public authorities. Access to data, information security and
robustness, together with the ethical use of AI, are principles of central
importance in the future regulatory approach.”
In October 2021, a national summit on AI and democracy, entitled
“the Future of Democracy Summit” was initiated to focus on the question
of how digitalisation and AI can be used in a democratically sustainable
way.1611
The Trellborg Controversy
The automation of government services has been underway in
Sweden since the 1970s. By 2019, “more than 80% of all government
decisions that the National Audit has reviewed were automated. This
involves 121 million decisions by 13 authorities.”1612 Various benefits, such
as Welfare payments, from parental benefits to dental care subsidies, are
allocated without any human intervention.
As for municipalities, who are in charge of social services, a 2019
report published by the Union for Professionals, found that “only 16 out of
a total of 290 municipalities have implemented RPA [Robotic Processing
Automation] in their administration of social benefits.”1613 The Trellborg
Municipality was the only one to implement solely automated decisionmaking.
In 2019 the journalist Freddi Ramel, and Simon Vinge, chief
economist at the Union for Professionals, challenged the Trelleborg
automated decision system. According to AlgorithmWatch, the Swedish
International Bar Association, Guidelines and Regulations to Provide Insights on
Public Policies to Ensure Artificial Intelligence’s Beneficial Use as a Professional Tool
(Sept. 2020), https://www.ibanet.org/Document/Default.aspx?DocumentUid=f5099a331e70-4a32-839d-589236b7568d.
1611
Future of Democracy, Sustainable citizenship in a digital age – Future of Democracy
Summit https://www.futureofdemocracy.se/summit
1612
Nord News, The Swedish National Audit Office: Automatic government decisions are
becoming more common (Nov. 19, 2020), https://nord.news/2020/11/19/the-swedishnational-audit-office-automatic-government-decisions-are-becoming-more-common/
1613
Lupita Svensson, ”Tekniken är den enkla biten” Om att implementera digital
automatisering i handläggningen av försörjningsstöd (2019),
https://akademssr.se/sites/default/files/files/LupitaSvensson.pdf
1610
403
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Parliamentary Ombudsman has so far failed to determine whether the
municipality provided “meaningful information” as required by Article 15
of the GDPR.1614 However, Ramel obtained access to the source code after
a court ruled that the code was a public record under the Swedish Freedom
of Information Act. The Trelleborg municipality subsequently undertook an
investigation.
Independent Oversight
The Swedish Government acknowledged the General Data
Protection Regulation (GDPR) as “an important part of the AI
framework.”1615 Sweden’s Authority for Privacy Protection’s (IMY) “role
is to uphold the protection of personal data, monitoring that they are handled
correctly and do not fall into the wrong hands.”1616 It remains active. In
2019, it issued its first fine in a case involving facial recognition. The IMY
concluded that a school that conducted a pilot using facial recognition to
keep track of students' attendance in school violated the GDPR and imposed
a fine on the municipality of approximately 20,000 euros.1617 The IMY has
also held the police accountable for its unlawful use of facial recognition
technology, which is detailed further in the facial recognition section
below.1618
In May 2019, the Ministry of Infrastructure launched three
assignments to strengthen the country’s open access efforts. 1619 It assigned
Sweden’s Lantmäteriet to analyze the consequences of free access to
valuable amounts of data, a special investigator to analyze “the need for
constitutional amendments and ensure appropriate national regulation”, and
AlgorithmWatch, Central authorities slow to react as Sweden’s cities embrace
automation of welfare management (2020),
https://automatingsociety.algorithmwatch.org/report2020/sweden/sweden-story/
1615
Government Offices of Sweden, National Approach to Artificial Intelligence (May
2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb49433587574c513a837fac/
national-approach-to-artificial-intelligence.pdf6.
1616
Swedish Authority for Privacy Protection, Welcome to IMY, https://www.imy.se/en/
1617
Swedish Authority for Privacy Protection, Facial recognition in school renders
Sweden's first GDPR fine, 21 August 2019, https://www.imy.se/en/aboutus/arkiv/nyhetsarkiv/facial-recognition-in-school-renders-swedens-first-gdpr-fine/
1618
EDPB, Swedish DPA: Police unlawfully used facial recognition app (Feb. 11, 2021),
https://edpb.europa.eu/news/national-news/2021/swedish-dpa-police-unlawfully-usedfacial-recognition-app_en
1619
Swedish Government, The government is gathering strength around artificial
intelligence and open data (May 2, 2019),
www.regeringen.se/pressmeddelanden/2019/05/regeringen-kraftsamlar-kring-artificiellintelligens-och-oppna-data.
1614
404
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Sweden’s Agency for Digital Administration (DIGG) to “increase
the public administration's ability to make open data available and to
conduct open and data-driven innovation.”1620 This has resulted in the
launch of Sweden’s data portal with new functionality for APIs, and the
establishment of principles, guidelines, and recommendations “in order to
increase the public administration's ability to make open data available.”1621
Much of this work was based on DIGG’s piloted projects with business,
academia, and civil society at challengesgov.se.
In the January 2020 report, Sweden’s Agency for Digital
Administration (DIGG) recommended that the Government establish a
center with expertise in AI; develop a platform for collaboration, codevelopment and innovation; develop an AI guide; create legal conditions
to facilitate experimental activities; develop vocational and role-specific
training in AI; and develop a national data strategy for public
administrations.1622 DIGG also set up an expert group on AI for public
administration, mainly composed of academics, which aims to provide
advises to DIGG in the fulfilment of its mission.1623
AI Sweden also established a Legal Expert Group, which consists of
legal experts from its partners. The Group discusses legal questions related
to AI and data and they “collaborate in trying to create, for example, white
papers, guidelines and/or common interpretations and solutions for legal
issues that could benefit all partners within AI Sweden.”1624 In 2021, the
Legal Expert Group met numerous times and discussed matters such as the
practical use of applied AI, the concept of Federated Learning, questions
regarding anonymization and pseudonymization, the EU Proposal for an AI
Act, intellectual property rights, and standardized agreements for sharing
data. AI Sweden and the Legal Expert Group also initiated a collaboration
with the Swedish Authority for Privacy Protection (IMY). Future
Swedish Government, The government is gathering strength around artificial
intelligence and open data (May 2, 2019),
www.regeringen.se/pressmeddelanden/2019/05/regeringen-kraftsamlar-kring-artificiellintelligens-och-oppna-data.
1621
DIGG, Öppna data, datadriven innovation och AI (Open data, data-driven innovation
and AI), 29 Jan. 2021, https://www.digg.se/publicerat/publikationer/2021/oppna-datadatadriven-innovation-och-ai
1622
DIGG, Promote the ability of public administration to use AI, 13 Jan. 2020,
https://www.digg.se/publicerat/publikationer/2020/framja-den-offentliga-forvaltningensformaga-att-anvanda-ai
1623
DIGG, Referensgrupp inom AI, https://www.digg.se/om-oss/regeringsuppdrag/oppnadata-datadriven-innovation-och-ai#referensgrupp_inom_ai, [MT].
1624
AI Sweden, Legal Expert Group, https://www.ai.se/en/legal-expert-group
1620
405
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
discussions will concentrate on using AI in a legal context, such as using
natural language processing (NLP) as a tool in legal work. 1625
Lastly, the independent Equality Ombudsman (DO) plays a role in
ensuring the absence of discrimination which could result from a biased
algorithm.1626
Foreign Policy and AI
As for the international landscape, Peter Eriksson, the Swedish
Minister for Housing and Digital development, signed the declaration on
“AI in the Nordic-Baltic region” establishing a collaborative framework on
“developing ethical and transparent guidelines, standards, principles and
values to guide when and how AI applications should be used” and “on the
objective that infrastructure, hardware, software and data, all of which are
central to the use of AI, are based on standards, enabling interoperability,
privacy, security, trust, good usability, and portability.”1627 This Declaration
has recently been reinforced by the Ministerial Declaration Digital North
2.0.1628 In November 2021, the Ministers released another joint statement
announcing a focus on digital inclusion, striving to implement measures to
make digital services more accessible to all Swedish inhabitants and
ensuring that those who do not possess the necessary level of skills get the
opportunity to acquire them. 1629
Public Participation
As for public participation, the Government states in the National
Approach that “For Sweden to reap the benefits of AI, all sectors of society
must be involved.” It is, however, debatable to what extent different groups
in society are actually involved. According to AlgorithmWatch, most of the
funding and strategic development takes place in the universities and as
AI Sweden, Update from AI Sweden's Legal Expert Group (Jan. 17, 2022,
https://www.ai.se/en/news/update-ai-swedens-legal-expert-group
1626
Diskriminerings Ombudsmannen, Welcome to the Equality Ombudsman (Oct. 20,
2020), https://www.do.se/other-languages/english/. See also Emma Lundberg, Automated
decision-making vs indirect discrimination – Solution or aggravation (2019),
https://www.diva-portal.org/smash/get/diva2:1331907/FULLTEXT01.pdf.
1627
Nordic Cooperation, AI in the Nordic-Baltic region (May 14, 2018),
https://www.norden.org/en/declaration/ai-nordic-baltic-region.
1628
Nordic Cooperation, Ministerial Declaration Digital North 2.0 (Sept. 29, 2020),
https://www.norden.org/en/declaration/ministerial-declaration-digital-north-20.
1629
Nordic Cooperation, Common statement on the importance of promoting digital
inclusion as a central part of the digital transformation in the Nordic-Baltic region, 26
Nov. 2021, https://www.norden.org/en/declaration/common-statement-importancepromoting-digital-inclusion-central-part-digital.
1625
406
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
support for business environments.”1630 Nevertheless, AlgorithmWatch also
describes the “addAI initiative” which “is a collaboration between experts
in academia, government and companies to discuss and explore the impact
of smart algorithms and AI on society through the organisation of
workshops and participation in public events.”
In October 2021, the “Future of Democracy Summit” hosted
stakeholders from business, civil society, academia, and government to
discuss sustainable AI and democracy.1631
Facial recognition
In March 2020, the data protection officer for the Swedish police
undertook an investigation to determine whether the police may have used
ClearView AI, an AI product for mass surveillance enabled by facial
recognition.1632 The Swedish police confirmed that they have used
Clearview AI, after previously denying use of the face surveillance tool.1633
Subsequently, the Swedish DPA “initiated an inspection to find out whether
Swedish authorities use the face recognition technology provided by the US
company Clearview AI.”1634 The DPA noted that the European Data
Protection Board “will produce guidance on how law enforcement
authorities should approach facial recognition technology. Sweden is one of
the driving countries in the world.” In 2021, the Swedish DPA found that
the Swedish Police Authority had processed personal data in breach of the
Swedish Criminal Data Act when using Clearview AI to identify
individuals. The investigation concluded that Cleaview 1635AI had been
used by the Police on several occasions and sometimes without any prior
authorisation. The Police had failed to implement sufficient organisational
Anne Kuan and Julia Velkovia, Automating Society: Sweden (Jan. 29, 2019),
https://algorithmwatch.org/en/automating-society-sweden/.
1631
Future of Democracy, Sustainable citizenship in a digital age – Future of Democracy
Summit https://www.futureofdemocracy.se/summit
1632
SVT NYHETER, Intern utredning: Polisen får inte använda kritiserad AI-tjänst –
skulle bryta mot lagen (Mar. 6, 2020) [MT], https://www.svt.se/nyheter/inrikes/svenskpolis-forbjuds-att-anvanda-kontroversiella-ai-tjansten
1633
Mikael Grill Peterson and Linea Carlén, Polisen bekräftar: Har använt
omdiskuterade Clearview AI, SVT NYHETER, (March 11, 2020) [MT],
https://www.svt.se/nyheter/inrikes/ekot-polisen-bekraftar-anvandning-av-kontroversiellapp
1634
The Swedish Data Protection Authority, The Data Inspectorate initiates supervision
due to Clearview AI, https://www.datainspektionen.se/nyheter/datainspektionen-inledertillsyn-med-anledning-av-clearview-ai/.
1635
EDPB, Swedish DPA: Police unlawfully used facial recognition app (Feb. 11, 2021),
https://edpb.europa.eu/news/national-news/2021/swedish-dpa-police-unlawfully-usedfacial-recognition-app_en
1630
407
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
measures to ensure and be able to demonstrate that the processing of
personal data in this case has been carried out in compliance with the
Criminal Data Act. No impact assessment had been concluded. As a result,
a 250,000 Euros fine was imposed.
In 2019, the Swedish DPA did approve the use of facial recognition
technology by the police to help identify criminal suspects. However, such
an authorization relates to the use of biometric templates in databases under
the control of public authorities and established under Union or Member
States law.1636 As Andrea Jelinek, Chair of the European Data Protection
Body, subsequently emphasized, “[t]he possible use of a service such as
offered by Clearview AI by law enforcement authorities would, however,
be fundamentally different, in that it would imply, as part of a police or
criminal investigation, the sharing of personal data with a private party
outside the Union and the biometric matching of such data against the
latter’s mass and arbitrarily populated database of photographs and facial
pictures accessible online.”1637 She also clearly questions the legality of the
use of Clearview AI by public authorities.
In 2019, the Swedish DPA issued its first fine in a case involving
facial recognition. A school in northern Sweden conducted a pilot using
facial recognition to keep track of students' attendance in school. The
Swedish DPA concluded that the test violates the GDPR and imposed a fine
on the municipality of approximately 20,000 euros. The school processed
sensitive biometric data unlawfully and failed to do an adequate impact
assessment including seeking prior consultation with the Swedish DPA. The
school based the processing on consent but the Swedish DPA considers that
consent was not a valid legal basis given the clear imbalance between the
data subject and the controller.1638
In 2021, the Swedish DPA concluded that the Swedish Police
Authority breached the Swedish Criminal Data Act with its use of
Clearview AI to identify individuals, having used it on several occasions
Official Journal of the European Union, Directive EU 2016/680 of the European
Parliament and of the Council (Apr. 27, 2016), https://eur-lex.europa.eu/legalcontent/EN/TXT/PDF/?uri=CELEX:32016L0680&from=EN
1637
EDPB response to MEPs Sophie in ‘t Veld, Moritz Körner, Michal Šimečka, Fabiene
Keller, Jan-Christoph Oetjen, Anna Donáth, Maite Pagazaurtundúa, Olivier Chastel,
concerning the facial recognition app developed by Clearview AI (June 10, 2020)
https://edpb.europa.eu/sites/edpb/files/files/file1/edpb_letter_out_20200052_facialrecognition.pdf.
1638
https://www.datainspektionen.se/nyheter/2019/facial-recognition-in-school-rendersswedens-first-gdpr-fine/.
1636
408
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
without any prior authorisation. 1639 The Police had failed to implement
sufficient organizational measures to ensure that the processing of personal
data in this case has been carried out in compliance with the Criminal Data
Act. IMY imposed a 250,000 Euros fine.
Lethal Autonomous Weapons
Beginning in 2013, Swedish NGOs called for Sweden to endorse
an official ban of LAWS.1640 As one NGO coalition stated recently “A future
where machines themselves decide over life and death, what and who is to
be attacked in an armed conflict, is not the future we want. But the fact is
that we are on our way there - and development is fast.”1641 Previously,
leaders in the Swedish government declared that “Sweden must take a
leading role in the work for a ban on deadly autonomous weapon systems.”
However, the position adopted by Sweden so far seems to be more nuanced.
The Swedish government has emphasized human control and said,
“that multilateralism remains our only chance to address our many common
challenges and to ensure international peace and security.”1642 At the 75th
UN General Assembly meeting in October 2020, Sweden’s Ambassador
stated “Sweden is of the strong conviction that human control over the use
of force always must be upheld.” She also expressed Sweden’s support to
the 11 LAWS Guiding Principles.1643 Earlier, Sweden also explained that
the “specific measures required for human control will thus need to be
context dependent and assessed on a case-by-case basis” and referred to a
report, supported by Sweden, together with Germany, Switzerland and the
Netherlands.1644 The Swedish government has also set up a working group
EDPB, Swedish DPA: Police unlawfully used facial recognition app (Feb. 11, 2021),
https://edpb.europa.eu/news/national-news/2021/swedish-dpa-police-unlawfully-usedfacial-recognition-app_en
1640
Campaign to Stop Killer Robots, Building Awareness in Sweden (Oct. 13, 2020),
https://www.stopkillerrobots.org/2013/10/sweden-outreach/;
1641
Amnesty International Sweden, Sweden Must Stand Against Killer Robots (Sept. 21,
2020) [GT], https://www.amnesty.se/aktuellt/sverige-maste-sta-upp-mot-mordarrobotardebattartikel-publicerad-209-2020/.
1642
Government of Sweden, Ministry of Foreign Affairs, Statement delivered by
Ambassador Anna Karin Eneström, Permanent Representative of Sweden to the United
Nations at the General Debate of the First Committee, 75th session of the UN General
Assembly, United Nations (Oct. 14, 2020),
https://reachingcriticalwill.org/images/documents/Disarmamentfora/1com/1com20/statements/14Oct_Sweden.pdf
1643
Geneva Internet Platform, GGE on lethal autonomous weapons systems,
https://dig.watch/process/gge-laws#view-14508-1
1644
Stockholm International Peace Research Institute, Limits on Autonomy in Weapon
Systems: Identifying Practical Elements of Human Control (June 2020),
1639
409
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
on autonomous weapons. According to the NGO the Swedish Peace and
Arbitration Society, this “working group is defense-oriented, with a
majority of its members coming from defense-related authorities and
institutions.”1645
OECD AI Principles
Sweden endorsed the OECD AI Principles. In 2021, the OECD
noted that Sweden published a document outlining its national approach to
AI in 2019. “The purpose of this document was to identify an overall
direction for AI-related work in Sweden and lay the foundation for future
priorities.” Sweden also described an AI governance structure with “three
Ministries – Ministry of Infrastructure; Ministry of Enterprise and
Innovation; and Ministry of Education and Research – are responsible for
AI policies and independent agencies under these Ministries implement
these policies.”1646
Algorithmic Transparency
Sweden is subject to the General Data Protection Regulation which
established rights to “meaningful information about the logic involved” as
well as about “the significance and the envisaged consequences.”1647 The
Swedish Data Protection Authority is competent to handle complaints in
this regard.1648 In 2019, the Equality Ombudsman Agneta Broberg warned
that the sanctions available under the Discrimination Act are not effective
to tackle the challenges of AI and discriminatory algorithms.1649
Following the Trelleborg episode concerning automated decisions
by municipalities, the Union for Professionals called for the creation of an
https://www.sipri.org/publications/2020/other-publications/limits-autonomy-weaponsystems-identifying-practical-elements-human-control-0
1645
Swedish Peace and Arbitration Society, Questions and Answers about Killer Robots,
[MT] https://www.svenskafreds.se/vad-vi-gor/nedrustning/stoppamordarrobotar/faqmordarrobotar/
1646
OECD, State of Implementation of the OECD AI Principles: Insights from National
AI Policies (June 2021), https://www.oecd.org/digital/state-of-implementation-of-theoecd-ai-principles-1cd40c44-en.htm
1647
[GDPR Art. 22, Art. 13.2.f].
1648
The Swedish Data Protection Authority, https://www.datainspektionen.se/otherlang/in-english/.
1649
Diskriminerings Ombudsmannen, Skyddet mot diskriminering behöver ses över
(Protection Against Discrimination Needs to be Reviewed), (Feb. 21, 2020),
https://www.do.se/om-do/pressrum/aktuellt/aktuellt-under-2020/skyddet-motdiskriminering-behover-ses-over/; Diskriminerings Ombudsmannen, Annual Report
2019, https://www.do.se/globalassets/om-do/do-arsredovisning-2019.pdf.
410
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
algorithm ombudsman. 1650 “The Union has, among other things, conducted
a survey that shows that the requirement for an Algorithm Ombudsman
has broad support among the public - and that transparency and openness
are absolutely crucial for there to be trust in algorithms and automation.” In
this regard, a 2020 survey organized by BEUC, the European Consumer
Organization, which involved the Swedish Consumer Organization,
revealed that “more than half of Sweden's consumers feel that artificial
intelligence (AI) is used to manipulate them. And over 60 percent wish they
could say no to automated decision making.”1651
Lastly, another independent government agency, the Equality
Ombudsman (DO), may also play a part in ensuring the absence of
discrimination which could result from a biased algorithm.1652 The case of
Freddi Ramel v. the Trelleborg municipality previously mentioned also
makes clear that the principle of public access does cover the source code
of the software used for automated decisions and can be vindicated before
the relevant administrative court.
Human Rights
Sweden is a signatory to many international human rights treaties
and conventions, among which the Universal Declaration of Human Rights
and the Council of Europe’s European Convention on Human Rights and
the Convention 108+, the Modernized Convention for the protection of
individuals with regard to the processing of personal data. Sweden typically
ranks among the top nations in the world for the protection of human rights
and transparency.1653
According to Freedom House, “Sweden is a parliamentary
monarchy with free and fair elections and a strong multiparty system. Civil
liberties and political rights are legally guaranteed and respected in practice,
and the rule of law prevails.”1654 It consequently ranks among the top
Union for Professionals, Algorithm Policy in a Digital World,
https://akademssr.se/opinion/algoritmpolitik.
1651
https://www.sverigeskonsumenter.se/nyheter-press/nyheter-ochpressmeddelanden/las-mer-om-undersokningen-har/ [MT]; BEUC, Artificial intelligence:
what consumers say, https://www.sverigeskonsumenter.se/media/kbgf3wya/beuc-ai.pdf.
1652
Diskriminerings Ombudsmannen, Welcome to the Equality Ombudsman (Oct. 20,
2020), https://www.do.se/other-languages/english/. See also Emma Lundberg, Automated
decision-making vs indirect discrimination – Solution or aggravation (2019),
https://www.diva-portal.org/smash/get/diva2:1331907/FULLTEXT01.pdf.
1653
According to Freedom House, Sweden’s Global freedom score is 100/100,
https://freedomhouse.org/country/sweden/freedom-world/2020.
1654
Freedom House, Freedom in the World 2021– Sweden (2021),
https://freedomhouse.org/country/sweden/freedom-world/2021
1650
411
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
nations for political rights and civil liberties, earning a perfect 100/100
score.
Evaluation
Sweden endorsed the OECD AI Principles and is committed to
developing trustworthy AI. Sweden ranks at the top among nations for the
protection of political rights and civil liberties and has proactive
ombudsman institutions and an active data protection agency. However, the
lack of a clear strategy to involve citizens in the debate over the future of
AI, Sweden’s opposition, along with other Nordic countries, to a strong
regulatory framework for AI raises concern about Sweden’s ability to both
support technical innovation and avoid ethical risks. There are also concern
about the use of the controversial ClearviewAI facial recognition
technology in Sweden.
412
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Switzerland
National AI Strategy
The Swiss government recently announced AI Guidelines for the
Federal Government.1655 The AI Guidelines are intended to ensure a
coherent government policy for AI. The AI Guidelines emphasize putting
people at the center; Framework conditions for the development and use of
AI, Transparency, traceability and Explainability; Accountability; Safety;
Active participation in shaping the governance of AI; and Involvement of
all affected national and international actors. Specific AI guidelines will be
formulated for education and science. Future AI work will be undertaken
by the Federal Office of Communications OFCOM together with the federal
agencies concerned.1656
The AI Guidelines follow from the Digital Switzerland Strategies.
These strategies encompass the Swiss governments principles and key
objectives for the digital transformation across all sectors.165716581659 Although
these reports do not have the sole focus of AI, the federal government has
taken further action to focus on AI. Following the Digital Switzerland
Strategy 2018, the federal government identified several areas for further
enquiry:
• International law and the use of AI in public opinion and
decision making
• How the use of AI in the federal administration can be
improved
Der Bundesrat, Leitlinien, “Künstliche Intelligenz” für den Bund (Nov. 2020) [DT],
https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2020/11/leitlinie_ki.pdf.download.pdf
/Leitlinien%20Künstliche%20Intelligenz%20-%20DE.pdf
1656
Der Bundesrat, Leitlinien “Künstliche Intelligenz” für die Bundesverwaltung
verabschiedet (Nov. 25, 2020) [DT],
https://www.admin.ch/gov/de/start/dokumentation/medienmitteilungen.msg-id81319.html)
1657
Schweizerischer Eidgenossenschaft, Digital Switzerland» Strategy, September 2020,
https://www.bakom.admin.ch/dam/bakom/en/dokumente/informationsgesellschaft/strateg
ie/strategie_digitale_schweiz.pdf.download.pdf/Strategie-DS-2020-EN.pdf
1658
Schweizerischer Eidgenossenschaft, Digital Switzerland» Strategy, September 2018,
https://www.bakom.admin.ch/dam/bakom/en/dokumente/informationsgesellschaft/strateg
ie/Strategie_DS_Digital_2-EN-barrierenfrei.pdf.download.pdf/Strategie_DS_Digital_2EN-barrierenfrei.pdf
1659
Schweizerischer Eidgenossenschaft, Digital Switzerland» Strategy, April 2016,
https://www.bakom.admin.ch/dam/bakom/en/dokumente/bakom/digitale_schweiz_und_i
nternet/Strategie%20Digitale%20Schweiz/Strategie/Strategie%20Digitale%20Schweiz.p
df.download.pdf/digital_switzerland_strategy_Brochure.pdf
1655
413
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Department of Education, Research and Innovation (SBFI)
was given the task of preparing stakeholders for the digital
transformation through their policy work.
• An Interdepartmental Working Group on Artificial Intelligence
was established to pursue strategic objectives for the federal
government.1660
These areas of further enquiry resulted in three reports prepared by the
federal government with the following focuses:
• Artificial Intelligence in Cyber Security and Security Policy1661
• International Committees and Artificial Intelligence1662
• Artificial Intelligence, the Media and the Public1663
The report on “Artificial Intelligence in Cyber Security and Security
Policy” gives an overview of how AI is influencing national security and
how the military and government are dealing with this. It further lists
considerations that need to be made in this regard. This includes how
fundamental and human rights are affected, how legal and ethical
considerations can be integrated and what new regulatory measures need to
be implemented. 1664
The report on International Committees and Artificial Intelligence
also gives an overview of different international organizations and their
efforts in AI policy. It further goes on to give recommendations of concrete
action in Swiss foreign policy. The report mentions the importance of the
•
Schweizerischer Eidgenossenschaft: Staatssekretariat für Bildung, Forschung und
Innovation SBFI, Künstliche Intelligenz, https://www.sbfi.admin.ch/sbfi/de/home/bfipolitik/bfi-2021-2024/transversale-themen/digitalisierung-bfi/kuenstliche-intelligenz.html
1661
Schweizerischer Eidgenossenschaft: Eidgenössisches Department für Verteidigung,
Bevölkerungsschutz und Sport, Künstliche Intelligenz in der Cybersicherheit und
Sicherheitspolitik, https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/k-i_cs.pdf.download.pdf/k-i_c-s_d.pdf
1662
Schweizerischer Eidgenossenschaft, International Gremien und Künstliche
Intelligenz, August 2019, https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/ig_k-i.pdf.download.pdf/i-g_ki_d.pdf
1663
Schweizerischer Eidgenossenschaft: Eidgenössisches Department für Umwelt,
Verkehr, Energie und Kommunikation UVEK, Künstliche Intelligenz, Medien und
Öffentlichkeit, August 2019,
https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/i-g_k-i.pdf.download.pdf/ig_ki_d.pdf
1664
Schweizerischer Eidgenossenschaft: Eidgenössisches Department für Verteidigung,
Bevölkerungsschutz und Sport, Künstliche Intelligenz in der Cybersicherheit und
Sicherheitspolitik, 2019. https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/ki_c-s.pdf.download.pdf/k-i_c-s_d.pdf
1660
414
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Swiss governments taking a position that upholds existing Swiss values like
the respect of human rights, the rule of law, democracy and liberal values.1665
The report on Artificial Intelligence, the Media and the Public
outlines the challenges associated with AI and mass media. It describes
current regulations and areas that could be improved. The report mentions
the importance of ensuring transparency, accountability and
traceability/comprehensibility when AI is deployed in journalism, in the
media or in social media.1666
The 2019 Report of the Interdepartmental Working Group on
Artificial Intelligence summarizes three reports by the federal government.
The report attempts to give an overview of AI, the current legal situation
and then considers AI in 17 different policy areas. The report emphasizes
the need for transparency, fairness or non-discrimination, accountability
and compliance with fundamental and human rights. The group points out
that the more human or fundamental rights are involved in a topic, the more
transparency and comprehensibility is required. They came to the
conclusion that no fundamental change to the Swiss regulatory framework
is necessary due to the fact that the legal principles of Swiss technology
policy are formulated in a technology-neutral manner which allows them to
be applied to AI systems. However, made several recommendations:
• AI should be monitored continuously as the report is only a
snapshot of the current situation and respective legal action
need to be taken when new developments are identified
• The Swiss government should engage in more international
information and knowledge sharing on AI governance
• AI policy should be integrated into the “Digital Switzerland”
Strategy
• Clarification of the 17 policy areas are necessary.1667
1665
Schweizerischer Eidgenossenschaft, International Gremien und Künstliche
Intelligenz, August 2019, https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/ig_k-i.pdf.download.pdf/i-g_ki_d.pdf
1666
Schweizerischer Eidgenossenschaft: Eidgenössisches Department für Umwelt,
Verkehr, Energie und Kommunikation UVEK, Künstliche Intelligenz, Medien und
Öffentlichkeit, August 2019,
https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/k-i_m-o.pdf.download.pdf/ki_m-o_d.pdf
1667
Schweizerische Eidgenossenschaft: Interdepartementalen Arbeitsgruppe «Künstliche
Intelligenz», Herausforderungen der künstlichen Intelligenz: Bericht der
interdepartementalen Arbeitsgruppe «Künstliche Intelligenz» an den Bundesrat,
https://www.sbfi.admin.ch/dam/sbfi/de/dokumente/2019/12/bericht_idag_ki.pdf.downloa
d.pdf/bericht_idag_ki_d.pdf
415
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
These recommendations were then integrated in the Digital Switzerland
Strategy 2020.1668 The Strategy 2020 emphasizes: Putting people at the
forefront, providing room for development, facilitating structural change,
and networking the shaping of transformation processes. 1669 And the
following key objectives:
• Enabling equal participation for all and strengthening solidarity
• Guaranteeing security, trust and transparency
• Continuing to strengthen people's digital empowerment and
self-determination
• Ensuring value creation, growth and prosperity
• Reducing the environmental footprint and energy consumption
In comparison to the 2018-2020 Strategy, the 2020-2022 Strategy
emphasizes “the aspects of data and environment.”1670 Transparency,
sustainable development and equal opportunities and participation have
been key objectives encompassed in these reports since the first version was
released in 2016.16711672
Swiss Foreign Policy and AI
One of the thematic focus areas of the Swiss Foreign Policy Strategy
2020-2023 is “digitalization.” The Ministry writes, “The focus is on
people’s needs. The rule of law and universal human rights – such as
freedom of expression and information and the right to privacy – must also
be guaranteed online. It is important to defend liberties such as press
freedom.” Furthermore, the Ministry seeks to “position Geneva as the
location for global digitalisation and technology debate” and to promote
Schweizerische Eidgenossenschaft: Federal Office of Communications, Digital
Switzerland, July 2020, https://www.bakom.admin.ch/bakom/en/homepage/digitalswitzerland-and-internet/strategie-digitale-schweiz.html
1669
Schweizerischer Eidgenossenschaft, Digital Switzerland Strategy, 2020,
https://www.digitaldialog.swiss/en/
1670
Schweizerischer Eidgenossenschaft: Federal Office of Communications OFCOM,
Digital Switzerland, July 2020,
https://www.bakom.admin.ch/bakom/en/homepage/digital-switzerland-andinternet/strategie-digitale-schweiz.html
1671
Schweizerischer Eidgenossenschaft: Federal Office of Communications OFCOM,
Digital Switzerland: Strategy, November 2018,
https://www.bakom.admin.ch/bakom/en/homepage/digital-switzerland-andinternet/strategie-digitale-schweiz/strategy.html
1672
Schweizerischer Eidgenossenschaft, Digital Switzerland» Strategy, September 2018,
https://www.bakom.admin.ch/dam/bakom/en/dokumente/informationsgesellschaft/strateg
ie/Strategie_DS_Digital_2-EN-barrierenfrei.pdf.download.pdf/Strategie_DS_Digital_2EN-barrierenfrei.pdf
1668
416
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
sustainable development using digital technologies, digital selfdetermination and cyber diplomacy.1673
In 2018 an expert group on the future of data processing and data
security published 51 recommendations for the federal government. The
federal government and its ministries adopted 31 of them. These included:
• “The Confederation and the cantons adapt the powers and
resources of the data protection authorities to enable them to
perform their statutory tasks of sensitization, consultation and
supervision comprehensively and effectively.
• “In cooperation with the cantons, the Confederation creates forms
of cooperation between data protection supervisory authorities
(e.g., competence center).”
• “In implementing the e-government strategy for Switzerland, the
Confederation and the cantons will ensure that the "offline"
population group is not socially excluded by digitization.”
• “The Confederation, cantons and municipalities promote open and
participatory systems and processes (…) in order to achieve social
goals such as digital transformation, resilience and sustainability
more quickly.”
• “The Confederation and the cantons ensure that students at upper
secondary schools and all students develop the basic skills and
competencies necessary for handling and shaping digital
technologies and transformation.”
• “The Confederation and the cantons are committed to ensuring that
the protection of fundamental values, human rights and human
dignity is also secured in the digital age and that informational
self-determination is promoted.”
• “The Confederation will ensure sufficient transparency,
traceability, comprehensibility and accountability of digital
processes and algorithms to create a trust-based digital economy
and society.”
Many of these recommendations coincide with ongoing activities.
For example, an association of the cantonal data protection authorities is in
place, the federal government is working with the cantons and universities
to integrate digital skills and knowledge into their respective curriculums
and the revision of the Data Protection Act which was approved by the
Schweizerische Eidgenossenschaft: Federal Department of Foreign Affairs FDFA,
Foreign Policy Strategy 2020-2023, January 2020,
https://www.eda.admin.ch/dam/eda/en/documents/publications/SchweizerischeAussenpol
itik/Aussenpolitische-Strategie-2020-23_EN.pdf
1673
417
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Swiss parliament in 2020.1674 If it is not challenged in a referendum then it
is set to come into force in 2021.1675 The revision improves transparency for
citizens, gives the Swiss Federal Data Protection and Information
Commissioner more competencies and resources and aligns Swiss data
protection law with the European Data Protection Regulation.1676
Independent AI oversight
The Federal Data Protection and Information Commissioner
(FDPIC) is the “competent authority for data processing by federal bodies
and private persons, including enterprises.” Furthermore, data processing
by cantonal or communal authorities is supervised by cantonal and
communal data protection commissioners.1677 The revision of Swiss Data
Protection Act ascribed more competencies and resources to the FDPIC
which should allow for more comprehensive oversight over the regulation
of the data protection laws. 1678
Further, in 2019 the Federal Council approved the proposal to
establish a national human rights institution (NHRI). The NHRI is the result
of a pilot project called the Swiss Centre of Expertise in Human Rights
(SCHR). According to the Federal Council “The NHRI will be independent,
Schweizerischer Eidgenossenschaft: Eidgenössisches Department für Umwelt,
Verkehr, Energie und Kommunikation UVEK, Bericht zu den Empfehlungen der
Experten- gruppe zur Zukunft der Datenbearbeitung und Datensicherheit: Kenntnisnahme
und weiteres Vorgehen, October 2019,
https://www.bakom.admin.ch/dam/bakom/en/dokumente/informationsgesellschaft/datenp
olitik/empfehlungen_experten_gruppe.pdf.download.pdf/Bericht%20zu%20den%20Emfe
hlungen%20der%20Expertengruppe.pdf
1675
Swiss IT Magazine, Entwurf zur Totalrevision des Datenschutzgesetzes angenommen
(Sept. 28, 2020),
https://www.itmagazine.ch/artikel/73072/Entwurf_zur_Totalrevision_des_Datenschutzge
setzes_angenommen.html
1676
Schweizerischer Eidgenossenschaft: Eidgenössisches Justiz- und Polizeidepartment
EJDP, Den Datenschutz verbessern und den Wirtschaftsstandort stärken (Sept. 2017),
https://www.bj.admin.ch/ejpd/de/home/aktuell/news/2017/2017-09-150.html
1677
Schweizerischer Eidgenossenschaft: Federal Data Protection and Information
Commissioner, Data Protection – Switzerland,
https://www.edoeb.admin.ch/edoeb/en/home/the-fdpic/links/data-protection--switzerland.html
1678
Schweizerischer Eidgenossenschaft: Eidgenössisches Justiz- und Polizeidepartment
EJDP, Den Datenschutz verbessern und den Wirtschaftsstandort stärken (Sept. 2017),
https://www.bj.admin.ch/ejpd/de/home/aktuell/news/2017/2017-09-150.html
1674
418
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
include members from across society and receive an annual financial
support from the Confederation.”1679
Public Participation
Switzerland is a direct democracy which means that citizens have
the right to decide on policy directly, either by referendums or citizeninitiated referendums.16801681 Further, policy revisions or proposals of
importance go through a consultation procedure with relevant stakeholders
to include their opinions and needs and therefore to minimize the chance of
a referendum.1682
Further specifically in technology policy, the website for the
“Digital Switzerland” Strategy (www.digitaldialog.swiss) provides a
summary of the Strategy and lists related initiatives and committees.
Updates on how the Strategy is being implemented are also published on
this website.1683 Through the digital dialogue website, organisations,
companies, municipalities and cantons can propose measures for a Digital
Switzerland in the action plan. You can directly upload a proposal on this
website and it will be reviewed in the context of the “Digital Switzerland”
action plan.1684
In 2018, the Federal Council established the opendata.swiss website,
“the Swiss public administration’s central portal for open government
data.”1685 The website, managed by the Federal Statistical Office, “supports
organisations in publishing their open data” and “continuously monitors the
quality of the catalogue.”
Finally, the “Plateforme Tripartite Suisse” is an information hub and
platform to exchange dialogue. It was founded in light of the “World
Summit on the Information Society” in 2003 to prepare for this conference.
It now “serves as a national forum for the informal exchange of information
and experience on WSIS implementation and follow-up activities. It is open
to all interested representatives from the administration, the business sector,
The Federal Council, National human rights institution to be established in
Switzerland (Dec. 13, 2019), https://www.admin.ch/gov/en/start/documentation/mediareleases.msg-id-77508.html
1680
Schweizerischer Eidgenossenschaft: Bundeskanzlei, Referenden,
https://www.bk.admin.ch/bk/de/home/politische-rechte/referenden.html
1681
Schweizerischer Eidgenossenschaft: Bundeskanzlei, Volksinitiativen,
https://www.bk.admin.ch/bk/de/home/politische-rechte/volksinitiativen.html
1682
Schweizerischer Eidgenossenschaft: Der Bundesrat: Das Portal der Schweizer
Regierung, Vernehmlassungen (July 2019),
https://www.admin.ch/gov/de/start/bundesrecht/vernehmlassungen.html
1683
Digital Switzerland Strategy, https://www.digitaldialog.swiss/en/
1684
Digital Switzerland Strategy, https://digitaldialog.typeform.com/to/ulwdzc
1685
opendata.swiss, Portal, https://opendata.swiss/en/about/
1679
419
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
civil society and the internet community at the national level and meets on
an ad-hoc basis.”1686
AI Events in Geneva
Switzerland is a hub for international policy as many international
organizations are based in Geneva. The United Nations hosted the first
annual AI for Good Global Summit in 2017.1687 Co-organized by ITU and
the XPRIZE Foundation, the event convened 500 representatives of
government, industry, academic and research institutes, United Nations
agencies and civil society to “explore the potential of AI to accelerate
progress in fields such as health, education, energy and the protection of our
environment.” The most recent AI for Good Summit was held online.
In October 2020, the Swiss Federal Institute of Technology (ETH)
in Zürich launched the Center for AI. The research center comprises 29
professorships, a new executive director and a fellowship program.1688 The
vision for this research center is to “lead the way towards trustworthy,
accessible, and inclusive AI systems for the benefit of society.” Among the
key goals:
• To “contribute to international networks (Europe & beyond)”
• To “provide a stimulating, transdisciplinary, and inclusive
environment”
• To “address ethical, societal, and policy implications”
• To “engage with the general public on AI topics1689
AI and Criminal Justice
According to AlgorithmWatch, the Swiss government is using AI in
the penal system. The application helps to triage inmates, which is the first
of several steps to develop the inmate’s release plan. Further, the Federal
government uses a system for automatic vehicle detection and traffic
monitoring at state borders. Finally, the Federal Customs Administration
uses a different system, like a data processing system, to help assess goods
coming into the country. Some cantons use a police software to predict
domestic burglaries. According to AlgorithmWatch, “it should be noted that
Schweizerischer Eidgenossenschaft : Federal Office of Communications OFCOM,
May 2007, https://www.bakom.admin.ch/bakom/en/homepage/ofcom/internationalactivities/un-world-summit-on-the-information-society/wsis.html
1687
International Telecommunications Union, Artificial Intelligence
https://www.itu.int/en/ITU-T/AI/Pages/201706-default.aspx
1688
ETH Zürich, New Centre for AI research (Oct. 20, 2020), https://ethz.ch/en/newsand-events/eth-news/news/2020/10/pr-new-centre-for-ai-research.html
1689
ETH Zürich: ETH AI Center, About Us, https://ai.ethz.ch/about-us.html
1686
420
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the use of predictive policing in Switzerland is currently limited to a
relatively small and clearly defined area of preventive police work.”1690
Algorithmic Transparency
Switzerland is outside the European Union and is therefore not
directly subject to the GDPR. The Federal Act on Data Protection (FADP)
was revised as of September 2020 to comply with the modernized Council
of Europe Convention 108. 1691 Switzerland ratified the Council of Europe
Convention in 2019.1692 Article 9(1)(c) of the Convention gives every
individual the right “to obtain, on request, knowledge of the reasoning
underlying data processing where the results of such processing are applied
to him or her.”1693 The Swiss Data Protection Act establishes a Data
Protection and Information Commissioner (FDPIC) with independent
supervisory authorities. The Act creates obligations to undertake privacy
impact assessments in certain circumstances. A key amendment increased
transparency in data processing.1694
Article 21 of the new FADP introduces the "Duty to inform in the
case of an automated individual decision."1695 The FADP states that: (1)
“The person responsible shall inform the person concerned of a decision
that is based exclusively on automated processing that is associated with a
legal consequence for them or significantly affects them (automated
individual decision) and (2) “On request, it shall give the data subject the
opportunity to state his or her position. The data subject may request that
the automated individual decision be reviewed by a natural person."
Further, article 25(g) establishes a data subject's right to algorithmic
transparency in the case of an automated individual decision: "In any case,
the following information will be communicated to the data subject: g. if
AlgorithmWatch, Automating Society 2020, (Oct.
2020), https://automatingsociety.algorithmwatch.org/report2020/switzerland/)
1691
Linklaters, Data Protected – Switzerland (Oct. 2020), https://www.linklaters.com/enus/insights/data-protected/data-protected---switzerland
1692
Council of Europe, https://www.coe.int/en/web/conventions/full-list//conventions/treaty/223/signatures
1693
Council of Europe, 128th Session of the Committee of Ministers (May 17-18, 2018),
https://search.coe.int/cm/Pages/result_details.aspx?ObjectId=09000016807c65bf
1694
Lexology, Revision of Swiss data protection act adopted (Sept. 25, 2020),
https://www.lexology.com/library/detail.aspx?g=ebc8ce19-0fee-457d-a94fa0625e4805b8
1695
Swiss Parliament, Vorlage der Redaktionskommission für die Schlussabstimmung,
Bundesgesetz über den Datenschutz (DSG) (Sept. 25, 2020) (“Presentation by the
editorial board for the final vote on the Data Protection Act”) [DT]ra,
https://www.parlament.ch/centers/eparl/curia/2017/20170059/Schluzssabstimmungstext
%203%20NS%20D.pdf
1690
421
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
applicable, the existence of an automated individual decision and the logic
on which the decision is based."
OECD AI Principles
Switzerland endorsed the G20 AI Principles. Regarding
implementation of the AI Principles, the OECD notes Switzerland’s active
involvement in relevant international organisations and processes.
“Particularly important for Switzerland is to ensure that fundamental and
established values and norms such as human rights are respected and that
all relevant stakeholders are involved in decision-making.”1696
Human Rights
Switzerland is a signatory to many international human rights
treaties and conventions. Switzerland typically ranks among the top nations
in the world for the protection of human rights and transparency.1697 In 2021,
Freedom House gave Switzerland the score of 96/100, unchanged from
2021, and noted that “the government is generally transparent in its
operations. In recent years, an increasing number of cantonal governments
have passed transparency laws that make government data more accessible
to citizens.”
Evaluation
Switzerland has newly established a national set of guidelines on
ethics that are aimed at the public administration. Further, across most
reports and initiatives, ethics have been considered, integrated and
implemented in the governments work on AI policy. However, there is no
clear regulatory strategy for the private sector. There is also, at the moment,
no express support for the Universal Guidelines for AI. Switzerland is one
of many countries that endorsed the UNESCO Recommendation on AI
Ethics.
G20 Digital Economy Task Force, Examples of National AI Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf.
1697
Freedom House, Freedom in the World 2021: Switzerland (2021),
https://freedomhouse.org/country/switzerland/freedom-world/2021
1696
422
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Taiwan
AI Action Plan
“Beyond sparking a scientific and technological revolution, artificial
intelligence (AI) will fundamentally transform human life and industry and
create boundless business opportunities.” This sentence officially
introduces the 2018 Taiwan Government’s four-year AI Action Plan1698
which aims to propel Taiwan “into the ranks of the world's leading smart
nations.”
With a total budget of 1.1 billion EUR over 2018-2021, “guided by
the principles of deregulation, open access and technology investment,” the
AI Action Plan is designed to “sharpen Taiwan's advantages, prioritize
innovation and real-world implementation, and develop software and
hardware in tandem, thereby injecting greater momentum into Taiwan's
industries.”
To this end, the Action Plan focuses on five action areas: i)
developing AI talent; ii) promoting Taiwan's lead role in AI by expanding
its world-leading position in the semiconductor chip industry; iii) building
Taiwan into an AI innovation; iv) liberalizing laws and opening test grounds
to ease restrictions on innovative technologies; and v) Transforming
industry with AI.
The AI Action Plan follows on the five-year AI strategy developed
by the Ministry of Science and Technology1699 (MOST) to “cultivate AI
technology specialists and create an environment for AI scientific research.”
This “Grand Strategy for a Small Country1700” has a budget of 490 million
EUR over 2017-2021 and builds on Taiwan’s “strengths and potential
advantages, such as semiconductors and information and communications
technology.”
The AI Strategy aims to “develop select fields for the future,
including the internet of things, security solutions and driverless vehicles”
1698
Government of Taiwan, Executive Yuan, AI Taiwan Action Plan (Aug. 7, 2019),
https://english.ey.gov.tw/News3/9E5540D592A5FECD/1dec0902-e02a-49c6-870de77208481667. The Executive Yuan is an executive branch of the Taiwan Government. It
is a Council headed by the premier which includes the vice-premier, ministers,
chairpersons of commissions, and ministers without portfolio. See also: Government of
Taiwan, AI Taiwan, https://ai.taiwan.gov.tw/#actionplan
1699
The Ministry of Science and Technology (MOST) is one of the ministries under the
Executive Yuan in Taiwan and is responsible for the scientific and technological
innovation of Taiwan.
1700
Executive Yuan, R. O. C. (Taiwan). AI innovation: Grand strategy for a small
country (-Major Policies Detail)
https://english.ey.gov.tw/News3/9E5540D592A5FECD/edadb735-e6a6-43e1-ac931959602bb3ec
423
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and has five “key facets”: i) R&D services with the creation of an AI cloud
service and high-speed computing platform; ii) Value-added innovation
with the establishment of four AI innovation research centers; iii) Creativity
and practice with an AI Robot Makerspace; iv) Industrial pilot program with
an AI semiconductor "moonshot" project; and v) Social participation with
three "Formosa Grand Challenge" technology competitions to uncover
talent, develop technology and stimulate creativity. Ethical questions are
also targeted in Taiwan – as one of the many aspects related to AI.1701
AI Core Values
MOST announced in September 2019 AI Technology R&D
Guidelines “in a bid to create a reliable environment conforming to
international trends of AI R&D and to provide directions for Taiwan AI
researchers to follow.”1702 When presenting the Guidelines Science and
Technology Minister Chen Liang-gee said that his ministry “has the
responsibility of helping humans be able to trust” AI1703 and that “those who
provide digital tools must be ethical."1704 He also explained that the whole
world is still watching the evolution of artificial intelligence and that it is
right now “more appropriate to adopt guidelines than sanctions.”
The AI R&D Guidelines are based on three core values1705: (1)
Human-centered (the human being should be at the heart of research, an AIbased society should respect human dignity, rights and freedom, and
application of AI is to prompt human welfare and hike human living
standards); (2) Sustainable development (AI R&D should seek balance
among economic growth, social progress and environmental protection to
reach co-existence and common prosperity among human being, society
and environment); (3) Diversity and inclusion (AI R&D is to create an AIbased human society of diverse value concepts and backgrounds via
interdisciplinary dialog mechanisms).
1701
Netherland Innovation Network, Artificial Intelligence; an overview of policies and
developments in Taiwan (Mar. 2020), https://www.rvo.nl/sites/default/files/2020/04/AIDevelopments-in-Taiwan.pdf
1702
Bryan Chuang, Adam Hwang, MOST announces AI R&D guidelines, DigiTIMES
(Sept. 24, 2019), https://www.digitimes.com/news/a20190923PD209.html
1703
Shirley Lin, Taiwan first in world to set R&D guidelines for AI, Radio Taiwan
International (Sept. 23, 2019), https://en.rti.org.tw/news/view/id/2001895
1704
Radio Taiwan International, The French Ministry of Science defines a research
framework for artificial intelligence (Sept. 23, 2019) (original in French) [GT],
https://fr.rti.org.tw/news/view/id/90832
1705
Bryan Chuang, Adam Hwang, MOST announces AI R&D guidelines, DigiTIMES
(Sept. 24, 2019), https://www.digitimes.com/news/a20190923PD209.html
424
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI R&D Guidelines
“AI research and development must be people-oriented,” the
Minister said, asking that research teams retain the source codes and AI
training materials so that the work can be traced. The complexity of AI
means that it is vulnerable to misuse, which countries are seeking to mitigate
by establishing standards for its development, he said. In particular, because
AI technology systems learn from data, they can perpetuate and amplify
human biases, he said.1706 “After an artificial intelligence program is written,
it evolves based on the data provided to it. If the data is discriminatory, the
program will be discriminatory. If the data is deviant, it will be deviant."
More precisely, eight guidelines derive from Taiwan’s AI core
values, including i) Common good and well-being; ii) Fairness and nondiscrimination; iii) Autonomy and control; iv) Safety; v) Privacy and data
governance; vi) Transparency and traceability; vii) Explainability; and viii)
Accountability and communication.1707
The official press release points to the fact that the core of AI
technology is its people-oriented nature, so researchers must safeguard
human rights and preserve human dignity and that the guideline’s eight
criteria emphasize promoting shared benefits and common well-being, with
researchers who should aim to develop systems that are free of
discrimination. Likewise, AI tools should support human policies, and
people using the tool.1708
“Human-centered AI” for Taiwan AI start-up iKala means the
involvement of people in AI development and applications, as well as the
creation of ethical, humane AI. As an illustration, iKala Co-founder and
CEO Cheng cites the example of an innovative picture-as-a-service
(PicaaS) technology which automatically edits product pictures to allow
companies to circulate clean photos of their products. However, following
complaints that it could potentially be used to infringe on image owners’
intellectual property rights, Cheng and his team re-trained the software to
recognize and reject copyrighted images. “That’s the kind of responsible AI
Radio Taiwan International, The French Ministry of Science defines a research
framework for artificial intelligence (Sept. 23, 2019) (original in French) [GT],
https://fr.rti.org.tw/news/view/id/90832
1707
Bryan Chuang, Adam Hwang, MOST announces AI R&D guidelines, DigiTIMES
(Sept. 24, 2019), https://www.digitimes.com/news/a20190923PD209.html
1708
Overseas Community Affairs Council, AI Taiwan: Ministry releases guidelines for AI
research (Sept. 24, 2019),
https://ai.taiwan.gov.tw/news/ministry-releases-guidelines-for-ai-research/
1706
425
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
we want to be working on,” says Cheng. “Putting humans in the equation –
not just stealing and not just replacing people.”1709
Medical Data and AI Ethics
Two research projects on AI Ethics in the medical and biomedical
areas are ongoing. One examines the Ethical, Legal, and Societal Issues
Surrounding Artificial Intelligence-Assisted Medical Care (ELSI-AIM1710)
and is in its second year. Another one (NCKU AI Biomedical Research
Center on AI Ethics) focuses on AI for biomedical research with a multidisciplinary team of clinicians, biomedical, AI experts, legal and ethical
advisors. 1711The Taiwan Biobank created in 2012 is a repository of tissues/
information but is not allowed to directly carry out research. Information on
the participants, all voluntary, whose samples are included in the biobank
will link several sources of data: national identification number, National
Health Insurance system, cancer registry, and cause of death registry.1712 Its
operations are regulated by detailed legislation.
Current discussion seems to focus on how the Taiwan Biobank can
be transformed by leveraging digital technologies. For some, greater
participant engagement and the uptake of Information Technology (IT) and
Artificial Intelligence (AI) applications can be used in partnership with
vertical and horizontal integration as part of a four-pronged approach to
promote biobank sustainability, and facilitate the biobank’s
Jeremy Olivier, Taiwan Tests the Limits with Artificial Intelligence, Taiwan Business
(May 15, 2020), https://topics.amcham.com.tw/2020/05/taiwan-tests-limits-
1709
ai/
Joint Research Center for AI Technology and All Vista Healthcare sponsored by:
Ministry of Science and Technology http://mahc.ntu.edu.tw/en/research_view.php?id=13
This project includes four subprojects: (1) The deliberation of ethical issues on artificial
intelligence-assisted medicine; (2) legal and policy implications of artificial intelligence
in medicine; (3) the implications of societal issues on artificial intelligence-assisted endof-life physician-patient communication: opportunities and challenges; and (4) artificial
intelligence-based medicine assisted system from analytical design to practical
application.
1711
NCKU - MoST AI Biomedical Research Center The MoST AI Biomedical Research
Center is located at the National Cheng Kung University (NCKU). The research center
focuses on AI for biomedical research. Currently there are fourteen projects ongoing with
a focus in four areas: smart medicine, smart healthcare, smart biotechnology and ethics
and humanities.
1712
Michael Cheng-tek, Taiwanese Experience of Data-Sharing in Biobanking (PPT
slides)
1710
426
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
transformation.1713 Others seem more cautious,1714 pointing to key issues
raised by the current ethical governance1715 of the Taiwan Biobank, namely
i) the handling of ethnicity, including the special requirements that it
imposes with respect to obtaining participant consent, and ii) transparency
(and accountability) around the undertaking’s governance.
Covid-19 and Big Data Analytics
In January 2020, Taiwan integrated its national health insurance
database with its immigration and customs database to begin the creation of
big data for analytics and allow case identification by generating real-time
alerts during a clinical visit based on travel history and clinical
symptoms.1716
"The combination of these two sets of data allows us to generate
alerts to tell healthcare staff when a patient returns from a risk area," said
Yu-Lun Liu, doctor in the intelligence department of the Taiwan Center for
disease control (CDC). The identified patients are then treated through a
separate circuit, limiting contact with other patients. 1717 To strengthen this
new data set, the authorities are “working with telephone providers, on the
basis of roaming data, to identify people whose last stopover is not
necessarily an area at risk, but who have made trips with stops in areas
affected by the epidemic. " “The authorities are developing a model for
processing public video surveillance images to estimate the proportion of
masked people. "This artificial intelligence-based model has allowed us to
1713
Journal of Translational Medicine - Transformation of the Taiwan Biobank 3.0:
vertical and horizontal integration by Jui-Chu Lin, Wesley Wei-Wen Hsiao and Chien-Te
Fan
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7406956/
1714
August 2018 – Journal of Law, Technology and Society - Biobank Governance: The
Cautionary Tale of Taiwan Biobank by Shawn H.E. Harmon, Shang-Yung Yen and ShuMei Tang
https://script-ed.org/article/biobank-governance-the-cautionary-tale-of-taiwan-biobank/
1715
Taiwan Biobank established an Ethics and Governance Council (EGC) to act as an
independent guardian of Taiwan Biobank’s Ethics and Governance Framework, and to
advise the Competent Authority (the MOHW) on its revision from time to time. Cited
above: https://script-ed.org/article/biobank-governance-the-cautionary-tale-of-taiwanbiobank/
1716
March 3, 2020 - How Taiwan Used Big Data, Transparency and a Central Command
to Protect Its People from Coronavirus by Beth Duff-Brown
https://fsi.stanford.edu/news/how-taiwan-used-big-data-transparency-central-commandprotect-its-people-coronavirus
1717
L’Usine Digital, Covid-19 : comment Taïwan s'est appuyé sur la technologie pour
contenir l'épidémie (March 20, 2020),
https://www.usine-digitale.fr/article/covid-19-comment-taiwan-s-est-appuye-sur-latechnologie-pour-contenir-l-epidemie.N943431
427
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
see a rapid increase in the number of people wearing masks. We have
chosen to stop their export and increase local production," recalls Yu-Lun
Liu. "
In March 2020, the BBC reported that when the phone belonging to
an American University student in Taiwan, who was subject to 14 days'
quarantine after returning from Europe, ran out of battery power, in less
than an hour he had received phone calls from four different local
administrative units, a text message notifying him he would be arrested if
he had broken quarantine, and a visit from two police officers. The phone
tracking system uses phone signals to triangulate locations of the more than
6,000 people subject to home quarantine; an alert is sent to the authorities
if the phone is turned off for more than 15 minutes.1718
Autonomous vehicles
In November 2018, the Legislative Yuan passed the Act for
Unmanned Vehicle Technology Innovative Experiments, which was
enacted by the President in December 2018. Entered into force in May 2019,
the Act frees autonomous vehicles and drones from limits by some traffic
regulations in their test runs.1719 Moreover, the regulations specifically call
for AI-boosted algorithmic unmanned platforms.1720 Taiwan CAR
(Connected, Autonomous, Road-test) Lab,1721 the nation's first closed field
for testing self-driving cars, also opened for use in 2019.
1718
BBC, Coronavirus: Under surveillance and confined at home in Taiwan (March 24,
2020), https://www.bbc.co.uk/news/technology-52017993
1719
GNSS Asia, Taiwan’s Executive Yuan Approves Bill Promoting Unmanned Vehicle
Experimentation (May 24, 2019), https://gnss.asia/new/taiwans-executive-yuanapproves-bill-promoting-unmanned-vehicle-experimentation/
1720
Yisuo Tzeng, Prospect for Artificial Intelligence in Taiwan’s Defense, Jewish Policy
Center (Winter 2019), https://www.jewishpolicycenter.org/2019/01/11/prospect-forartificial-intelligence-in-taiwans-defense/
1721
Taiwan Car Lab,
http://taiwancarlab.narlabs.org.tw/index_en.html#:~:text=The%20Taiwan%20CAR%20(
Connected%2C%20Autonomous,evaluation%20of%20self%2Ddriving%20vehicles.&tex
t=Taiwan%20CAR%20Lab%20is%201.75%20hectare.
428
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Facial Recognition
Facial recognition is implemented in Taiwan in various sectors, such
as banks,1722 retail stores,1723 airports,1724 and law enforcement.1725 In June
2019, the Taiwan Railways Administration announced that, due to privacy
concerns, its surveillance system trial would not include facial
recognition.1726 The artificial intelligence–based surveillance will still be
capable of monitoring passenger behavior, including trespassing, loitering
in restricted areas and prohibited acts.1727
More recently, Liao Wei-min, associate professor at Taiwan
National Chung Hsing University’s Department of Law called for
legislation regarding facial recognition and data collection.1728 He wrote,
“What is essentially a beneficial technology is deeply problematic given the
lack of precise and targeted legislation, and this is the fault not of civil
servants, but their politically appointed masters. Something needs to be
done to address this.”
Algorithmic Transparency
The Taiwan Personal Data Protection Act (PDPA) regulates “the
collection, processing and use of personal data so as to prevent harm on
personality rights, and to facilitate the proper use of personal data.” 1729 The
PDPA applies to AI applications, but does not explicitly provisions
Luana Pascu, Public Bank of Taiwan rolls out Kneron edge AI facial recognition,
Biometric Update (Sept. 24, 2019), https://www.biometricupdate.com/201909/publicbank-of-taiwan-rolls-out-kneron-edge-ai-facial-recognition
1723
Telpo, 7-Elevm Open the 2nd Face Recognition Unstaffed Store in Taiwan (Nov. 20,
2018). https://www.telpo.com.cn/blog/7-eleven-taiwan-face-recognition-store.html
1724
Gorilla, Taiwan's International Airports Implement Gorilla Biometrics Technology to
Improve Airport Operations and Security (Dec. 23, 2016), https://www.gorillatechnology.com/Press-Room/Taiwan's-International-Airports-Implement-GorillaBiometrics-Technology-to-Improve-Airport-Operations-and-Security
1725
AsiaOne, Privacy not violated by facial recognition technology, says Taiwan police
agency (May 28, 2014), https://www.asiaone.com/asia/privacy-not-violated-facialrecognition-technology-says-taiwan-police-agency
1726
Focus Taiwan, TRA to cut facial recognition feature from surveillance system trial
(Nov. 6, 2019), https://focustaiwan.tw/society/201911060011
1727
IAPP, Taiwan Railways Administration excludes facial recognition from surveillance
trial (Nov. 7, 2019), https://iapp.org/news/a/taiwan-railways-administration-excludesfacial-recognition-from-surveillance-trial/
1728
Liao Wei-min, Legislation needed for advances in surveillance, Taipei Times (Dec.
28, 2019), http://www.taipeitimes.com/News/editorials/archives/2019/12/28/2003728301
1729
Laws and Regulations Database of the Republic of China, Personal Data Protection
Act. https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=I0050021
1722
429
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
regarding algorithmic transparency such as those found in the GDPR. No
specific laws or regulations governing civil liability regarding AI. 1730
Use of AI for digital democracy
For the last few years, Taiwan has organized public debates via the
citizen-run vTaiwan
platform.1731 vTaiwan’s (for Virtual
Taiwan) algorithms highlight where there is consensus in a debate while
minimizing the voices at the most extreme ends.1732 This system, is officially
and routinely part of the law-making process of Taiwanese institutions,
involving thousands of citizens in varying degrees.1733
The vTaiwan process giving weight to the citizen voice and that has
led to real regulatory innovations at four stages: (1) informing the public;
(2) collecting the strategic approaches through Pol.is,1734 an opensource self-learning algorithm; (3) deliberating; and (4) observing decisionmaking.1735
“When people started using Polis, we found that it became a
consensus-generating mechanism,” Megill said to Wired.1736 To bring the
groups closer together, Polis has reengineered many of the features we take
for granted on social media. No reply button – hence no trolling. No echochambers, replaced by an attitudes map showing you where you are in
relation to everyone else. The platform does not highlight the most divisive
statements, but gives more visibility to the most consensual ones. The ones
that get attention are those that find support not only in one cluster, but
across other groups, too.
The outcomes of vTaiwan have been put in front of Parliament, by
government, to form the core of 11 pieces of laws and regulation, with eight
more waiting to go on everything from revenge porn to fintech regulation.
Global Legal Insights, AI, Machine Learning & Big Data 2020 | Taiwan,
https://www.globallegalinsights.com/practice-areas/ai-machine-learning-and-big-datalaws-and-regulations/taiwan
1731
vTaiwan, About, https://info.vtaiwan.tw/
1732
Walter Kerr, Taiwan Is Beating Political Disinformation. The West Can Too, Foreign
Policy (Nov. 11, 2020), https://foreignpolicy.com/2020/11/11/political-disinformationtaiwan-success/
1733
bluenove, vTaiwan : making citizens the key to public debate (May 28, 2018),
https://bluenove.com/en/blog/vtaiwan-making-citizens-the-key-to-public-debate/
1734
The Computational Democracy Project, Designing Future Democracies, h
https://github.com/pol-is/polis-documentation/blob/master/README.md
1735
Bluenove, vTaiwan : making citizens the key to public debate (May 28, 2018),
https://bluenove.com/en/blog/vtaiwan-making-citizens-the-key-to-public-debate/
1736
Wired, Taiwan is making democracy work again. It's time we paid attention (Nov.
26, 2019), https://www.wired.co.uk/article/taiwan-democracy-social-media
1730
430
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Data Protection
Personal data protection in Taiwan is essentially subject to the 2015
Personal Information Protection Act (PDPA) which applies to the public
and the private sectors.1737 It is supplemented by the 2016 Enforcement
Rules of the Personal Data Protection Act.1738 Local and national
government authorities enforce these Act.
The Taiwan Government considers amending the PDPA to meet
GDPR standards to obtain an adequacy status decision from the EU and
held several public hearings in 2019 to solicit public comments. Among the
various topics discussed during the public hearings, the government is
contemplating the adoption of data breach notification obligations and
cross-border data transfer restrictions similar to those under GDPR. The
government is also planning to establish an independent data protection
authority.1739 In July 2020, Taiwan’s digital minister said that she supported
the idea of establishing a dedicated agency for personal data protection
before the electronic identification cards (eID) are rolled out next year.1740
OECD AI Principles
Although Taiwan is not an OECD member country many of its AI
policies align with the OECD AI Principles. The Ministry of Science and
Technology noted in the announcement of the AI R&D Guidelines that
“many countries and organizations have established ethics standards for AI
R&D, such as the EU's Ethics Guidelines for Trustworthy AI, OECD's
Principles on Artificial Intelligence and IEEE's Ethically Aligned DesignVersion II.”1741 In some respects, the AI R&D Guidelines of Taiwan go
beyond the OECD Principles and reflect the broader goals of the Universal
Guidelines for AI.
Taiwan, Personal Data Protection Act (Dec. 30, 2016),
https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=I0050021#:~:text=The%20P
ersonal%20Data%20Protection%20Act,proper%20use%20of%20personal%20data.&text
=%22data%20subject%22%20refers%20to%20an,is%20collected%2C%20processed%2
0or%20used.
1738
Taiwan, Enforcement Rules of the Personal Data Protection Act (March 2, 2016),
https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=I0050022
1739
OneTrust, Taiwan - Data Protection Overview (July 2020),
https://www.dataguidance.com/notes/taiwan-data-protection-overview
1740
Huang Tzu-ti, Taiwan’s digital minister says personal data protection agency needed
for digital ID: Measure to introduce eID has been met with fierce opposition from
academics, experts, Taiwan News (July 30, 2020),
https://www.taiwannews.com.tw/en/news/3976854
1741
Digitimes, MOST announces AI R&D guidelines (Sept. 24, 2019),
https://www.digitimes.com/news/a20190923PD209.html.
1737
431
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Fundamental Rights
Taiwan is a signatory to many international human rights treaties
and conventions and ranks high in the world for the protection of human
rights and transparency.1742 Regarding transparency, Freedom House notes
that “the 2005 Freedom of Government Information Law enables public
access to information held by government agencies, including financial
audit reports and documents about administrative guidance. Civil society
groups are typically able to comment on and influence pending policies and
legislation.”
Evaluation
Although Taiwan is not an OECD member country many of its AI
policies align with the OECD AI Principles. There is also a robust public
debate about the use of AI for facial recognition, medical data, and
autonomous vehicles. But privacy concerns arise with the integration of
government data sets while the government has not established an
independent data protection agency that could oversee AI applications from
a privacy perspective.
1742
Freedom House Report: Taiwan (2020), https://freedomhouse.org/country/taiwan
432
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Thailand
National AI Strategy
Thailand does not have a national AI strategy. Digitalization and AI
objectives are covered under 20-year national strategy and various plans
developed centrally by the National Council for Peace and Order (NCPO)
government that took over the government with a military coup in 2014.
• 20-Year National Strategy (2017-2036): provides a vision of
“Thailand as a developed country with security, prosperity, and
sustainability in accordance with the principles of the Sufficiency
Economy Philosophy” Government commits that the targets and
indicators will also have to abide by the internationally accepted
2030 Sustainable Development Goals.”1743
• Thailand Digital Economic and Society Development Plan
(2017-2021): The plan defines Digital Thailand as a “transformed
Thailand that maximizes the use of digital technologies in all
socio-economic activities in order to develop infrastructure,
innovation, data, human capital, and other digital resources that
will ultimately drive the country towards wealth, stability, and
sustainability.”1744 Plan highlights creating a knowledge-driven
digital society by building participation, ensuring inclusive and
equal usage; creating open government; building trust and
confidence in the use of digital technology and updating laws and
regulations as some of the strategies.
• Thailand Digital Government Development (TDG) Plan (20172021): Developed by Electronic Government Agency (EGA),
objective is to strategically digitalize Thai government agencies, to
deliver best citizen-centric services, with a high level of efficiency
and transparency1745. One of the goals of the plan is to enable
people to access accurate public information conveniently, to
enhance transparency, people’s confidence and trust in the
government, and the successful participation of civil society.
Government of Thailand, The Twelfth National Economic and Social Development
Plan (2017-2021), https://www.nesdc.go.th/nesdb_en/ewt_dl_link.php?nid=4345
1744
Thailand Digital Economy and Society Development Plan (2016),
https://www.itu.int/en/ITUD/RegionalPresence/AsiaPacific/Documents/Events/2016/AprDigital2016/S2_Present_Pansak_Siriruchatapong.pdf
1745
Thailand Digital Government Development Plan 2017-2021,
http://jfcct.bypronto.com/wp-content/uploads/sites/1871/2018/05/Digital-GovernmentDevelopment-Plan-2017-2021-executive-version.pdf
1743
433
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
12th National Economic and Social Development Plan (20172021): focuses on using artificial intelligence (AI) and embedded
technology to create a “just and quality society which leaves no
one behind, based on the collaboration of all Thai stakeholders.”1746
• Thailand 4.0 (2016): This policy aims to develop Thailand into a
smart and technology-driven economy. Digitalization and AI
adoption sit in the core as new engines of growth transform the
society and industries.1747 The policy includes Smart Cities and
Digital Park Thailand, which is an economic digital innovation
zone that serves as the hub of the ASEAN sub-region.
The national policy and plans are formulated with an emphasis on
digital technologies empowering people to increase their ‘citizen well-being
and quality of life.’ Increasing the transparency and accountability of public
agencies is mentioned several times in these plans. However, it is not clear
how the goals will be adopted by the agencies and how the actions will be
coordinated across different levels of government.
There is no explicit mention of concepts such as fundamental rights,
human rights, rule of law, fairness with respect to AI in the plans. 12th
National Economic and Social Development Plan acknowledges that
Thailand “has high inequality and a lack of fairness”1748 across society. The
only mention of any AI ethics guidelines is acknowledging the existence of
EU Ethics Guidelines for Trustworthy AI, and the World Government
Summit’s Ethical AI Systems Design through a reference within a TDG
Plan summary document1749. EGA, as the agency responsible to implement
standards, models, measures, principles and approaches in the form of
digital technology, only mentions openness, integrity and collaboration
under its eight core values.1750
In 2019, the Ministry of a Digital Economy and Society, academics
and experts from Mahidol University and Microsoft Thailand joined
together to draft Digital Thailand – Draft AI Ethics Guidelines, to serve as
a manual and provide ethical codes for AI development in government
•
Government of Thailand,
https://www.nesdc.go.th/nesdb_en/ewt_dl_link.php?nid=4345
1747
Royal Thai Embassy, Thailand 4.0, https://thaiembdc.org/thailand-4-01748
The Twelfth National Economic and Social Development Plan (2017-2021),
https://www.nesdc.go.th/nesdb_en/ewt_dl_link.php?nid=4345
1749
Digital Government Development Agency, How Government is Transforming with
AI.
https://www.dga.or.th/upload/download/file_310433b825a546dcfd59203b423ca175.pdf
1750
Digital Government Development Agency, DGA Core Value,
https://www.dga.or.th/en/profile/2110/
1746
434
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
agencies, private firms, regulatory bodies, researchers, designers,
developers, and users. The draft guidelines cover six aspects of
development: competitiveness and sustainable development; legal
regulations and international ethical standards; operational codes and
duties; security and privacy; equality, diversity, and fairness; and
credibility.1751 As of the time of writing of this report, the guidelines are still
in draft format with no published date to launch.
The Office of National Higher Education Science Research and
Innovation Policy Council (NXPO) established five technical working
groups to explore ethical issues in genetic engineering and technology;
artificial intelligence, robotics and big data; climate change and
environment; research ethics; and communications and youth engagement
in science and technology policy development.1752 One of the recent
outcomes is Bangkok Statement on the Ethics of Science and Technology
and Sustainable Development1753, a statement calling for concerted effort of
all stakeholders to take actions on ethics of science and technology so that
new technologies can be progressed and fully developed to benefit
mankind.
AI System for Surveillance
The AI System for Surveillance and Criminal Analysis in Public is
piloted in east Bangkok. The pilot project links with security cameras at
crime hotspots under Huai Kwang police jurisdiction. Its facial recognition
compares faces against photos in a database of arrest warrants, while its
behavior analysis aims to prevent petty crime.1754 Thai government positions
the pilot program as a public safety tool under its wider Thailand 4.0 and
Smart Cities initiatives. The government also plans to build five Smart
Cities within 3 years.
Anti-fake News Centre
In November 2019, Thailand launched its “Anti-fake News Centre.”
Minister of Digital Economy and Society defines as “fake news” as any
viral online content that misleads people or damages the country’s image.
The Centre is staffed by around 30 officers at a time, who review online
National News Bureau of Thailand, Digital Ministry outlines AI ethics (Oct. 24,
2019), https://thainews.prd.go.th/en/news/detail/TCATG191024113200588
1752
NXPO, Sandbox Act and Guidelines (2020), https://www.nxpo.or.th/th/en/sandboxact-and-guideline/
1753
Bangkok Statement on the Ethics of Science and Technology and Sustainable
Development, http://www.stethicsconference2019.net/bkk-statements
1754
The Nation Thailand, Bangkok police to pilot AI surveillance system (July 25, 2019),
https://www.nationthailand.com/news/30373672
1751
435
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
content, gathered through “social listening” tools. Coupled with a law
prohibiting criticism of the monarchy, the Centre allows the government to
potentially censor or suppress any news it finds broadly affecting “peace
and order, good morals, and national security”1755 without the need for
evidence.
Digital ID
Thailand is currently working on legislation that would replace
physical ID cards with the Digi-ID which will be the backbone of the ecommerce transactions in the country. It is planned to use blockchain to
securely exchange user’s data but also require and facial recognition
verification in an effort towards a “self-sovereign” digital identity
management system.1756 It remains unclear how the government conducted
the risk or impact assessment on the mandatory use of biometric data.
Public Participation
Under the Digital Government Plan (2017-21), “Creating
Government Data that Easily Accessed and Improve Government
Transparency and Public Participation” is defined as one of the four
goals.1757 Two of the indicators that plan commits to measure itself against
are “Promoting Country’s Open Data Index to place in the World’s top 25”
and “Creating e-Gov Act”. However, all the plans and majority of initiatives
relating to AI have been developed by the central government rather than
any meaningful public participation.
Fundamental Rights and OECD/G20 AI Principles
Thailand has experienced 19 constitutional changes in less than a
century. The government acknowledges that Thailand is both a destination
and transit country for human trafficking linked to illegal immigration, child
labor and prostitution.1758 However, most of the plans put the responsibility
of respecting human rights on the citizens and not the government. The
ongoing 2020 protests are citizen criticizing the government and requesting
reform of monarchy. In response, the Thai government has extended
Reuters, Thailand unveils 'anti-fake news' centre to police the internet (Nov. 1, 2019),
https://fr.reuters.com/article/governmentFilingsNews/idUKL3N27G4KR
1756
Thailand, Digital Identity for All, https://www.ndid.co.th/
1757
Thailand Digital Government Development Plan 2017-2021,
http://jfcct.bypronto.com/wp-content/uploads/sites/1871/2018/05/Digital-GovernmentDevelopment-Plan-2017-2021-executive-version.pdf
1758
Office of the National Economic and Social Development Board Office of the Prime
Minister, Thailand, The Twelfth National Economic and Social Development Plan (20172021), https://www.nesdc.go.th/nesdb_en/ewt_dl_link.php?nid=4345
1755
436
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
emergency powers of the existing emergency decree. Under the 2017
Constitution, members of the NCPO are protected from prosecution for
human rights violations committed during NCPO rule.1759 This protection is
concerning given the reports of torture, extrajudicial executions and
enforced disappearances against, human rights defenders.
The recent Thailand Cybersecurity Act gives the government the
authority to monitor and seize data and equipment without a court order in
the name of cybersecurity risk and denies anyone targeted by the law in the
cases of a crisis or critical threat the right to any appeal.1760 Despite
continuous promises of reform, Thai authorities continue to suppress and
prosecute citizens criticizing the monarchy or the military. The combination
of AI policing, Fake-News monitoring and Cybersecurity Act creates
further concerns on fundamental rights. A group of international public and
private experts and NGOs launched the study Thailand’s Cybersecurity Act:
Towards a Human-Centered Act Protecting Online Freedom and Privacy,
While Tackling Cyber Threat to help build the discourse on the necessity of
applying a human-rights based approach to cybersecurity legislation.1761
Freedom House gives Thailand low marks for political and civil
liberties (32/100).1762 Freedom House noted that in 2019, “Thailand’s status
improved from Not Free to Partly Free due to a slight reduction in
restrictions on assembly and tightly controlled elections that, despite
significant flaws, ended a period of direct rule by military commanders.”
Thailand passed Gender Equality Act in 2015. However, the legislation still
allows for exceptions to gender discrimination on grounds of religion and
national security.1763
Thailand has endorsed Universal Declaration of Human Rights in
1948. Thailand has not signed the OECD AI Principles, Universal
Guidelines for AI. Thailand has endorsed the UNESCO Recommendation
on AI Ethics. Thailand is also the only country from Southeast Asia to
Constitute, Thailand’s Constitution of 2017,
https://www.constituteproject.org/constitution/Thailand_2017.pdf?lang=en
1760
Voice of America, Rights Groups Urge Thai Government to Curb Powers in New
Cybersecurity Act (Sept. 24, 2019), https://www.voanews.com/east-asia-pacific/rightsgroups-urge-thai-government-curb-powers-new-cybersecurity-act
1761
Manushya Foundation, Thailand’ Cybersecurity Act: Toward a Human Centered Act
Protecting Online Freedom and Privacy, While Tackling Cyber Treats (Sept. 2019),
https://www.manushyafoundation.org/study-on-cybersecurity-act
1762
Freedom House, Global Freedom Scores: Thailand.
https://freedomhouse.org/country/thailand/freedom-world/2021
1763
United National, Human Rights Treaties,
https://tbinternet.ohchr.org/_layouts/15/treatybodyexternal/Download.aspx?symbolno=C
CPR/C/THA/CO/2&Lang=En
1759
437
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
benefit from an OECD Country Programme which comprises 15 projects
drawing from four key strategic pillars: good governance and transparency,
business climate and competitiveness, “Thailand 4.0” and inclusive growth.
It includes peer reviews, capacity-building activities, inclusion in the
OECD’s statistical tools, participation in eight OECD Committees or their
subsidiary bodies and adherence to nine OECD legal instruments.1764
Data Protection
Thailand’s Personal Data Protection Act (PDPA) is the country’s
first consolidated law on data protection, framing the collection, use, and
disclosure of personal data, drawing key concepts and principles from the
EU General Data Protection Regulation (GDPR), and establishing a
Personal Data Protection Committee.1765 The Act came into force in part on
May 28, 2019 but two successive grace periods, the last one from July 2020,
postponed full implementation to give time to a broad range of government
agencies and businesses time to prepare for compliance.1766
On July 17, 2020, the Thai government issued an interim
Notification of Standards for Maintenance of Security of Personal Data to
act as a stop-gap to ensure that personal data is protected until the deferred
provisions of the PDPA become effective in 2021 and compliance with the
PDPA becomes mandatory.1767 Under the Notification, certain data
controllers must immediately implement basic security controls and
measures, including, among others, administrative, technical and physical
safeguards for personal data security and staff training and awareness.
The Digital Government Act requires establishment of government
data exchange platform to establish secure standards for transfer of data.
Agencies are mandated to keep data open to the public. However, the actual
availability of the open data through this platform across all agencies is not
complete.
The Sandbox Act provides an experimental environment set by
Ministry of Digital Economy and Society (MDES) to test technologies
OECD, A Solid Partnership between Thailand and the OECD,
http://www.oecd.org/southeast-asia/countries/thailand/
1765
Government Gazette, Personal Data Protection Act B.E. 2562 (2019) (May 27, 2019)
(unofficial translation), https://thainetizen.org/wp-content/uploads/2019/11/thailandpersonal-data-protection-act-2019-en.pdf
1766
OneTrust Data Governance, Thailand: Government approves Royal Decree
postponing PDPA (May 20, 2020), https://www.dataguidance.com/news/thailandgovernment-approves-royal-decree-postponing-pdpa
1767
OneTrust, Thailand-Data Protection Overview (Nov. 2020),
https://www.huntonprivacyblog.com/2020/05/29/delayed-implementation-of-thailandspersonal-data-protection-act/
1764
438
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
under the strict supervision of the regulators to safeguard public safety and
privacy, without being required to abide by full regulatory requirements.1768
Lethal Autonomous Weapons
Thailand expressed concern at the “wide and understudied
implications” of lethal autonomous weapons systems and affirmed “the
importance of respecting and evolving international humanitarian law.” It
has not commented on calls to ban such weapons and retain meaningful
human control over the use of force. Thailand is not a Convention on
Conventional Weapons (CCW) state party.1769
Evaluation
Thailand has launched an ambitious plan for AI. The Bangkok
Statement on Ethics is significant as is the work of the NXPO in the field of
AI ethics. But the absence of protections for fundamental rights as the
country seeks to expand national identification and systems for facial
recognition is troubling. Legal safeguards should precede AI deployment to
ensure trustworthy AI. Thailand should ensure that Personal Data Protection
Act goes into effect with an independent data protection authority.
UN High Contracting Parties and Signatories CCW:
https://www.un.org/disarmament/the-convention-on-certain-conventional-weapons/highcontracting-parties-and-signatories-ccw/
1769
439
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Turkey
National AI Strategy
In 2020, the Digital Transformation Office (DTO) of the Turkish
government outlined a National Artificial Intelligence Strategy (NAIS).1770
The goal is to “promote the effective use of big data and artificial
intelligence in public sector, through a human-centered approach . . . in
collaboration with universities, the private sector and NGOs. Further, the
strategy is meant to address fundamental principles such as human-centered
development, fairness, transparency, trustworthiness, accountability, and
commitment to ethical values. An intended output of the Strategy is also to
increase nationwide awareness on data sharing and privacy and AI
applications. Finally, the Strategy will contribute to implementing the G20
AI Principles, especially on human-centered values and fairness.
In August 2021, the Turkish government published the National
Artificial Intelligence Strategy 2021-2025.1771 The NAIS was prepared in
line with the Digital Türkiye vision and the National Technology Initiative.
The NAIS Strategy is based on six strategic priorities:
1) Training AI Experts and Increasing Employment in the Domain,
2) Supporting Research, Entrepreneurship and Innovation,
3) Facilitating Access to Quality Data and Technical Infrastructure,
4) Regulating to Accelerate Socioeconomic Adaptation,
5) Strengthening International Cooperation, and
6) Accelerating Structural and Labor Transformation.
Further, the National Strategy sets out the following values to guide
implementation,
• Respect for Human Rights, Democracy and Rule of Law
• Flourishing the Environment and Biological Ecosystem
• Ensuring Diversity and Inclusiveness, as well as international
human rights law, standards, and principles
• Living in Peaceful, Just and Interconnected Societies,
The Digital Transformation Office will pursue several key projects
including, Federated Learning and Differential Privacy technologies “with
the purpose of ensuring the privacy and security of data,” making Black
Box algorithms explainable, and preventing misleading artificial
intelligence algorithms.
Presidency of the Republic of Turkey, Digital Transformation Office, Artificial
Intelligence, https://cbddo.gov.tr/en/artificial-intelligence
1771
Presidency of The Republic of Turkey, National Artificial Intelligence Strategy 20212025 (Aug. 20, 2021). https://cbddo.gov.tr/en/nais
1770
440
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The National AI Strategy follows a February 2020 workshop on AI,
organized by the Ministry of Industry and Technology, and TÜBİTAK (The
Scientific and Technological Research Council of Turkey).1772 At that
meeting, Dr. Ali Taha Koç, the DTO Director, emphasized the need to
“develop more reliable artificial intelligence systems.” He also said that “an
artificial intelligence ecosystem that is not based on ethical principles will
not be successful or sustainable.” He listed several principles that should
govern the use of artificial intelligence to alleviate privacy concerns,
including focused on people, fairness, aiming to make a positive
contribution to social welfare, transparent, reliable, accountable, able to
derive value from data, in line with national and ethical values.
In October 2020, at the Turkish-Hungarian Artificial Intelligence
and High-Technology Conference, the Minister of Industry and Technology
Mustafa Varank said that the National AI Strategy makes “special emphasis
on the most important aspects of AI policies such as talent development,
scientific research, ethics and inclusion and digital infrastructure."1773
Public Participation
In the development of the NAIS, interviews were conducted with
public institutions, universities, private sector organizations, NGOs, and
international organizations on physical and online platforms, and domain
experts were asked to provide their evaluations. January 2020, to develop
the AI Technology Roadmap, Turkey established a stakeholder Working
Group, comprised of academia, private sector and major umbrella
NGOs.1774 The Working Group operates under the Science, Technology and
Innovation Policy Council of the Turkish Presidency, via the technical
contribution of TUBITAK. The Working Group will identify frontier
scientific themes and priority sectoral applications of AI technologies. The
working group will help ensure effective intergovernmental coordination.
Presidency of the Republic of Turkey, Digital Transformation Office, Workshops on
National AI Strategy and AI Institute Organized (Feb. 19, 2020),
https://cbddo.gov.tr/en/news/4701/ulusal-yapay-zeka-stratejisi-ve-yapay-zeka-enstitusucalistaylari-duzenlendi
1773
Daily News, Turkey to reveal artificial intelligence strategy (Oct. 16, 2020),
https://www.hurriyetdailynews.com/turkey-to-reveal-artificial-intelligence-strategy159189
1774
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1772
441
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
In 2019, Turkey hosted the proceedings of the 2019 3rd International
Conference on Advances in Artificial Intelligence1775 and have since
contributed to multiple AI related conferences.
Turkey’s Industry and Technology Strategy 2023 includes sectoral
and R&D competency mapping on AI technology and AI and machine
learning, with a view to strengthening Turkey’s capacity of scientific
research and product development. In 2019 Turkey’s Ministry of Industry
and Technology published the 2023 Industry and Technology Strategy,
taking a holistic approach to the fields of industry and technology, and
aiming to ensure wide participation and to mobilize society.1776 The
Strategy consists of five main pillars: High Technology and Innovation,
Digital Transformation and Industry Move, Entrepreneurship, Human
Capital, and “Infrastructure.
Specific to AI, the Strategy will see preparation of an R&D
competency map to analyze the present state of in AI technology (in
particular AI and machine learning) in Turkey, notably the capacity of
scientific research and product development. Majority of Turkey’s
international R&D collaborations are covered under EU Horizon 2020 and
2021-2027 Programs and Digital Europe Program which necessitates
certain harmonization with EU legislation and requirements. The NAIS
proposes the establishment of sectoral Co-Creation Laboratories” within the
TÜBİTAK Artificial Intelligence Institute for multi-stakeholder
development and testing of sectoral AI applications consisting of productoriented targets in areas such as software, aviation and space.1777 The
Institute held a stakeholder workshop in February 2020.
Public participation is still being ensured after the entry into force
of the Turkish NAIS. As per the Strategy, a Steering Committee was
established to carry out the implementation process, which will be chaired
by the Vice President. The governance mechanism embraces AI Ecosystem
Advisory Group and working groups as well, where all relevant
stakeholders will be represented. Action plans, in which the implementation
details of the Turkish NAIS will be laid out, are being prepared by the
relevant Ministries under the coordination of the Steering Committee.
1775
ACM, ICAAI 2019: Proceedings of the 2019 3rd International Conference on
Advances in Artificial Intelligence (2019),
https://dl.acm.org/doi/proceedings/10.1145/3369114
1776
Turkey’s Ministry of Industry and Technology, 2023 Sanayi ve Teknoloji Stratejisi
(Sept. 18, 2019), https://www.sanayi.gov.tr/strateji2023/sts-ktp.pdf
1777
Daily News, Turkey to reveal artificial intelligence strategy (Oct. 16, 2020),
https://www.hurriyetdailynews.com/turkey-to-reveal-artificial-intelligence-strategy159189
442
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The Turkish NAIS is accessible by the public through the official
website of the Digital Transformation Office of the Presidency of Türkiye,
both in English and Turkish.1778
Open Data Project
Turkey’s Open Data Project will establish an open data portal so
citizens, researchers, public institutions and organizations, and state
affiliates can “leverage data produced by public resources.”1779 As a
platform for the datasets needed for developing AI technologies and
applications, the initiative will contribute to fostering a digital ecosystem
for AI.
The Open Data Project will be launched under the responsibility of
Turkey’s Presidency of Digital Transformation Office. The main focus is to
establish an open data portal, but the project will also manage the regulatory
and legislative steps for participation in the Open Government Partnership.
Organisations seeking to take part in the Open Government Partnership will
need to meet certain regulatory and legislative requirements. The initiative
also aims to provide effective coordination in preparing the labor market for
digital transformation. The project also proposes creation of National Data
Dictionary and data sharing via Open Government Data Portal.1780
Data Protection
The Law on the Protection of Personal Data was published in April
2016.1781 The law established the Turkish Data Protection Authority
(KKVK), an independent regulatory authority. The Authority is composed
of the Personal Data Protection Board and the Presidency. The mission of
the Authority is to provide the protection of personal data and to develop
public awareness in line with the fundamental rights related to privacy and
freedom stated in the Constitution. Turkey ratified Convention 108 of the
Council of Europe and its data protection law originates from European
Text of the Turkish NAIS, https://cbddo.gov.tr/SharedFolderServer/Genel/File/TRUlusalYZStratejisi2021- 2025.pdf; English translation of the Turkish NAIS,
https://cbddo.gov.tr/SharedFolderServer/Genel/File/TRNationalAIStrategy20212025.pdf; Presidential Circular No. 2021/18,
https://www.resmigazete.gov.tr/eskiler/2021/08/20210820-22.pdf
1779
British Embassy Ankara, Open Data in Turkey (March 2020),
http://www.novusens.com/s/2462/i/UK-Turkey_Open_Data_Writeup_ENG.pdf
1780
Presidency of the Republic of Turkey, Digital Transformation Office, Open Data:
Generating Value from Data for Our Country, https://cbddo.gov.tr/en/opendata/aboutthe-project/
1781
KVKK, Data Protection in Turkey,
https://www.kvkk.gov.tr/SharedFolderServer/CMSFiles/5c02cb3c-7cc0-4fb0-b0a785cb90899df8.pdf
1778
443
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Union Directive 95/46/EC. According to the KVKK, the Protection of
Personal Data law ensures:
• That data is Processed lawfully and fairly; Accurate and where
necessary, kept up to date; Processed for specified, explicit and
legitimate purposes; Relevant, limited and proportionate to the
purposes for which they are processed; and Retained for the period
of time determined by the relevant legislation or the period deemed
necessary for the purpose of the processing.
• That explicit consent is required by an individual for data
collection and data transfer. Further, data transfer outside of
Turkey is strictly regulated.
• That individuals have the right to access and complain regarding
data collection.
• That data collectors know to what extent they need to protect data
and regulations for responses to individual complaints.
The Personal Data Protection Board will implement and interpret the
privacy law in line with GDPR. Currently, efforts are also underway to
adopt a new privacy law, based on the GDPR.1782
There are two other institutions that could provide independent
oversight of AI practices. The Human Rights and Equality Institution of
Türkiye, affiliated with the Ministry of Justice, was established by Law No.
6701 in 2016.1783 The Human Rights Institution has public legal entity status
and administrative and financial autonomy. The Human Rights Institution
was established on the basis of the principle of human dignity and has
authority for the protection and promotion of human rights; guaranteeing
individuals’ right to equal treatment, prevention of discrimination in the
exercise of rights and freedoms; and opposing torture and ill-treatment.
The Ombudsman Institution of the Republic of Türkiye was
established in 2012 with the Law on the Ombudsman Institution No. 6328
as a constitutional public entity affiliated with the Grand National Assembly
of Turkey.1784 The Ombudsman Institution has its own private budget and
headquarters in Ankara and one office in Istanbul. According to the
Turkish Ministry of Justice, Action Plan on Human Rights: Free Individual, Strong
Society; More Democratic TURKEY (March 2021) (“6.7 The Law on Protection of
Personal Data will be harmonized with the European Union standards.”),
https://inhak.adalet.gov.tr/Resimler/SayfaDokuman/5320211949561614962441580_insa
n- haklari-EP-v2_eng.pdf.
1783
The Human Rights and Equality Institution of Türkiye, https://www.tihek.gov.tr/en
1784
Grand National Assembly of Turkey, Ombudsman Institution,
https://www.ombudsman.gov.tr/English/index
1782
444
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Ombudsman Law, the Institution shall be responsible for examining,
investigating, and submitting recommendations to the Administration with
regard to all sorts of acts and actions as well as attitudes and behaviors of
the Administration upon complaint on the functioning of the Administration
within the framework of an understanding of human rights-based justice and
in the aspect of legality and conformity with principles of fairness. Among
its functions, the Ombudsman Institution aims to increase the service
quality of the administration, internalize principles of good administration,
improve human rights standards, strengthen the culture of seeking legal
remedies, and form a transparent and accountable administration.
OECD AI Principles
Turkey is signatory to the OECD AI Principles. According to the
OECD, Turkey’s AI Technology Roadmap is a multistakeholder effort that
supports implementation of the G20 AI Principles on inclusive growth,
robustness and accountability.1785 The NAIS explicitly highlights humancentric AI principles adopted by the OECD, G20, EU and UNESCO.1786
The NAIS also lists several priority AI Principles, including
Proportionality, Safety and Security, Fairness, Privacy, Transparency and
Explainability, Responsibility and Accountability, Data Sovereignty, and
Multi-stakeholder governance. The NAIS also refers to several platforms
order to operationalize these principles within public and private
institutions, including the AI Maturity Model and the AI Project
Management Guide. A Public AI Platform will facilitate the preparation
process of AI systems. The Trustworthy AI Seal approach will encourage
the use of reference models in application development and operation.
Universal Guidelines for AI
Turkey has actively participated in the Council of Europe Ad Hoc
Committee on Artificial Intelligence (CAHAI) and served as a coordinator,
as well as a lead drafter in a working group for “Red-Lines” under the Legal
Frameworks working group. Turkey actively championed the restriction of
certain uses of AI that would be against the spirit of fundamental rights
during the discussions. This position reflected the Turkish NAIS which
emphasized that, "The scope of lifestyle choices, beliefs, ideas, expressions,
or personal experiences, including the discretionary use and design of AI
systems, should in no way be restricted at any stage of the lifecycle of AI
systems. The production, development and implementation of AI
OECD G20 Digital Economy Task Force, Examples of AI National Policies (2020),
https://www.mcit.gov.sa/sites/default/files/examples-of-ai-national-policies.pdf
1786
Presidency of The Republic of Turkey (Aug. 20, 2021)
1785
445
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
technologies should not result in discrimination in any way, and datasets
should be audited in this regard.”1787
Human Rights
According to Freedom House, Turkey is “not free.1788 Turkey
receives low scores for political rights and civil liberties (32/100).
Regarding transparency, Freedom House reports, “Although Turkey has an
access to information law on the books, in practice the government lacks
transparency and arbitrarily withholds information on the activities of state
officials and institutions.” In October 2020, a law entered into force
requiring all domestic and foreign social network providers that serve more
than one million users in Turkey to have local representation in Turkey. The
law gives authority to Information and Communication Technologies
Authority to order the provider the removal or blocking of content within
48 hours if related to violation of personality and privacy rights, related to
listed crimes, threats to public order or security, or other pressing and
immediate dangers.1789
Algorithmic Transparency
The Turkish Constitution establishes rights for privacy and for data
protection.1790 Further, using personal data in algorithms is considered
“processing of personal data.” Processing of personal data is defined under
as “any operation which is performed on personal data, wholly or partially
by automated means or non-automated means which provided that form part
of a data filing system, such as collection, recording, storage, protection,
alteration, adaptation, disclosure, transfer, retrieval, making available for
collection, categorization, preventing the use thereof.”1791
Dr. Ali Taha Koç, president of DTO, also acknowledged the
importance of transparency, security, and accountability for AI. He said,
“To eliminate the concerns that may arise, this system must first be humancentered, it must be fair, it must increase social welfare, it must be
transparent, reliable, accountable, value-based, and dependent on national
and ethical values. In our AI strategy, which should focus on human and
ethical values, we have obligations such as creating a sustainable and
NAIS at 65.
Freedom House, Freedom in the World 2021 – Turkey (2021),
https://freedomhouse.org/country/turkey/freedom-world/2021
1789
Law No:5651 on the Regulation of Publications on the Internet and Combatting
Crimes Committed by Means of Such Publications, https://perma.cc/T97C-AM9H
1790
Article 20.
1791
Law No. 6698, Article 3/1€.
1787
1788
446
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
production-based environment in Turkey by building an AI ecosystem,
paving the way for work on AI in our country by completing the framework
of data access, sharing and increasing the efficiency of all businesses and
business processes in the public sector by expanding the use and application
of AI technologies, sustaining this AI ecosystem by bringing up and
educating qualified manpower, increasing the human benefit of each AI
system to be produced, and ensuring its well-being."1792 The National AI
Strategy states that “legal regulations that directly concern individual rights
and freedoms should be made in a way that sets an example to the universal
legal system.” However same document also suggests AI for predictive
policing would be permissible.
Lethal Autonomous Weapons
Turkey has participated in every Convention on Certain
Conventional Weapons (CCW) meeting on lethal autonomous weapons
systems between 2014 and 2019. Turkey expressed concern “at the
increasing impact of such [lethal autonomous] weapons worldwide,
especially through the perpetration of terrorist acts” and importance to the
humanitarian aspect of the matter and supporting “need for human control
and accountability for such weapon systems” nevertheless also stating
“taking into consideration that yet such weapon systems do not exist and it
is an issue which is still hypothetical, we hesitate on the accuracy of a
general prohibition pre-emptively.”1793 Most recently, a United Nations
Security Council report suggests that the attack drone Kargu-2, made by a
Turkish company, was used in Libya for autonomous attacks on human
targets.1794
The NAIS states that “legal regulations that directly concern
individual rights and freedoms should be made in a way that sets an example
to the universal legal system.”1795 However same document also suggests in
addition to real-time event support, law enforcement agencies are using AI
Şule Guner, Experts to map out Turkey's strategy on AI centered on ethics and data
protection, Daily Sabah (Mar. 4, 2020), https://www.dailysabah.com/life/experts-to-mapout-turkeys-strategy-on-ai-centered-on-ethics-and-data-protection/news
1793
Human Rights Watch, Stopping Killer Robots: Country Positions on Banning Fully
Autonomous Weapons and Retaining Human Control – Turkey (Aug. 10, 2020),
https://www.hrw.org/report/2020/08/10/stopping-killer-robots/country-positions-banningfully-autonomous-weapons-and
1794
Maria Cramer, A.I. Drone May Have Acted on Its Own in Attacking Fighters, U.N.
Says: A United Nations report suggested that a drone, used against militia fighters in
Libya’s civil war, may have selected a target autonomously, New York Times (June 4,
2021), https://www.nytimes.com/2021/06/03/world/africa/libya-drone.html
1795
NAIS at 50.
1792
447
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
applications in order to predict future events and risks.1796 In January 2022,
the Ministry of Defense announced several R&D projects including those
involving machine learning capabilities, noting that Turkey “has to establish
its technological independence.”1797
Evaluation
Turkey is an emerging market for AI, and a regional leader in AI.
The National Assembly of Turkey has proposed the creation of roadmaps
and R&D in different technology sectors, particularly AI. There have been
multiple statements by Turkish officials regarding their 2020 AI Strategy
and commitment to human-centered development. Overall, despite
investment, participation in AI related conferences, and proposed plans and
sector roadmaps, there has been little policy action and most directives,
particularly the AI Strategy, are still in “planning” phases with no official
publications to date. The only official policy related to AI to date is The
Law on the Protection of Personal Data.
NAIS at 54.
1797 Demiroren Haber Ajansi (Jan 27, 2022), İsmail Demir: Teknolojik bağımsızlığımızı
sağlamak mecburiyetindeyiz. https://www.dha.com.tr/gundem/ismail-demir-teknolojikbagimsizligimizi-saglamak-mecburiyetindeyiz-2014833
1796
448
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
United Arab Emirate
National AI Strategy
In 2017, the UAE became the first country to appoint a Minister of
State for Artificial Intelligence.1798 The responsibilities include enhancing
the government performance levels by investing in the latest technologies
of artificial intelligence and applying them in various sectors. The United
Arab Emirates also created the UAE Council for Artificial Intelligence and
Blockchain to facilitate the government’s implementation of AI policies.1799
In October 2017, the UAE Government launched the UAE Strategy
for Artificial Intelligence (AI).1800 The strategy aligns with UAE Centennial
2071,1801 which has a truly ambitious goal to make the UAE the best country
in the world by 2071. AI will play a significant role in education, economy,
government development, and community happiness through various AI
implementations in various sectors, including energy, tourism, and
education to name a few. AI Strategy outlines eight strategic objectives,
namely:
1) “Build a reputation as an AI destination.
2) Increase the UAE competitive assets in priority sectors
through deployment of AI.
3) Develop a fertile ecosystem for AI.
4) Adopt AI across customer services to improve lives and
government.
5) Attract and train talent for future jobs enabled by AI.
6) Bring world-leading research capability to work with target
industries.
7) Provide the data and supporting infrastructure essential to
become a test bed for AI.
8) Ensure strong governance and effective regulation.” 1802
UAE Artificial Intelligence Office, Omar Sultan Al Olama has been appointed as
Minister of State for Artificial Intelligence (Oct. 20, 2017), https://ai.gov.ae/about/
1799
UAE Artificial Intelligence Office, UAE adopts formation of Council for Artificial
Intelligence (Mar. 5, 2018), https://ai.gov.ae/ai_council/
1800
UAE Strategy for Artificial Intelligence - The Official Portal Of the UAE Government
(Oct. 12, 2021), https://u.ae/en/about-the-uae/strategies-initiatives-and-awards/federalgovernments-strategies-and-plans/uae-strategy-for-artificial-intelligence
1801
UAE Centennial 2071, https://u.ae/en/about-the-uae/strategies-initiatives-andawards/federal-governments-strategies-and-plans/uae-centennial-2071
1802
Artificial Intelligence Office, UAE National Strategy for AI 2031,
https://ai.gov.ae/strategy/
1798
449
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The UAE was among the UNESCO member states adopted the firstever global agreement on Artificial Intelligence Ethics.1803 By addressing
issues of transparency, accountability and privacy, the pact promotes human
rights and contributes to the accomplishment of the Sustainable
Development Goals, encompassing chapters on data governance, education,
culture, employment, health, and the economy.
The UAE also has a multi-national strategy with India which was
birthed by signing a Memorandum of Understanding (MoU) for India UAE Artificial Intelligence Bridge in 2018. The UAE-India collaboration
seeks to evaluate the dynamic nature of innovation and technology by
convening a UAE-India AI Working Group (TWG) between the UAE
Ministry for Artificial Intelligence, Invest India and Start-up India.1804
OECD/G20 Principles
The Dubai AI Ethics Guidelines describe the key principles of a fair,
transparent, accountable, and explainable AI system. 1805 Launched in
January 2019, the AI Principles and Guidelines for the Emirate of Dubai
demonstrate Dubai’s broader approach to ethical AI. Accompanying the
Principles and Guidelines is an Ethical AI Self-Assessment Tool built to
enable AI developer or operator organisations to evaluate the ethics level of
an AI system. The AI Ethics Guidelines provide an assessment (from proof
of concept to production) of the ethical issues that may arise throughout the
development process and how specific AI applications could be improved
to ensure fairness, transparency, accountability and explainability. The tool
also aims to ensure careful adoption of AI in which innovation potential is
optimised and where economic and social value is captured. Below are the
OECD AI principles outlined in these guidelines;
• “Inclusive growth, sustainable development and well-being
• Human-centred values and fairness
• Transparency and explainability
• Robustness, security and safety
• Accountability
UNESCO, UNESCO Member States Adopt the First Ever Global Agreement On the
Ethics Of Artificial Intelligence (Nov. 25, 2021), https://en.unesco.org/news/unescomember-states-adopt-first-ever-global-agreement-ethics-artificial-intelligence.
1804
AI Ethicist, National AI Strategies, https://www.aiethicist.org/national-strategies
1805
Digital Dubai, AI Principles & Ethics, https://www.digitaldubai.ae/initiatives/aiprinciples-ethics
1803
450
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
• Providing an enabling policy environment for AI” 1806
However, the UAE is not a member of the OECD and has not formally
endorsed the OECD AI Principles. Dubai established a governing AI ethics
board, comprised of government policy, academic, legal and industry
experts to oversee and guide the strategic development of the AI Ethics
Guidelines. The Executive Council of Dubai has directed government
entities to use the principles and guidelines when considering AI
development, and entities including the Road and Transport Authority and
the Dubai Police have formally acknowledged their adoption of the selfassessment tool when developing AI.1807
The UAE Cabinet formed the UAE Council for Artificial
Intelligence (AI) that will oversee AI integration in government
departments and the education sector. The council is tasked with proposing
policies to create an AI-friendly ecosystem, encourage advanced research
in the sector and promote collaboration between the public and private
sectors, including international institutions to accelerate the adoption of
AI.1808
Public Participation
The government has set up The UAE National Program for Artificial
Intelligence1809 which is a comprehensive and consolidated compilation of
resources that highlight the advances in AI and Robotics. The UAE, through
the Ministry of Intelligence, launched an initiative to develop \egislation,
policies and initiatives for a responsible and efficient adoption of artificial
intelligence (AI) within the private sector. The initiative is called 'Think AI'
which was established to facilitate a series of roundtables, workshops and
panel discussions to enable the participation of more than 100 government
officials, representatives from the private sector and experts from local and
international organisations. The discussions aim to develop comprehensive
dialogs and ideas that support the UAE's efforts towards accelerating the
adoption of artificial intelligence in various key sectors such as
infrastructure, governance and legal legislation, the development of
The OECD AI Policy Observatory, AI Principles and Ethics for the Emirate of Dubai
(2019), https://oecd.ai/en/dashboards/policyinitiatives/http:%2F%2Faipo.oecd.org%2F2021-data-policyInitiatives-26783
1807
Digital Dubai, AI Principles & Ethics, https://www.digitaldubai.ae/initiatives/aiprinciples-ethics
1808
UAE’s Government portal, https://u.ae/en/about-the-uae/digital-uae/artificialintelligence-in-government-policies
1809
UAE, Empowering the Next Generation of Coders, https://ai.gov.ae/
1806
451
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
appropriate infrastructure, and the strengthening the countries position as a
global hub for artificial intelligence.1810
Free courses are being run for UAE residents to raise awareness and
understanding of AI technologies.1811 Additionally, the government has
endeavoured to upskill the student population and government employees
by providing relevant trainings to them. 1812
Further, the UAE has a digital participation policy to provide insight
to participants on what can be posted on the public platform and to facilitate
a healthy environment for sharing information and enabling meaningful
discussion on topics concerning the services of the UAE Digital
Government and life in the UAE in general. 1813Dubai has an AI principles
and ethics URL available for public access. Dubai’s Ethical AI Toolkit is
particularly helpful for three main types of user: Government Entities,
Private Sector Entities and Individuals. 1814
Data Protection
In November 2021, the UAE adopted sweeping legislative reforms,
including the Personal Data Protection Law, modelled after the GDPR. The
Personal Data Protection Law constitutes an integrated framework to ensure
the confidentiality of information and protect the privacy of community
members by providing proper governance for optimal data management and
protection, in addition to defining the rights and duties of all concerned
parties.1815 The provisions of the law apply to the processing of personal
data, whether all or part of it through electronic systems, inside or outside
the country. The law prohibits the processing of personal data without the
consent of its owner, with the exception of some cases in which the
processing is necessary to protect the public interest, or that the processing
is related to the personal data that has become available and known to all by
an act of the data owner, or that the processing is necessary to carry out any
of the legal procedures and rights. The law defines the controls for the
processing of personal data and the general obligations of companies that
have personal data and defines their obligations to secure personal data and
maintain its confidentiality and privacy. It also defines the rights and cases
‘UAE Government Launches “Think AI” Initiative’ (wam),
http://wam.ae/en/details/1395302745072
1811
‘UAE Artificial Intelligence Office, AI Summer Camp 4.0, https://ai.gov.ae/camp/
1812
UAE Artificial Intelligence Office, Learn AI https://ai.gov.ae/learn/
1813
UAE, Digital participation policy, https://u.ae/en/footer/digital-participation-policy
1814
Digital Dubai, AI Principles and Ethics, https://www.digitaldubai.ae/initiatives/aiprinciples-ethics
1815
Emirates News Agency, UAE adopts largest legislative reform in its history (Nov.
27, 2021), https://www.wam.ae/en/details/1395302997239
1810
452
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
in which the owner has the right to request correction of inaccurate personal
data, restrict or stop the processing of personal data. The law sets out the
requirements for the cross-border transfer and sharing of personal data for
processing purposes.
However, the law does not apply to government data, government
authorities that control or process personal data, or personal data processed
by the security and judicial authorities.1816 The law also does not cover
processing of health, banking, and credit data which is subject to sectorspecific legislation and companies and institutions located in free zones
which have specific data protection laws, such as the Dubai International
Finance Centre (DIFC) and the Abu Dhabi Global Market (ADGM).
Commentators have also noted significant divergences with the GDPR,
including more limited legal basis, with a focus on consent as the primary
legal basis, and less comprehensive transparency requirements.1817
Less onerous transparency requirements (only certain limited
information will be required to be provided prior to processing) and no
specific privacy notice requirement
The UAE has also established the UAE Data Office that aims at
ensuring the full protection of personal data. The office, which will be
affiliated with the Cabinet, is responsible for a wide range of tasks that
include proposing and preparing policies and legislations related to data
protection, proposing and approving the standards for monitoring the
application of federal legislation regulating this field, preparing and
approving systems for complaints and grievances, and issuing the necessary
guidelines and instructions for the implementation of data protection
legislations.
Others UAE laws that provide general rights to privacy include:
1) The UAE Constitution addresses privacy by providing that
freedom of communication by post or other means of
communication and the secrecy thereof is guaranteed in
accordance with the law;1818
1816
Al Tamini, UAE’s New Federal Data Protection Law (Dec. 6, 2021),
https://www.tamimi.com/news/uaes-new-federal-data-protection-law/
1817
Latham and Watkins, UAE Publishes First Data Protection Law (Dec. 14, 2021),
https://www.globalprivacyblog.com/legislative-regulatory-developments/uae-publishesfirst-federal-data-protection-law/
1818
‘United_Arab_Emirates_2004.Pdf’,
https://www.constituteproject.org/constitution/United_Arab_Emirates_2004.pdf
453
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
2) The UAE Penal Code prohibits those who have access to an
individuals’ personal data from disclosing or publicizing that
information;1819 and
3) The Cyber Crimes Law (Federal Law No. 5 of 2012 relating to
Combating Information Technology Crimes, as amended by
Federal No. 12 of 2016 and Emiri Decree No. 2 of 2018)
prohibits invading the privacy of another person via
technological means, without their consent.1820
Algorithmic Transparency
The UAE’s Artificial Intelligence Guide touches upon the subject of
transparency with respect to AI and its use in companies. The Guide
recommends that companies evaluate corporate policy to ensure the right
guidelines are in place for any AI implementation to be ethical, fair,
accountable, transparent, and explainable. The intention of the
recommendation is to ensure that the AI solution is not only innovative but
also delivers human benefit and happiness.
This outlook on AI is supplemented by the Ethical AI Toolkit
published by the Dubai Data Establishment, Smart Dubai Office in 2019
which defines guiding principles for ethical AI focusing on four domains:
ethics, security, humanity, and inclusiveness. Within the purview of ethics,
the AI systems are expected to be fair, transparent, accountable, and
understandable. However, the recently adopted Personal Data Protection
Law does not appear to establish a legal right to algorithmic transparency.
Human Rights
The UAE has signed the Universal Declaration of Human Rights
(UDHR), however there have been cases of unexplained and arbitrary
arrests, enforced disappearances for months to unknown locations,
criminalisation of basic freedom of speech, torture and other ill-treatment
of prisoners resulting in false confessions, unfair trials without legal
representation and citizenship revocation and deportations without
legitimate reasons.1821
The Freedom House has given the UAE quite a low score (17/100)
with regards to political rights and civil liberties, and rated the country “Not
‘Federal Law No. 3 of 1987: The Penal Code’ (DataGuidance, 4 January 2018),
https://www.dataguidance.com/legal-research/federal-law-no-3-1987-penal-code
1820
UAE, Cyber Laws, https://u.ae/en/resources/laws
1821
ICFUAE, Human Rights in the UAE,
http://icfuae.org.uk/sites/default/files/Human%20Rights%20Violations%20Briefing_1309-2017.pdf
1819
454
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Free.” Freedom house reports that in UAE, political parties are banned, and
all executive, legislative, and judicial authority ultimately rests with the
seven hereditary rulers. The civil liberties of both citizens and noncitizens,
who make up an overwhelming majority of the population, are subject to
significant restrictions.1822
Facial Recognition
The UAE’s digital transformation has led to the adoption of an AIpowered facial recognition system. Digitisation will not only lead to growth
but criminal activities will experience a significant increase too. Retail,
finance, security, airport clearance, and public transport are some of the
areas where facial recognition will be employed. In 2021, the UAE
Government announced that it will employ biometric face recognition
(Facial ID) to register customers in the "UAE Pass" application, as part of
the plan to launch the first digital national ID for citizens and residents.
Currently, the number of people registered on the UAE Pass app is over
1.38 million, including 628,000 individuals with verified accounts.1823
Predictive Policing
In 2017 the Dubai Police released the 2018-31 strategic plan for
artificial intelligence.1824 Among other things, the plan works towards
integrating AI techniques and into criminal investigation in the field of
forensic work, and in police operations to predict crimes. AI techniques will
be used for forecasting crime, for crowd management and to enhance traffic
safety and road security. The AI techniques will also be utilised in crisis and
disaster management.1825 The UAE deploys some of the world’s most
advanced surveillance technologies to pervasively monitor public spaces,
internet activity, and even individuals’ phones and computers, in violation
of their right to privacy, freedom of expression, association, and other
rights. Using CCTV cameras, license plate detection, and facial recognition,
UAE authorities aim to keep tabs on all residents. In 2018, Dubai Police
announced an artificial intelligence surveillance program called Oyoon,
which utilizes tens of thousands of cameras with facial recognition software
Freedom House, United Arab Emirates: Freedom in the World 2021 Country Report,
https://freedomhouse.org/country/united-arab-emirates/freedom-world/2021
1823
MoCA, UAE, UAE Government to employ biometric face recognition to register
customers under 'UAE Pass' app (Apr. 7, 2021),
https://www.moca.gov.ae/en/media/news/uae-government-to-employ-biometric-facerecognition-to-register-customers-under-'uae-pass'-app
1824
OPENGOV, Dubai Police Releases 2018-31 Strategic Plan for Artificial Intelligence,
(Dec. 21, 2017), https://opengovasia.com/dubai-police-releases-2018-31-strategic-planfor-artificial-intelligence/
1825
Government of Dubai,
1822
455
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and microphones that feed back into one central command center and can
be used to track and analyze movements in key areas, and even issue verbal
warnings to those suspected of wrongdoing. 1826 Abu Dhabi has a system
similar to Oyoon called Falcon Eye. The Falcon Eye system provides a
comprehensive central monitoring system that receives live feeds of various
surveillance systems installed throughout the city, provides smart alerts, and
allows fast access to events and incidents.1827
Autonomous Weapons
The UAE leads the Gulf region in AI developments. The UAE also
wishes to advance in autonomous weapons. Given the knowledge Israel
acquired in the field, particularly in the defense sector, there is a good
chance that their AI collaboration will grow with the new diplomatic
agreement between the UAE and Israel. For example, in 2018 the Emirati
and Israeli state-owned weapons makers agreed to design unmanned vessels
capable of anti-submarine warfare together.1828
Evaluation
The UAE has a national AI strategy that is open and public
participation is encouraged. Further the country has established a body to
facilitate the government’s implementation of AI policies overseen by the
Minister of State for Artificial Intelligence. While UAE is not a member of
the OECD and has therefore not formally endorsed the OECD AI Principles,
its AI policies reflect elements found in the OECD framework. However,
UAE also has a poor track record on human rights. There is no express
support for the Universal Guidelines for AI, or a stand on Lethal
Autonomous Weapons. Although UAE is one of the leading countries in the
Gulf region with regards to AI and strives to be the world’s most prepared
country for Artificial Intelligence, there are still steps that should be taken
such as towards strengthening human rights, creation of a data protection
1826
Human Rights Watch, UAE: Tolerance Narrative a Sham: Censorship; Surveillance;
Prison or Barred Entry for Critics, (Oct. 1, 2021),
https://www.hrw.org/news/2021/10/01/uae-tolerance-narrative-sham-0
1827
Dubai Police General HQ, Dubai Police launch “Oyoon” AI Surveillance
Programme (Jan. 28, 2018),
https://www.dubaipolice.gov.ae/wps/portal/home/search/!ut/p/z1/04_Sj9CPykssy0xPLM
nMz0vMAfIjo8zi_T29HQ2NvA18LQJNTQwCPUIN_Hy8QowMTIz0w8EKDHAARwP
9KEL6o1CVuJt4OxkEuoZa-IX4BsZGBhBFeCxIjg1T78gN8Igy8RREQC7VCIU/?1dmy&urile=wcm%3apath%3a/wps/w
cm/connect/DubaiPolice_en/DubaiPolice/Media-Center/News/A70
1828
Reuters, UAE, Israel To Jointly Develop Unmanned Military, Commercial Vessels
(Nov. 18, 2021), https://www.reuters.com/world/middle-east/uae-israel-jointly-designunmanned-military-commercial-vessels-2021-11-18/
456
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
legislation and an oversight authority and rapid adoption of facial
recognition without clear legal basis to support limits on lethal autonomous
weapons.
457
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
United Kingdom
National AI Strategy
In September 2021, the UK government launched its first National
Artificial Intelligence Strategy.1829 This comes on the back of a raft of
related plans, strategies and roadmaps including the National Data Strategy
(2020),1830 a Plan for Digital Regulation (2021)1831 and the UK Innovation
Strategy (2021).1832 The UK holds a strong position in AI and is currently
ranked third in the world for private venture capital investment and is home
to a third of Europe’s total AI companies.1833 The AI Strategy sets out a tenyear plan with the vision, “to remain an AI and science superpower fit for
the next decade.” The UK AI Strategy has three main pillars: (1) investing
and planning for the long-term requirements of the UK’s AI ecosystem; (2)
supporting the transition to an AI-enabled economy across all sectors and
regions of the UK; and (3) ensuring that the UK gets the national and
international governance of AI technologies right in order to encourage
innovation, investment and protect the public and the country’s fundamental
values.
The first of these pillars focuses on the need to invest in the skills
and resources that lead to AI innovation with the aim of increasing the type,
frequency and scale of AI discoveries which are developed and exploited in
the UK. To achieve this, the UK will continue to invest in developing,
attracting and training the best people; develop a new approach to research
development and innovation; increase international collaboration and
research; and improve access to data and compute.
The second pillar aims to ensure that the benefits of AI innovation
are shared across all sectors and regions of the UK economy. Here, the UK
GOV.UK, National AI Strategy (Sept. 2021),
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/1020402/National_AI_Strategy_-_PDF_version.pdf
1830
GOV.UK, National Data Strategy (Dec. 2020),
https://www.gov.uk/government/publications/uk-national-data-strategy/national-datastrategy
1831
GOV.UK, Digital Regulation: Driving Growth and Unlocking Potential (July 2021)
https://www.gov.uk/government/publications/digital-regulation-driving-growth-andunlocking-innovation/digital-regulation-driving-growth-and-unlocking-innovationa
1832
GOV.UK, Policy Paper: UK Innovation Strategy: leading the future by creating it
(July 2021), https://www.gov.uk/government/publications/uk-innovation-strategyleading-the-future-by-creating-it
1833
GOV.UK, Policy Paper: UK Innovation Strategy: leading the future by creating it
(July 2021), https://www.gov.uk/government/publications/uk-innovation-strategyleading-the-future-by-creating-it
1829
458
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
aims to support UK AI businesses; better understand the drivers for
organizational adoption of AI; and leverage the public sector’s capacity to
create demand for AI and market for new services. This pillar follows on
from earlier industrial and digital strategies.1834
The third pillar centers on the government’s aim to build the most
trusted and pro-innovation system for AI governance in the world. This
centers on the UK having a clear, proportionate and effective framework for
regulating AI; UK regulators having the flexibility and capabilities to
respond to the challenges of AI; and ensuring organizations can confidently
innovate and adopt AI technologies with the right tools and infrastructure
to address AI risks and harms.
Overall, the strategy covers a wide range of issues and additionally
sets out a timeline summarizing key actions for the short, medium and long
term. Common themes that emerge within these goals are the importance of
the UK maintaining and building on its reputation for research, innovation
and ethics; embracing digital and data developments; creating a proinnovation regulatory environment; and adopting a global approach.
From a regulatory standpoint, the UK currently takes the view that,
“blanket AI-specific regulation, at this stage, would be
inappropriate…[and] that existing sector specific regulators are best placed
to consider the impact on their sector of any subsequent regulation which
may be needed.”1835 Instead, the UK regulates many aspects of the
development and use of AI through ‘cross-sector’ legislation and different
regulators. This includes coverage in data protection through the
Information Commissioner’s Office; competition through the Competition
and Markets Authority; and in human rights and equality through the
Equality and Human Rights Commission. Saying that, the Strategy does
question whether this approach is adequate. Accordingly, there are plans for
the Office for Artificial Intelligence to develop a national position on
GOV.UK, Industrial Strategy: Artificial Intelligence Sector Deal (2017)
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/702810/180425_BEIS_AI_Sector_Deal__4_.pdf; GOV.UK, AI Sector Deal
(May 2019), https://www.gov.uk/government/publications/artificial-intelligence-sectordeal/ai-sectordeal#:~:text=This%20Sector%20Deal%20is%20the,to%20%C2%A3342%20million%20
from; GOV.UK, National AI Strategy (Sept. 2021)
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/1020402/National_AI_Strategy_-_PDF_version.pdf
1835
GOV.UK, National AI Strategy (Sept. 2021)
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/1020402/National_AI_Strategy_-_PDF_version.pdf.
1834
459
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
governing and regulating AI, which will be set out in a White Paper in early
2022.
In the UK, responsibility for overseeing implementation of the
National AI Strategy sits with the Office for Artificial Intelligence, a joint
Department for Business, Energy and Industrial Strategy and Department
for Digital, Culture, Media and Sport unit. Other AI specific UK bodies and
structures include the AI Council, the Alan Turing Institute, the Centre for
Data Ethics and Innovation and most recently the newly announced UK AI
Standards Hub.1836
The UK is engaged internationally in the development of AI
governance in line with the values of fairness, freedom and democracy. This
engagement includes working with partners to shape AI governance under
development including the EU AI Act and the potential Council of Europe
legal framework. The UK has also proactively worked with the OECD, the
Council of Europe and UNESCO and helped to found the Global
Partnership on AI (GPAI). Additionally, UK defence has a strong record of
collaborating with international partners and allies. This includes
engagement with NATO supporting the AI Partnership for Defence. On the
bi-lateral front, in September 2020 the UK and the U.S. signed a to establish
a bilateral dialogue on their shared vision for driving technological
breakthroughs in AI and to explore an AI R&D ecosystem that “promotes
the mutual wellbeing, prosperity, and security of present and future
generations.” The Declaration mentions, as one objective to protect “against
efforts to adopt and apply these technologies in the service of
authoritarianism and repression.” 1837
Public Participation
The guidance of the Information Commissioner’s Office, as well as
all above-mentioned documents are publicly accessible. The ICO guidance
sets space for online feedback that individuals and organizations may wish
to provide on how the guidance can be implemented in practice.
GOV.UK, Press Release: New UK initiative to shape global standards for artificial
Intelligence (Jan 2022) https://www.gov.uk/government/news/new-uk-initiative-toshape-global-standards-for-artificial-intelligence
1837
Gov.UK, OAI, Declaration of the United States of America and the United Kingdom
of Great Britain and Northern Ireland on Cooperation in AI Research and Development
(Sept. 25, 2020), https://www.gov.uk/government/publications/declaration-of-the-unitedstates-of-america-and-the-united-kingdom-of-great-britain-and-northern-ireland-oncooperation-in-ai-research-and-development/declaration-of-the-united-states-of-americaand-the-united-ki
1836
460
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Parliamentary hearings on AI are regularly held, and are accessible to watch
online.
Facial Recognition
Human rights organizations have long criticized the UK government
for the almost unparralled deployment of CCTV. (Chongqing, China has
now overtaken London as the most surveilled city in the world.1838) Early
in 2020, London’s Metropolitan Police deployed live facial recognition.
The Met says its use of the controversial technology will be targeted to
“specific locations where intelligence suggests we are most likely to locate
serious offenders.”1839 Silkie Carlo, the director of Big Brother Watch,
called the move “an enormous expansion of the surveillance state and a
serious threat to civil liberties in the UK.” There is currently no law in the
UK governing facial recognition. More worryingly, between 2016-2019,
Met Police recognition technology was 93% inaccurate, with 3,000+ people
wrongly identified by police facial recognition.1840
AI Grading Controversy
A widely reported controversy over the use of AI in the UK public
sector erupted in the Summer of 2020. The UK used an algorithm to
estimate exam results. Nearly 40 percent of students saw their grades
reduced after the government reevaluated exams, known as “A-levels.”1841
The software model incorporated school’s past results and student’s earlier
results on mock exams. The calculations favored elites.1842 As the BBC
explained, the algorithm “locks in all the advantages and disadvantages and means that the talented outlier, such as the bright child in the low-
Matthew Keegan, Big Brother is watching: Chinese city with 2.6m cameras is world's
most heavily surveilled, The Guardian (Dec. 2, 2019)
https://www.theguardian.com/cities/2019/dec/02/big-brother-is-watching-chinese-citywith-26m-cameras-is-worlds-most-heavily-surveilled.
1839
Vikram Dodd, Met police to begin using live facial recognition cameras in London,
The Guardian (Jan. 24, 2020),
https://www.theguardian.com/technology/2020/jan/24/met-police-begin-using-live-facialrecognition-cameras
1840
Big Brother Watch. https://bigbrotherwatch.org.uk/campaigns/stop-facialrecognition/
1841
Adam Satariano, British Grading Debacle Shows Pitfalls of Automating Government,
New York Times (Aug. 20, 2020) (“The uproar over an algorithm that lowered the grades
of 40 percent of students is a sign of battles to come regarding the use of technology in
public services.”), https://www.nytimes.com/2020/08/20/world/europe/uk-englandgrading-algorithm.html
1842
Karla Adams, The UK used an algorithm to estimate exam results. The calculations
favored elites, The Washington Post (Aug. 17, 2020), https://www.washingtonpost.com
1838
461
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
achieving school, or the school that is rapidly improving, could be delivered
an injustice.”1843
As the Open Data Institute pointed out, a student would have
received a high grade in math only because historically someone from her
school had received a high school although the same student was predicted
at B or C.1844 The new evaluation method was triggered by the corona virus
since in-person exams had to be cancelled and the government sought to
standardize college admissions. Wired UK reported that some researchers
stated that “[r]ather than the algorithm getting it wrong, …it was simply the
wrong algorithm.”1845 However, others thought that the application of
Article 22 of the General Data Protection Regulation (prohibition of
decisions solely made by automated decision making) was at stake, albeit
disputed by the governmental agency that suggested the computergenerated score. Ultimately, protests in front of the British Parliament and
a pending lawsuit led the government to withdraw the system.1846
Karen Hao, a reporter with MIT Technology Review, wrote “The
problem began when the exam regulator lost sight of the ultimate goal—
and pushed for standardization above all else.”1847
NGO Perspectives on AI in the UK
NGO perspectives on the use of AI in the UK have previously
centered on the AI grading controversy but also with the automated process
for settling the status of EU nationals post-Brexit. The application launched
by the government to determine the status of EU nationals resident in the
UK was based on automated face recognition and automated data matching
across government departments. It displayed a number of errors including
Sean Coughlan, Why did the A-level algorithm say no?, BBC (Aug. 14, 2020),
https://www.bbc.com/news/education-53787203
1844
Open Data Institute, What can we learn from the qualifications fiasco? – The ODI
(Aug. 24, 2020), https://theodi.org/article/what-can-we-learn-from-the-qualificationsfiasco/
1845
Matt Burgess, The lessons we all must learn from the A-levels algorithm debacle,
WiredUK (Aug. 20, 2020) (“Unless action is taken, similar systems will suffer from the
same mistakes. And the consequences could be dire”)
https://www.wired.co.uk/article/gcse-results-alevels-algorithm-explained
1846
Daan Kolkman, “F**k the algorithm”?: What the world can learn from the UK’s Alevel grading fiasco, London School of Economic Blog (Aug. 26, 2020),
https://blogs.lse.ac.uk/impactofsocialsciences/2020/08/26/fk-the-algorithm-what-theworld-can-learn-from-the-uks-a-level-grading-fiasco/
1847
Karen Hao, The UK exam debacle reminds us that algorithms can’t fix broken
systems, MIT Technology Review (Aug. 20, 2020),
https://www.technologyreview.com/2020/08/20/1007502/uk-exam-algorithm-cant-fixbroken-system/
1843
462
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
for example the denial of a settled status (granted after 5 years of residence)
and thus the possibility to legally remain in the country of a French woman
who had worked for 15 years in the UK and was married to a British national
with two kids. Further, the data required by the app to determine the status
of applicants did not include child benefits or child tax credits, and thus
could be discriminatory towards women since 87% of child benefit
recipients were female. Further, the algorithm used to determine visa
applications known as 'the streaming tool' was found opaque and
discriminatory toward applicants from certain nationalities and race groups.
After a successful legal challenge, the government committed to a "redesign
of the process and the way in which visa applications are allocated for
decision-making."
Global Partnership on AI and OECD AI Principles
The UK is a member of the OECD and the G20 and therefore should
adhere to the OECD/G20 AI Principles. Nonetheless, it is interesting to
observe that the OECD/G20 AI Principles are not referred to in the National
AI Strategy detailed above. The UK is one of the founding members of the
GPAI.1848 The UK recently announced a £1m investment in GPAI’s data
trust research.1849
Data Protection
In July 2020, the UK’s Information Commissioner’s Office (ICO),
an independent governmental agency set up to “uphold information rights
in the public interest” published guidance to clarify how to assess the risks
to rights and freedoms that AI can pose from a data protection perspective;
and the appropriate measures that can be implemented to mitigate them.1850
The ICO is in effect the data protection watchdog of the UK set in
accordance with the EU’s Data Protection Directive as implemented by the
UK, and later replaced with the General Data Protection Regulation
(GDPA). The updated AI Guidance of the ICO states that it “aims to
mitigate the risks specifically arising from a data protection perspective . . .
without losing sight of the benefits such projects can deliver.” The
Gov.UK, Joint statement from founding members of the Global Partnership on
Artificial Intelligence (June 15, 2020),
https://www.gov.uk/government/publications/joint-statement-from-founding-membersof-the-global-partnership-on-artificial-intelligence.
1849
GOV.UK, National AI Strategy (Sept. 2021)
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_
data/file/1020402/National_AI_Strategy_-_PDF_version.pdf
1850
UK ICO, Guidance on AI and Data Protection (July 30, 2020),
https://ico.org.uk/media/for-organisations/guide-to-data-protection/key-data-protectionthemes/guidance-on-ai-and-data-protection-0-0.pdf.
1848
463
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
emphasis is thus on ensuring the fairness, transparency and lawfulness of
AI projects but also on data security and public awareness. Familiar
governance structures taken up from the data protection and privacy domain
(such as accountability and impact assessments but also the idea of privacy
by design) are exported to the field of AI. The well-known challenge of AI
to data minimization is mentioned by the ICO, albeit in a somewhat cursory
fashion.
Algorithmic Transparency
The UK Data Protection Act 20181851 includes specific provisions
on the right for individuals to intervene in automated decision-making.1852
The concept of Algorithmic Transparency derived from the UK’s data
protection framework is addressed in some detail in the AI Guidance
published by the Information Commissioner’s Office.
In November 2020, the CDEI published the final report of its review
into bias in algorithmic decision-making.1853 The Center focused on the use
of algorithms in significant decisions about individuals in four sectors
(policing, local government, financial services and recruitment). Key
recommendations include: 1) Government should place a mandatory
transparency obligation on all public sector organisations using algorithms
that have an impact on significant decisions affecting individuals; 2)
Organisations should be actively using data to identify and mitigate
bias. They should make sure that they understand the capabilities and
limitations of algorithmic tools, and carefully consider how they will ensure
fair treatment of individuals, and 3) Government should issue guidance that
clarifies the application of the Equality Act to algorithmic decisionmaking. This should include guidance on the collection of data to measure
bias, as well as the lawfulness of bias mitigation techniques (some of which
risk introducing positive discrimination, which is illegal under the Equality
Act).
More recently in May 2021, the Office for Artificial Intelligence
published the Ethics, Transparency and Accountability Framework for
Automated Decision-Making for public sector organizations on how to use
Legislation.gov.uk, Data Protection Act 2018,
https://www.legislation.gov.uk/ukpga/2018/12/contents/enacted
1852
Legislation.gov.uk, Data Protection Act 2018 (Sect. 95 – “Right to intervene in
automated decision-making”),
https://www.legislation.gov.uk/ukpga/2018/12/section/97/enacted
1853
Gov.UK, CDEI publishes review into bias in algorithmic decision-making (Nov. 27,
2020), https://www.gov.uk/government/publications/cdei-publishes-review-into-bias-inalgorithmic-decision-making
1851
464
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
automated or algorithmic decision-making systems in a safe, sustainable
and ethical way.1854 Transparency is also mentioned, albeit only in passing,
also in the AI Ethics & Safety Framework, issued by the Office for Artificial
Intelligence.1855 As a part of the Office’s recommendation on integrating
“responsible innovation” into AI projects, the organization OAI advises
advices that AI developers should “prioritise the transparency of how [they]
design and implement [their] model, and the justification and
interpretability of its decisions and behaviours.”
Human Rights
All UK AI initiatives must comply with the UK Human Rights Act
of 1998. Post-Brexit the UK remains a part of the European Convention on
Human Rights and is subject to the jurisdiction of the European Court of
Human Rights in Strasbourg. Therefore, even if the GDPR can no longer be
enforced in the UK through judgments of the Court of Justice of the
European Union in Luxembourg, the existence of structures like the UK’s
Information Commissioner’s Office and generally, the remnants of EU’s
Acquis Communautaire make it clear that the UK has in place a relatively
robust system of human rights protection. According to Freedom House,
UK receives high marks (93/100) for political rights and civil liberties1856.
The country is “is a stable democracy that regularly holds free elections and
is home to a vibrant media sector.” Currently, the public debate is often
about loss of jobs due to the growth of the robotics sector. Relatedly, in
2020 the Alan Turing Institute issued guidelines on AI and nondiscrimination/human rights.
Evaluation
The UK has endorsed the OECD/G20 AI Principles, and has a good
record on human rights. The UK has established several public bodies that
have issued policies and guidance on AI, including the Office for Artificial
Intelligence and the Centre for Data Ethics. Although the AI strategy is open
and public participation is encouraged, the UK suffered a significant public
backlash over the grading controversy in 2020. Laudable strategies such as
GOV.UK, Ethics, Transparency and Accountability Framework for Automated
Decision-Making (May 2021) https://www.gov.uk/government/publications/ethicstransparency-and-accountability-framework-for-automated-decision-making/ethicstransparency-and-accountability-framework-for-automated-decision-making
1855
GOV.UK, Guidance: Understanding artificial intelligence ethics and safety (Jun
2019) https://www.gov.uk/guidance/understanding-artificial-intelligence-ethics-andsafety
1856
Freedom House, Freedom in the World 2021 – United Kingdom,
https://freedomhouse.org/country/united-kingdom/freedom-world/2021
1854
465
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the development of AI-related workforce and education sector initiatives,
as well as the promise of data trusts need to be monitored closely. There is,
at the moment, no express support for the Universal Guidelines for AI.
466
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
United States
National AI Strategy
The United States promotes AI policies that seek to maintain
American leadership and to build alliances with other democratic countries.
Speaking to the United Nations General Assembly in September, 2021,
President Biden reaffirmed the U.S.’s commitment to “work together with
our democratic partners to ensure that new advances in areas from
biotechnology, to quantum computing, 5G, artificial intelligence, and more
are used to lift people up, to solve problems, and advance human freedom
— not to suppress dissent or target minority communities.”1857 President
Biden’s UN statement echoed similar comments to the G7 leaders at the
Munich Security Conference earlier in the year. At that meeting, President
Biden called for “rules that will govern the advance of technology and the
norms of behavior in cyberspace, artificial intelligence, and biotechnology,”
that will “lift people up” and not pin them down. Biden also urged the G7
nations to stand up for “democratic values.”1858
The current US position on AI is comprised of a 2020 Presidential
Executive Order, a 2019 Executive Order, OMB Guidance for Regulation
of AI Applications, the recommendations of a National Security
Commission on AI, and various initiatives and programs established by the
National AI Initiative Act (NAIIA).1859 The 2019 Executive
Order emphasized the need to maintain American leadership in Artificial
Intelligence, and sets out a range of policies and practices, including
funding, research, training, and collaboration.1860 The Executive Order also
describes the need protect “civil liberties, privacy, and American values.”
The Agency Guidance also underscores the desire to maintain American
leadership, and endorses such values as privacy, civil liberties, human
The White House, Remarks by President Biden Before the 76th Session of the United
Nations General Assembly (Sep. 21, 2021), https://www.whitehouse.gov/briefingroom/speeches-remarks/2021/09/21/remarks-by-president-biden-before-the-76th-sessionof-the-united-nations-general-assembly/.
1858
The White House, Remarks by President Biden at the 2021 Virtual Munich Security
Conference (Feb. 19, 2021), https://www.whitehouse.gov/briefing-room/speechesremarks/2021/02/19/remarks-by-president-biden-at-the-2021-virtual-munich-securityconference/
1859
William M. (Mac) Thornberry National Defense Authorization Act for Fiscal Year
2021, Division E (“National Artificial Intelligence Initiative Act”),
https://www.congress.gov/116/plaws/publ283/PLAW-116publ283.pdf
1860
The White House, Executive Order on Maintaining American Leadership in Artificial
Intelligence (Feb. 11, 2019), https://www.whitehouse.gov/presidential-actions/executiveorder-maintaining-american-leadership-artificial-intelligence/
1857
467
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
rights, the rule of law, and respect for intellectual property.1861 The Agency
Guidance outlines 10 principles, including Fairness and NonDiscrimination, Disclosure and Transparency, to promote innovation and
growth for AI.
The 2020 Executive Order on Promoting the Use of Trustworthy
Artificial Intelligence in the Federal Government reflects earlier goals set
in the 2019 Executive Order and established common guidance to
encourage federal agencies to use AI, consistent with nine principles. 1862
The 2020 Executive Order states that the “The ongoing adoption and
acceptance of AI will depend significantly on public trust.” The 2020
Executive Order repeatedly emphasizes the need to ensure that “the use of
AI remains consistent with all applicable laws, including those related to
privacy, civil rights, and civil liberties.” The Office of Management and
Budget is directed, by June 2021, to “post a roadmap for the policy guidance
that OMB intends to create or revise to better support the use of AI,
consistent with this order. This roadmap shall include, where appropriate, a
schedule for engaging with the public and timelines for finalizing relevant
policy guidance.”
Section 3 of the 2020 Executive Order describe Principles for Use
of AI in government. “When designing, developing, acquiring, and using
AI in the Federal Government, agencies shall adhere to the following
Principles:”
a) Lawful and respectful of our nation’s values
b) Purposeful and performance-driven
c) Accurate, reliable and effective
d) Safe, secure, and resilient
e) Understandable
f) Responsible and traceable
g) Regularly monitored
h) Transparent
i) Accountable
Members of the United States Congress have also proposed
legislation for a US national AI strategy. Representatives Robin Kelly (RIllinois)and
Will
Hurd
(R-Texas) introduced a Congressional
Office of Management and Budget, Memorandum for the Heads of Executive
Departments and Agencies (draft), https://www.whitehouse.gov/wpcontent/uploads/2020/01/Draft-OMB-Memo-on-Regulation-of-AI-1-7-19.pdf
1862
The White House, Executive Order on Promoting the Use of Trustworthy Artificial.
Intelligence in the Federal Government (Dec. 3, 2020),
https://www.whitehouse.gov/presidential-actions/executive-order-maintaining-americanleadership-artificial-intelligence/
1861
468
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Resolution calling for the creation of a US National AI Strategy.1863 The US
AI National AI Resolution, which passed the House in December 2020,
emphasizes global leadership, a prepared workforce, national security,
research and development, and Ethics, reduced bias, fairness, and privacy.
The Resolution would not establish any new agency to regulate AI nor does
it make clear which new obligations would exist for those who deploy AI
systems. But the Resolution does provide a detailed outline of a US national
AI strategy. Among other points, the Resolution states “Developing and
using artificial intelligence in ways that are ethical, reduce bias, promote
fairness, and protect privacy is essential for fostering a positive effect on
society consistent with core United States values.”1864 The Resolution also
acknowledges the OECD Principles on Artificial Intelligence. The
Bipartisan Policy Center has endorsed the Resolution, declaring “we must
embrace AI while protecting our civil liberties, modernizing our workforce
and education programs, and investing more in R&D.”1865
The National Artificial Intelligence Initiative Act marks one of the
most significant developments in U.S. AI policy. It directed the President
to establish the National Artificial Intelligence Initiative (NAII), with the
aim to “lead the world in the development and use of trustworthy artificial
intelligence systems in the public and private sectors.”1866 The NAIIA also
creates the National Artificial Intelligence Initiative Office, currently led by
Dr. Lynne Parker, within the Office of Science and Technology Policy to
coordinate and support the NAII. The Act includes provisions for funding
interdisciplinary AI education and workforce training, establishing AI
research institutes, and cooperating with allies on trustworthy AI
development.
The AI in Government Act of 2020, also passed in December 2020,
establishes an “AI Center of Excellence” to facilitate cohesive and
competent adoption of AI by the government “for the purposes of
benefitting the public and enhancing the productivity and efficiency of
Federal Government operations.”1867 Similar to the 2020 Executive Order,
Congresswoman Robin Kelly, ICYMI: Kelly, Hurd Call for Creation of National AI
Strategy (Sept. 18, 2020), https://robinkelly.house.gov/media-center/press-releases/icymihurd-kelly-call-for-creation-of-national-ai-strategy
1863
1864
https://hurd.house.gov/sites/hurd.house.gov/files/Resolution%20Text%20HURDTX_030
_xml.pdf
1865
Bipartisan Policy Center, BPC: National AI Strategy Resolution A Critical Step (Sept.
16, 2020), https://bipartisanpolicy.org/press-release/bpc-national-ai-strategy-resolution-acritical-step/
1866
NDAA FY 2021, Division E, Title LI, Sec. 5101.
1867
AI in Government Act, Sec. 103(a).
469
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
the AI in Government Act requires the OMB to “issue a memorandum” to
federal agencies regarding the government use of AI in ways that protect
“civil liberties, civil rights, and economic and national security,” along with
“best practices” for identifying and mitigating bias and discriminatory
impact in the use of AI.1868 As of early 2022, the OMB has not complied
with either the 2020 Executive Order or the AI in Government Act.1869
Finally, the White House and Congress have also paid close
attention to the draft AI Act in the European Union as a model for future
U.S. AI regulation. Jake Sullivan, the White House National Security
Advisor, has stated that “United States welcomes the EU’s new initiatives
on artificial intelligence” and that the U.S. will “work with our friends and
allies to foster trustworthy AI.”1870 Lynne Parker, Director of the National
AI Initiative Office and the OSTP’s Assistant Director of AI, has described
the EU AI Act as a “very good comprehensive approach that the U.S. should
consider.”1871
OMB AI Guidance for Agencies
In November 2020, the US Office of Management and Budget
issued Guidance for Regulation of Artificial Intelligence Applications.1872
The Guidance follows from the Executive on American Leadership in AI
and states that “when considering regulations or policies relayed to AI
applications, agencies should continue to promote advancements in
technology and innovation, while protecting American technology,
economic and national security, privacy, civil liberties and other American
values, including the principles of freedom, human rights, the rule of law,
and respect for intellectual property.” The Memorandum is explicitly
addressed to AI applications “developed and deployed outside of the federal
government.
Id. at Sec. 104(a).
CAIDP Statement to OMB (Oct. 10, 2021),
https://www.caidp.org/app/download/8350420263/CAIDP-OMB-Statement19102021.pdf
1870
Jake Sullivan [@JakeSullivan46], “The United States welcomes the EU’s new
initiatives on artificial intelligence. We will work with our friends and allies to foster
trustworthy AI that reflects our shared values and commitment to protecting the rights
and dignity of all our citizens.” Twitter (Apr. 21, 2021),
https://twitter.com/jakesullivan46/status/1384970668341669891.
1871
Dan Reilly, White House A.I. director says U.S. should model Europe’s approach to
regulation, Fortune (Nov. 10, 2021), https://fortune.com/2021/11/10/white-house-a-idirector-regulation/.
1872
OMB, Guidance for Regulation of Artificial Intelligence Applications (Nov. 17,
2020), https://www.whitehouse.gov/wp-content/uploads/2020/11/M-21-06.pdf
1868
1869
470
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
The OMB Guidelines on AI restate key goals for the Stewardship
of AI applications:
• Public Trust in AI
• Public Participation
• Scientific Integrity and Information Quality
• Risk Assessment and Management
• Benefits and Costs
• Flexibility
• Fairness and Non-discrimination
• Disclosure and Transparency
• Safety and Security
• Interagency Cooperation
The OMB Guidelines encourage communications to the public,
describing both the benefits and risks “in a manner that promotes public
trust and understanding of AI.” The Guidelines continue, “agencies should
communicate this information transparently by describing the underlying
assumptions and uncertainties regarding expected outcomes, both positive
and negative.”
There are provisions in the OMB AI Guidelines that are
controversial. The OMB recommends that agencies “promote public access
to government data and models where appropriate but fails to note whether
such government data is personal data or may be subject to protections
under federal law.
EU-U.S. Trade and Technology Council (TTC)
In June 2021, the U.S. and the European Union established the EUU.S. Trade and Technology Council (TTC) to “strengthen global
cooperation on technology, digital issues, and supply chains” and “with the
aim of promoting a democratic model of digital governance.”1873 At the
TTC’s inaugural meeting, the U.S. and EU acknowledged that “AI
technologies yield powerful advances but can also threaten our shared
values and fundamental freedoms” and committed to “develop and
implement AI systems that are innovative and trustworthy and that respect
universal human rights and shared democratic values.”
The U.S. and EU also expressed “significant concerns” with the use
of “social scoring systems with an aim to implement social control at scale.”
Noting that such uses of AI “pose threats to fundamental freedoms and the
The White House, U.S.-EU Trade and Technology Council Inaugural Joint Statement
(Sept. 29, 2021), https://www.whitehouse.gov/briefing-room/statementsreleases/2021/06/15/u-s-eu-summit-statement/
1873
471
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
rule of law,” the TTC stated its opposition to the use of AI for social scoring
or other “rights-violating systems.” Finally, the TTC outlined areas of U.S.EU cooperation, including “responsible stewardship of trustworthy AI”
through the OECD AI Recommendations, “measurement and evaluation
tools” to assess accuracy and bias, and the development of “AI technologies
designed to enhance privacy protections.”
The U.S. has made progress in support of the TTC’s goals. In
December 2021, the Biden Administration announced an initiative to
encourage development of “Democracy-Affirming Technologies,” that
support democratic values and governance.1874 Relatedly, the U.S. and UK
announced plans to promote Privacy Enhancing Technologies (PETs),
including low-data AI, the deletion of unnecessary data, and techniques for
robust anonymity.1875
Also in December 2021, the EU-U.S. Joint Technology Competition
Policy Dialogue was launched by the Federal Trade Commission, the
Department of Justice’s Antitrust Division, and the European Commission.
The Joint Dialogue is intended to align with the E.U.-US Trade and
Technology Council (TTC)’s goals through coordination “as much as
possible on policy and enforcement,” “especially in technology sectors,”
and by promoting “greater alignment” between the U.S. and EU.1876
Facial Recognition
There are wide-ranging protests in the United States against the
deployment of facial recognition technology. In May 2019, San Francisco
became the first city in the U.S. to ban the use of facial recognition
technology by city agencies.1877 The city supervisor said, “It’s
psychologically unhealthy when people know they’re being watched in
1874
The White House, Fact Sheet: Announcing the Presidential Initiative for
Democratic Renewal (Dec. 9, 2021) https://www.whitehouse.gov/briefingroom/statements-releases/2021/12/09/fact-sheet-announcing-the-presidential-initiativefor-democratic-renewal/
1875
The White House, US and UK to Partner on Prize Challenges to Advance PrivacyEnhancing Technologies (Dec. 8, 2021) https://www.whitehouse.gov/ostp/newsupdates/2021/12/08/us-and-uk-to-partner-on-a-prize-challenges-to-advance-privacyenhancing-technologies/
1876
the United States Department of Justice Antitrust Division and the United States
Federal Trade Commission. EU-U.S. Joint Technology Competition Policy Dialogue
Inaugural Joint Statement between the European Commission (Dec. 7, 2021)
https://www.ftc.gov/system/files/documents/public_statements/1598739/euus_joint_dialogue_statement_12721.pdf
1877
Kate Conger, Richard Fausset and Serge F. Kovaleski, San Francisco Bans Facial
Recognition Technology (May 14, 2019), https://www.nytimes.com/2019/05/14/us/facialrecognition-ban-san-francisco.html
472
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
every aspect of the public realm, on the streets, in parks.” Other cities,
including Cambridge, Oakland, and Portland followed. In October 2019,
California enacted a moratorium on the use of facial recognition technology
in police body cameras.1878 The bill prohibits the use of biometric
surveillance technology, which includes facial-recognition software, in
police body cameras. It also prohibits police from taking body-camera
footage and running it through facial-recognition software at a later time. It
does not prevent state and local police from using facial-recognition
technology in other ways, such as in stationary cameras, and it does not
apply to federal agencies such as the FBI.1879
In November 2021, Facebook announced it would shut down its
facial recognition system and delete the face scans of more than one billion
users it had gathered.1880 The announcement followed multiple lawsuits
alleging that the company had violated state and federal privacy laws as
well as repeated recommendations to the FTC that the company’s business
practices, including the collection and use of facial images, violated a 2011
settlement with the Commission.
In February 2022, and after months of criticism by advocacy groups,
the Internal Revenue Service announced that it would halt its program with
identity verification company ID.me to use facial recognition technology to
verify taxpayers’ identities.1881 The program would have required taxpayers
to take video selfies to verify themselves, raising concerns that citizens will
be “coerced into handing over their sensitive biometric information to the
government in order to access essential services.”
A bill introduced in the United States Congress would ban the use
of facial recognition by law enforcement agencies.1882 The would make it
illegal for any federal agency or official to “acquire, possess, access, or use”
biometric surveillance technology in the US. It would also require state and
1878
California Legislative Information, AB-1215 Law enforcement: facial recognition
and other biometric surveillance (Oct. 9, 2019),
1879
Rachel Metz, California lawmakers ban facial-recognition software from police body
cams (Sept. 13, 2019), https://www.cnn.com/2019/09/12/tech/california-body-cam-facialrecognition-ban/index.html
1880
Kashmir Hill and Ryan Mac, Facebook, Citing Societal Concerns, Plans to Shut
Down Facial Recognition System, New York Times (Nov. 2, 2021),
https://www.nytimes.com/2021/11/02/technology/facebook-facial-recognition.html
1881
Alan Rappeport and Kashmir Hill, I.R.S. to End Use of Facial Recognition for
Identity Verification: After a bipartisan backlash, the agency will transition away from
using a service from ID.me, New York Times (Feb. 7, 2022),
https://www.nytimes.com/2022/02/07/us/politics/irs-idme-facial-recognition.html
1882
Congress.gov, S.4084 - Facial Recognition and Biometric Technology Moratorium
Act of 2020 (June 25, 2020), https://www.congress.gov/bill/116th-congress/senatebill/4084
473
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
local law enforcement to bring in similar bans in order to receive federal
funding.1883 The bill was introduced by Senators Ed Markey Jeff Merkley,
and Representatives Pramila Jayapal and Ayanna Pressley.
In February 2022, the same members of Congress urged federal
agencies to end the use of Clearview AI Facial Recognition Technology.1884
In letters to the Department of Homeland Security (DHS), Department of
Justice (DOJ), Department of Defense (DOD), Department of Interior
(DOI), and Department of Health and Human Services (HHS), they wrote,
“Facial recognition tools pose a serious threat to the public’s civil liberties
and privacy rights, and Clearview AI’s product is particularly dangerous.
We urge you to immediately stop the Department’s use of facial recognition
technology, including Clearview AI’s tools.”
National Security Commission on AI
The US Congress established the National Security on AI in
2018.1885 The Commission has issued several reports and made
recommendations to Congress. The National AI Commission issued
an interim report in November 2019, which was criticized for its lack of
attention to democratic values.1886 In a more recent report Key
Considerations for Responsible Development and fielding of Artificial
Intelligence, the Commission recommends “Employ[ing] technologies and
operational policies that align with privacy preservation, fairness, inclusion,
human rights, and [the] law of armed conflict.”1887
In March 2021, the National Security Commission on AI released
the Final Report for an integrated U.S. strategy for “the coming era of AIaccelerated competition and conflict.”1888 The Final Report included
numerous recommendations for the U.S. government to use AI to protect
MIT Technology Review, A new US bill would ban the police use of facial
recognition (June 26, 2020), https://www.technologyreview.com/2020/06/26/1004500/anew-us-bill-would-ban-the-police-use-of-facial-recognition/
1884
Ed Markey, Senators Markey and Merkley and Reps. Jayapal & Pressley Urge
Federal Agencies to End Use of Clearview AI Facial Recognition Technology (Feb. 9,
2022), https://www.markey.senate.gov/news/press-releases/senators-markey-andmerkley-and-reps-jayapal_pressley-urge-federal-agencies-to-end-use-of-clearview-aifacial-recognition-technology
1885
National Security Commission on AI, Home, https://www.nscai.gov/home
1886
National Security Commission on AI, Interim Report (Nov, 2019),
https://www.epic.org/foia/epic-v-ai-commission/AI-Commission-Interim-Report-Nov2019.pdf
1887
National Security Commission on AI, Key Considerations and Responsible
Development and Fielding of Artificial Intelligence (July 22, 2020),
https://drive.google.com/file/d/1_zkNkT3Trz3rtFc8KVrEBNlg2R9MaUpi/view
1888
NSCAI Final Report, at 8.
1883
474
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
U.S. interests and to support AI innovation. The Final Report included
recommendations that the U.S. government improve public transparency in
its use of AI (including through AI risk and impact assessments), develop
and test “technical approaches to preserving privacy, civil liberties, and civil
rights,” and strengthen redress and due process mechanisms for victims of
AI-related harms. The report also called on the White House and State
Department to establish an “Emerging Technology Coalition (ETC) of
countries respectful of democratic values” to promote emerging
technologies “according to democratic norms and values,” including by
“building on” the OECD AI Principles and the work of the Global
Partnership on AI (GPAI).1889 Finally, the Final Report called on the U.S.
and ETC partners to create an “International Digital Democracy Initiative”
that would rely, in part, on the OECD AI Principles to develop AI
guidelines.1890
While the Final Report called for “baseline standards and safeguards
regarding facial recognition,” the Commission failed to address several
problems previously identified by CAIDP.1891 Beyond the lack of
opportunities for formal comment or input from the general public during
its drafting, the report failed to assess U.S. compliance with the OECD AI
Principles or G20 guidelines, support prohibitions on lethal autonomous
weapons or facial recognition technology, despite growing public concern
and widespread support in Congress.
NIST Risk Management Framework
The National Institute of Standards and Technology (NIST) within
the Department of Commerce announced the development of a voluntary
AI risk management framework (RMF) in July 2021. The framework aims
to “address risks in the design, development, use, and evaluation of AI
products, services, and systems.”1892 As a part of the broader National AI
Initiative, NIST hopes to produce a framework that can “develop along with
the technology,” “help[ing to] create and safeguard trust” in AI while
“permit[ting] the flexibility for innovation.”
Id. at 519-20.
Id. at 524.
1891
CAIDP Statement on draft final report of US National Security Commission on AI,
(Feb. 26, 2021). https://www.caidp.org/app/download/8297285563/CAIDP-NSCAI02262021.pdf
1892
NIST, AI Risk Management Framework, https://www.nist.gov/itl/ai-riskmanagement-framework; NIST, AI Risk Management Framework Concept Paper (Dec.
13, 2021),
https://www.nist.gov/system/files/documents/2021/12/14/AI%20RMF%20Concept%20P
aper_13Dec2021_posted.pdf
1889
1890
475
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Throughout 2021, NIST worked with and solicited input from the
public to develop the framework, identifying “characteristics of
trustworthiness” for AI systems: “accuracy, explainability and
interpretability, reliability, privacy, robustness, safety, security (resilience),
and mitigation of unintended and/or harmful bias, as well as of harmful
uses.” NIST anticipates producing a completed version 1.0 of the AI RMF
by early 2023.
JAIC
The Joint Artificial Intelligence Center (JAIC) is a research center
within the Department of Defense. The mission of the JAIC is to “transform
the DoD by accelerating the delivery and adoption of AI to achieve mission
impact at scale.1893 The JAIC has recently undertaken an ambitious agenda
to “accelerate the adoption of AI across every aspect of the military’s
warfighting and business operations.”1894 The new mission set is in contrast
to the JAIC’s introductory goal, which was to jumpstart AI in DoD through
pathfinder projects.
Algorithmic Transparency
The United States does not have an overarching privacy law, such
as the GDPR, nor is there a privacy agency, and there is no general law that
establishes a right of algorithmic transparency.
In April 2021, the FTC outlined a series of recommendations to
encourage transparency in the development and use of AI.1895 Pointing to
the agency’s authority under the FTC Act, the Fair Credit Reporting Act,
and the Equal Credit Opportunity Act, along with its January 2021
settlement requiring the photo app firm Everalbum to “delete models and
algorithms it developed by using the photos and videos uploaded by its
U.S. Department of Defense, Chief Information Officer, Vision: Transform the DoD
Through Artificial Intelligence. https://dodcio.defense.gov/About-DoDCIO/Organization/JAIC/
1894
Scott Maucine, JAIC entering new phase of life, will create teams to help DoD adopt
AI (Nov. 26, 2020), https://federalnewsnetwork.com/defense-main/2020/11/jaic-enteringnew-phase-of-life-will-create-teams-to-help-dod-adopt-ai/
1895
Federal Trade Commission, Aiming for truth, fairness, and equity in your company’s
use of AI (Apr. 19, 2021),
https://www.ftc.gov/news-events/blogs/business-blog/2021/04/aiming-truth-fairnessequity-your-companys-use-ai
1893
476
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
users,”1896 the FTC warned businesses using AI to “Hold yourself
accountable – or be ready for the FTC to do it for you.”1897
At the state level, the California Consumer Privacy Rights Act
(CPRA) updates the states privacy law and establishes a right to limit
algorithmic profiling. Businesses responding to requests for access are
required to include meaningful information around the logic behind the
decision-making processes and the likely outcome of the process with
respect to the consumer.”1898 A former U.S. federal official said the CPRA
would impose “new requirements for businesses to protect personal
information, including by ‘reasonably’ minimizing data collection, limiting
data retention, and protecting data security. It also strengthens
accountability measures by requiring companies to conduct privacy risk
assessments and cybersecurity audits, and regularly submit them to
regulators. In addition, it supplements the individual rights in the CCPA
with new notification requirements, clarifies that individuals have the right
to opt out of both the ‘sale’ and ‘sharing’ of personal information, and adds
protections for a new category of ‘sensitive data.’”1899
A separate California ballot initiative concerning AI-based profiling
for criminal justice was defeated. Proposition 25 would have removed the
right of people accused of a non-violent crime to secure their release by
posting bail or by order of a judge with an automated system of computergenerated predictive modelling. Civil rights groups favored Proposition 24
and opposed Proposition 25.1900 Alice Huffman, President of California
NAACP stated, that “Prop. 25 will be even more-discriminatory against
African Americans, Latinos and other minorities. Computer models may be
good for recommending songs and movies, but using these profiling
methods to decide who gets released from jail or who gets a loan has been
proven to hurt communities of color.” Regarding the California Privacy
1896
US FTC Requires Deletion of AI Models Developed from Data Unfairly Obtained,
CAIDP Update 2.03, https://dukakis.org/center-for-ai-and-digital-policy/caidp-update-usftc-requires-deletion-of-ai-models-developed-from-data-unfairly-obtained/
1897
Federal Trade Commission, Aiming for truth, fairness, and equity in your company’s
use of AI (Apr. 19, 2021), https://www.ftc.gov/news-events/blogs/businessblog/2021/04/aiming-truth-fairness-equity-your-companys-use-ai
1898
Briana Falcon, Devika Kornbacher, Prop 24 Gets A Yes: California Privacy Rights
Act To Become Law, J.D. Supra (Nov. 5, 2020),
https://www.jdsupra.com/legalnews/prop-24-gets-a-yes-california-privacy-21838/
1899
Cameron F. Kerry and Caitlin Chin, By passing Proposition 24, California voters up
the ante on federal privacy law, Brookings (Nov. 17, 2020),
https://www.brookings.edu/blog/techtank/2020/11/17/by-passing-proposition-24california-voters-up-the-ante-on-federal-privacy-law/
1900
Official Voter Information Guide,
https://vig.cdn.sos.ca.gov/2020/general/pdf/complete-vig.pdf
477
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Rights Act, Huffman stated, “Prop. 24 allows consumers to stop companies
from using online racial profiling to discriminate against them.”
The Algorithmic Accountability Act, introduced in February 2022,
would require large entities that deploy “automated decision systems” to
conduct impact assessments for “augmented critical decision processes,”
submit impact assessment summaries to the FTC, and mitigate the “material
negative impacts” of automated decision systems.1901
OECD AI Principles
The United States fully supported the OECD AI policy process,
endorsed the OECD AI Principles, and is a founding member of the Global
Partnership on AI. The OECD notes that the United States has taken several
steps to implement the AI Principles.
Since endorsing the OECD AI Principles in 2019,1902 the U.S. has
continued to voice its support. The U.S. and EU, in the joint statement on
the launched of the Trade and Technology Council, affirmed “their
commitment to . . . shared democratic values and respects universal human
rights, which they have already demonstrated by endorsing the OECD
Recommendation on AI,” which they “intend to continue to uphold and
implement.”1903 At a keynote address to the OECD in October 2021,
Secretary of State Antony Blinken stated that the “OECD’s Principles on
AI back in 2019—the first set of intergovernmental principles on the
topic—and the launch of the Global Partnership on AI in 2020, laid a
foundation for the world to build on.”1904
Public Participation and Access to Documents
The United States government provides access to all final policy
proposals concerning AI. Federal agencies have undertaken public
rulemakings and requested public comment. However, the National
Security Commission attempted to keep secret its deliberations. A federal
court later determined that the AI Commission had violated US open
Algorithmic Accountability Act, Sec. 3(b)(1)
US Mission to the OECD, White House OSTP’s Michael Kratsios Keynote on AI
Next Steps (May 21, 2019), https://usoecd.usmission.gov/white-house-ostps-michaelkratsios-keynote-on-ai-next-steps/
1903
The White House, U.S.-EU Trade and Technology Council Inaugural Joint Statement
(Sept. 29, 2021), https://www.whitehouse.gov/briefing-room/statementsreleases/2021/09/29/u-s-eu-trade-and-technology-council-inaugural-joint-statement/
1904
U.S. Department of State, Secretary Antony J. Blinken at OECD Opening and
Keynote Address (Oct. 5, 2021), https://www.state.gov/secretary-antony-j-blinken-atoecd-opening-and-keynote-address/; see CAIDP Update 2.38,
https://www.caidp.org/app/download/8352772763/CAIDP-Update-2.38.pdf
1901
1902
478
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
government laws and was required to make both its records and its meetings
open to the public.1905 Oddly, the AI Commission initially made agency
documents available on a proprietary platform rather than an agency
website.1906
In 2021, the United States began several new initiatives to promote
public participation in AI policy. In June, the White House launched the
National Artificial Intelligence Research Resource Task Force to “develop
a roadmap to democratize access to research tools that will promote AI
innovation and fuel economic prosperity.”1907 In July, the White House
Office of Science and Technology Policy (OSTP) and the National Science
Foundation (NSF) sought input on the implementation plan for a National
Artificial Intelligence Research Resource (NAIRR).1908 The Office of
Science and Technology Policy began a series of public meetings and
requests for information regarding AI policy. The AI Initiative Office
continues to regularly post and update AI policy publications, including
requests for information, concept papers and reports, ethical principles, and
1909
agency budgets, in its Publications Library on ai.gov.
Human Rights
The United States endorsed the Universal Declaration for Human
Rights, published a detailed annual report on human rights, and has
historically ranked high for the protection of human rights. But in 2021
Freedom House reported “in recent years its democratic institutions have
suffered erosion, as reflected in partisan pressure on the electoral process,
bias and dysfunction in the criminal justice system, harmful policies on
immigration and asylum seekers, and growing disparities in wealth,
economic opportunity, and political influence.” Freedom House scored the
United States at 83/100 in 2021, down from 86/100 in 2020.1910 On
1905
EPIC v. AI Commission, Seeking Public Access to the records and meetings of the
NSCAI, https://www.epic.org/foia/epic-v-ai-commission/
1906
National Security Commission on AI, Interim Report and Third Quarter
Recommendation (Oct. 2020) (federal agency report stored on a Google drive server),
https://drive.google.com/file/d/1jg9YlNagGI_0rid-HXY-fvJOAejlFIiy/view
1907
The White House, The Biden Administration Launches the National Artificial
Intelligence Research Resource Task Force (June 10, 2021),
https://www.whitehouse.gov/ostp/news-updates/2021/06/10/the-biden-administrationlaunches-the-national-artificial-intelligence-research-resource-task-force/
1908
Federal Register, Request for Information (RFI) on an Implementation Plan for a
National Artificial Intelligence Research Resource (July 23, 2021),
1909
National Artificial Intelligence Initiative Office, Publications Library,
https://www.ai.gov/publications/
1910
Freedom House, Freedom in the World 2021 – United States (2021),
https://freedomhouse.org/country/united-states/freedom-world/2021
479
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
transparency, Freedom House noted that the Trump administration operated
with “greater opacity than its immediate predecessors, for example by
making policy and other decisions without meaningful input from relevant
agencies and their career civil servants.”
The United States is not a member of the Council of Europe but did
sign and ratify the COE Convention on Cybercrime,1911 as COE
conventions are open for ratification by non-members states. The US could
ratify the COE Modernized Privacy Convention as well as any future COE
Convention on AI.
In October 2021, the Office of Science and Technology Policy
announced its intention to develop an AI “bill of rights” to “codify” the idea
that “[p]owerful technologies should be required to respect our democratic
values and abide by the central tenet that everyone should be treated
fairly.”1912 The OSTP also issued a request for information on current or
planned uses of AI-enabled biometric technologies,1913 along with six
public events on “the Bill of Rights for an Automated Society” in order “to
promote public education and engagement” on AI issues.1914 CAIDP board
members suggested at the time that the process of formulating an AI Bill of
Rights 1) aim for a small number of clear, powerful principles, 2) build on
prior initiatives, 3) proceed on a bipartisan basis, and 4) proceed without
delay.1915 Although the OSTP issued an update on its “continuing work” on
Council of Europe, Chart of signatures and ratifications of Treaty 185, Convention
on Cybercrime (Status as of Nov. 11, 2020),
https://www.coe.int/en/web/conventions/full-list/-/conventions/treaty/185/signatures
1912
Eric Lander and Alondra Nelson, Americans Need a Bill of Rights for an AI-Powered
World: The White House Office of Science and Technology Policy is developing
principles to guard against powerful technologies—with input from the public (Oct. 8,
2021), https://www.wired.com/story/opinion-bill-of-rights-artificial-intelligence/
1913
Office of Science and Technology Policy, Notice of Request for Information (RFI) on
Public and Private Sector Uses of Biometric Technologies (Oct. 8, 2021),
https://www.federalregister.gov/documents/2021/10/08/2021-21975/notice-of-requestfor-information-rfi-on-public-and-private-sector-uses-of-biometric-technologies
1914
The White House, Join the Effort to Create A Bill of Rights for an Automated
Society: OSTP Announces Public Events in November to Engage the American Public in
National Policymaking about AI and Equity (Nov. 10, 2021),
https://www.whitehouse.gov/ostp/news-updates/2021/11/10/join-the-effort-to-create-abill-of-rights-for-an-automated-society/
1915
Dr. Lorraine Kisselburgh and Marc Rotenberg, Next Steps on U.S. AI Bill of Rights,
The Washington Spectator (Nov. 2, 2021), https://washingtonspectator.org/ai-bill-ofrights/
1911
480
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI that “aligns with our democratic values” in February 2022,1916 progress
on the AI Bill of Rights and its impact remain unclear.
Evaluation
The US lacks a unified national policy on AI but President Biden,
and his top advisors, has expressed support for AI aligned with democratic
values. The United States has endorsed the OECD/G20 AI Principles. The
White House has issued two Executive Orders on AI that reflect democratic
values, a federal directive encourages agencies to adopt safeguards for AI,
The most recent Executive Order also establishes a process for public
participation in the development of federal regulations on AI though the
rulemaking has yet to occur. The overall US policy-making process remains
opaque and the Federal Trade Commission has failed to act on several
pending complaints concerning the deployment of AI techniques in the
commercial sector. But the administration has launched new initiatives and
encouraged the OSTP, NIST, and other agencies to gather public input.
There is widespread objection to the use of facial recognition, and both
Facebook and the IRS have cancelled facial recognition systems, following
widespread protests. But concerns remain about the use of facial
surveillance technology across the federal agencies by such US companies
as Clearview AI. The absence of a legal framework to implement AI
safeguards and a federal agency to safeguard privacy also raises concerns
about the ability of the US to monitor AI practices.
The White House, OSTP’s Continuing Work on AI Technology and Uses that Can
Benefit Us All (Feb. 3, 2022), https://www.whitehouse.gov/ostp/newsupdates/2022/02/03/ostps-continuing-work-on-ai-technology-and-uses-that-can-benefitus-all/
1916
481
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Uruguay
National AI Strategy
Strictly speaking, Uruguay does not have a general national AI
development strategy, but rather has an AI development strategy for the
digital government, officially called Artificial Strategy for Digital
Government (ASDG) 2020.1917 The preparation of the ASDG was led by
the Agency for Development of Electronic Government and Information
Society (AGESIC).1918 The ASDG is a part of the Digital Government
Agenda1919 and only focuses on digital government.
So far Uruguay has published four digital government agendas,
namely: Uruguay Digital Agenda 2008-2010, Uruguay Digital Agenda
2011-2015, Uruguay Digital Agenda 2020 and Uruguay Digital Agenda
2021-2025.1920 It should be kept in mind, however, that the issues associated
with AI are considered in the 2020 Digital Government Plan,1921 and were
included in Uruguay Digital Agenda 20201922 and Uruguay Digital Agenda
2021-2125.1923
Uruguay's digital government strategies aim, endorsing the
Declaration of Principles of the World Summit on the Information Society,
to advance implementation of the commitment “to build a people-centered,
inclusive and development-oriented Information Society, where everyone
can create, access, utilize and share information and knowledge, enabling
individuals, communities and peoples to achieve their full potential in
promoting their sustainable development and improving their quality of life,
premised on the purposes and principles of the Charter of the United
Uruguay Government, AGESIC, https://www.gub.uy/agencia-gobierno-electronicosociedad-informacion-conocimiento/comunicacion/publicaciones/estrategia-inteligenciaartificial-para-gobierno-digital/estrategia
1918
Id.
1919
Presidency of Uruguay, AGESIC, Estrategia de Inteligencia Artificial para el
Gobierno Digital, version 2.0, 2020, p.3.
1920
Uruguay Government, Agency for the Development of Electronic Government and
the Information Society, https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/politicas-y-gestion/programas/agenda-digital-del-uruguay
1921
Uruguay Government, Agency for the Development of Electronic Government and
the Information Society, https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/politicas-y-gestion/plan-gobierno-digital-2020
1922
Uruguay Government, Agency for the Development of Electronic Government and
the Information Society, https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/politicas-y-gestion/programas/agenda-digital-del-uruguay
1923
Id.
1917
482
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Nations and respecting fully and upholding the Universal Declaration of
Human Rights."1924
After a public-consultation process, Uruguay finally adopted its
Artificial Intelligence Strategy for the Digital Government (ASDG) in
January 2020’1925 The general objective of the ASDG is to promote and
strengthen the responsible use of AI in Public Administration. ASDG
comprises (I) nine general principles and (II) four pillars, each one with
specific objectives and areas of action.
The general principles are:
1) Purpose: AI must enhance the capabilities of human beings,
complementing them as much as possible, aiming to improve the
quality of people´s life, facilitating processes and adding value
to human activity;
2) General interest: AI-based solutions promoted by the State
should oriented toward protecting the general interest,
guaranteeing inclusion and equity;
3) Respect for human rights: Any technological solution that uses
AI must respect human rights, individual freedoms and
diversity;
4) Transparency: AI solutions used in the public sphere must be
transparent and comply with the regulations in force;
5) Responsibility: Technological solutions based on AI must have
a clearly identifiable person responsible for the actions derived
from the AI solution;
6) Ethics: When the application and/or development of AI-based
solutions present ethical dilemmas, they must be addressed and
resolved by human beings;
7) Added value: AI-based solutions should only be used when
adding value to a process;
8) Privacy by design: AI solutions should consider people´s
privacy from their design stage. Personal data-protection
principles in force in Uruguay are considered strategic
components; and
See Uruguay Digital Agenda 2008-2016 and World Summit on the Information
Society, Declaration of Principles: Building the Information Society: a global challenge
in the new Millennium, Geneva 2003- Tunis 2005,
https://www.itu.int/net/wsis/docs/geneva/official/dop.html
1925
Uruguay Government, Agency for the Development of Electronic Government and
the Information Society, https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/comunicacion/publicaciones/estrategia-inteligencia-artificialpara-gobierno-digital/estrategia
1924
483
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
9) Security: AI developments must comply, from their design, with
the basic principles of information security. The guidelines and
regulations related to cybersecurity in force in Uruguay that
apply to the development of AI are considered components of
this strategy.
The four pillars of the Artificial Intelligence Strategy for the Digital
Government are:
1) AI governance in public administration: Ensure that the
principles and comply with the recommendations outlined in the
strategy;
2) Capacity development for AI: Should focus on training civil
servants in different capacities related to AI;
3) Use and application of AI: Implies generating technical guides
for the good use of AI in public administration, promoting
algorithm transparency and designing specific action plans in
strategic sectors; and
4) Digital citizenship and AI: Prepare citizens to take advantage of
opportunities and face the challenges that AI poses, as well as to
generate the necessary confidence in people to develop and use
new technologies.
Global Partnership on AI
Uruguay is not a member of the Global Partnership on AI, however
it is a member of Digital Nations,1926 the Open Government Partnership,1927
Electronic Government Network of Latin America and the Caribbean.1928
Uruguay also has endorsed the Digital Agenda for Latin America and the
Caribbean eLAC20221929 in the context of Seventh Ministerial Conference
on the Information Society in Latin America and the Caribbean.1930
Global Privacy Assembly
Uruguay has been an accredited member of the Global Privacy
Assembly since 2009, and is represented by its national authorities called
https://www.leadingdigitalgovs.org/about
https://www.opengovpartnership.org/our-members/
1928
REDGEALC, https://www.redgealc.org/
1929
Presidency of Uruguay, https://www.gub.uy/uruguay-digital/politicas-ygestion/agenda-uruguay-digital-2025-agenda-digital-para-america-latina-caribe-2022
1930
Seventh Ministerial Conference on the Information Society in Latin America and the
Caribbean, https://conferenciaelac.cepal.org/7/en/documents/digital-agenda-elac2022
1926
1927
484
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Personal Data Regulatory and Control Unit. The 34th International
Conference was hosted in Canelones, Uruguay in 2012.1931
Public Participation
The Artificial Intelligence Strategy for Digital Government was
submitted to a public-consultation process based on a draft prepared by
Agency for the Development of Electronic Government and Information
Society, (AGESIC).1932 Likewise, AGESIC has published surveys on the
implementation of AI in the public sector on its website.1933
Data protection
The legal framework for personal data in Uruguay is built around
the following regulations:
• The Universal Declaration of Human Rights;
• The American Convention on Human Rights or Pact of San José
de Costa Rica;
• The Constitution of the Oriental Republic of Uruguay,
especially its article 72;
• Convention 108 For the protection of individuals with regards to
the Processing of Personal Data;
• Act N° 18.331 on Personal-Data Protection and Habeas Data
Action (LPDP) of August 11, 2008;
• Decrees N° 664/008 of December 22, 2008 and N° 414/2009 of
August 31, 2009;
• Articles 152 to 156 of Act N° 18.719 of December 27, 2010,
which introduce modifications to Act N° 18.331;
• Article 158 literal B) and C) of Act N° 18.719 of December 27,
2010; and
• Budget Act N° 19.670 (articles 37 to 40) dated October 25, 2019
and its regulatory Decree N° 64/020 dated February 21, 2020.
Global Privacy Assembly, History of the Assembly,
https://globalprivacyassembly.org/the-assembly-and-executive-committee/history-of-theassembly/
1932
AGESIC, Inició la Etapa IV: Al finalizar esta etapa, conocerás el documento final de
la Estrategia de Inteligencia Artificial para el Gobierno Digital,
https://www.gub.uy/participacionciudadana/consultapublica
1933
AGESIC, Iniciativas o experiencias en Inteligencia Artificial en la Administración
Pública (Sept. 21, 2021), https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/comunicacion/publicaciones/iniciativas-experienciasinteligencia-artificial-administracion-publica
1931
485
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Substantively and essentially, data protection in Uruguay is regulated by
Act 18.3311934 and its regulatory Decree N° 414/009, Regulating Law
18.331. Uruguay has a data-protection system that follows EU dataprotection rules, and has regulations that adapt its data-protection system to
the General Data Protection Regulation (Regulation (EU) 2016/679)
(GDPR). On August 21, 2012, the European Commission formally
approved Uruguay’s status as a country providing “adequate protection” for
personal data within the meaning of the European Data Protection Directive
(Article 25(6) of Directive 95/46/EC)1935.
Continuing with the process of full GDPR adaptation, Uruguay
passed Act N° 19.670. This Act includes provisions relating to data
protection that address, among others, the proactive responsibility principle
(which supposes the implementation of appropriate technical and
organizational measures such as privacy by design and privacy by default);
the obligation to designate a data-protection officer, and data-breach
notification rules. These provisions were further developed under the
regulatory Decree 64/0201936, which basically regulates the implementation
and enforcement of the provisions.
Data Protection Authority
Article 31 of the Act 18.331 on Personal-Data Protection and
Habeas Data establishes the Personal Data Regulatory and Control Unit
(URCDP)1937 as the country’s supervisory data-protection authority. The
URCDP is an autonomous entity of the Agency for the Development of
Electronic Government and Information Society.1938
Facial recognition
In November 2020, Uruguay began developing a facialidentification database for public-safety purposes under the Ministry of the
See in the officially Uruguay Acts Register,
https://www.impo.com.uy/bases/leyes/18331-2008
1935
2012/484/EU: Commission Implementing Decision of 21 August 2012 pursuant to
Directive 95/46/EC of the European Parliament and of the Council on the adequate
protection of personal data by the Eastern Republic of Uruguay with regard to automated
processing of personal data (notified under document C(2012) 5704), https://eurlex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32012D0484.
1936
See in the officially Uruguay Acts Register,
https://www.impo.com.uy/bases/decretos/64-2020
1937
Unidad Reguladora y de Control de Datos Personales, https://www.gub.uy/unidadreguladora-control-datos-personales/
1938
https://www.gub.uy/agencia-gobierno-electronico-sociedad-informacionconocimiento/
1934
486
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Interior. According to some civil-society organizations1939 “this system was
approved using the National Budget Act as an ‘omnibus law,’ thus
preventing proper discussion about the issue due to the tight deadlines for
approval of this type of law. Development of this database will be under the
responsibility of the Ministry of the Interior, using the database currently
under the control of the National Directorate of Civil Identification, the
organization in charge of issuing identification cards. The database will
include facial images of adults, first and last names, sex, date of birth,
nationality, and identification card number, as well as issue and expiration
date. The Ministry of the Interior has already purchased automated facial
recognition software and currently has a system of 8,433 cameras
distributed in the country in the 19 country´s departments, in addition to
private surveillance systems. The national government has admitted that the
intended use of this facial-identification database is automated surveillance
using facial-recognition algorithms.”1940
Human Rights
Uruguay is among the countries with a very high level of formal
adherence to the international human rights protection system, as it has
ratified practically all existing international instruments on the matter.
However, according to some reports, Uruguay has serious shortcomings
when it comes to effective compliance with such standards.1941 Likewise, a
2013 report by the Office of the United Nations High Commissioner for
Human Rights mentions the debts that the country maintains related to
certain human-rights categories contained in the treaties ratified by
Uruguay.1942 Impunity for crimes against humanity committed during the
DATYSOC, Organizaciones de la sociedad civil y académicas expresan su
preocupación por reconocimiento facial en el Proyecto de Ley de Presupuesto de
Uruguay (Nov. 17, 2020), https://datysoc.org/2020/11/17/organizaciones-de-la-sociedadcivil-y-academicas-expresan-su-preocupacion-por-reconocimiento-facial-en-el-proyectode-ley-de-presupuesto-de-uruguay/
1940
https://www.comprasestatales.gub.uy/consultas/detalle/id/744940; Uruguay: hacia
una población bajo vigilancia con reconocimiento facial, https://indela.fund/uruguayhacia-una-poblacion-bajo-vigilancia-con-reconocimiento-facial/ and
https://datysoc.org/2020/11/17/organizaciones-de-la-sociedad-civil-y-academicasexpresan-su-preocupacion-por-reconocimiento-facial-en-el-proyecto-de-ley-depresupuesto-de-uruguay/
1941
Institute of Legal and Social Studies of Uruguay (IELSUR) y United Nations
Development Program in Uruguay (PNUD), Estudio sobre armonización Legislativa
conforme a los tratados de derechos humanos ratificados por Uruguay u otras normas
legales con fuerza vinculante 35 (2006).
1942
United Nations, General Assembly, Human Rights Council, National report
submitted in accordance with paragraph 5 of the annex to Human Rights Council
resolution 16/21, Uruguay, A/HRC/WG.6/18/URY/2: chrome1939
487
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
military dictatorship (1973-1985) persists and is still an open issue. In 2020,
Uruguayan NGO Peace and Justice Service (SERPAJ) reported that “35
years after the democratic transition, the scenario of denial of justice for the
victims of crimes against humanity of the dictatorship of Uruguay shows
how the country is still far from developing and implementing satisfactory
and successful public policies regarding the search for justice for serious
human rights violations.”1943
Algorithmic Transparency
Uruguay has not developed specific regulations on algorithmic
transparency but is a signatory of the Convention 108+ (Convention for
protecting individuals with to the Processing of Personal Data), which
includes a broad provision regarding algorithm transparency (art. 9.1.c).1944
In addition, the government, through the Agency for the Development of
Electronic Government and the Information Society and Knowledge,
recently has promoted some studies on the impact of algorithms on decision
making.1945
OECD/G20 AI Principles
Uruguay has not endorsed the OECD/G20 AI principles.
Evaluation
Uruguay has focused its policies on digital government — and it has
done well. Proof of this is that according to the Digital Government of the
United Nations (UN) 2020 global-index report, Uruguay is the Latin
American regional leader and occupies the 26th place globally.1946
Consistent with the above, it chose to design an AI development strategy
for digital government (2020), which constitutes a positive first step toward
creating a more comprehensive AI regulatory framework. Even though
extension://efaidnbmnnnibpcajpcglclefindmkaj/viewer.html?pdfurl=https%3A%2F%2Fw
ww.uprinfo.org%2Fsites%2Fdefault%2Ffiles%2Fdocument%2Furuguay%2Fsession_18__january_2014%2Fa_hrc_wg.6_18_ury_1_e.pdf
1943 SERPAJ, Derechos Humanos en el Uruguay, Informe 2020, p. 22, available in:
http://www.serpaj.org.uy/destacados/presentacion-del-informe-anual-2020-de-serpajderechos-humanos-en-el-uruguay-2020/
https://www.coe.int/es/web/data-protection/-/uruguay-ratfies-convention-108Uruguay Government, Agency for the Development of Electronic Government and the
Information Society, https://www.gub.uy/agencia-gobierno-electronico-sociedadinformacion-conocimiento/comunicacion/publicaciones/guia-para-estudio-impactoalgoritmico
1946 UN, https://publicadministration.un.org/egovkb/en-us/Data/CountryInformation/id/185-Uruguay
1944
1945
488
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Uruguay's AI regulations are in the beginning stages, its legal data
protection system, which is adapted to the GPRD, presents an optimistic
view of the future for the country's AI regulations. However, it is concerning
that the government is promoting a facial-recognition policy without having
express regulations on the matter, especially when the OECD/G20 AI
principles are not yet signed.
489
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
COUNTRY EVALUATIONS
Evaluation Grid
Country
Q1
Q2
Q3
Q4
Q5
Q6
Q7
Q8
Q9
Q10
Q11
Q12
Score
Argentina
Y
P
Y
Y
P
P
Y
P
P
P
Y
Y
9.0
Australia
Y
P
Y
Y
Y
Y
Y
P
P
P
Y
N
9.0
Austria
Y
P
Y
Y
Y
Y
Y
Y
Y
P
Y
N
10.0
Bangladesh
N
N
Y
P
P
P
N
P
P
N
Y
N
4.5
Belgium
Y
P
Y
Y
P
P
Y
P
Y
P
Y
P
9.0
Brazil
Y
P
Y
Y
N
P
P
P
P
N
Y
N
6.5
Canada
Y
Y
Y
Y
Y
Y
Y
Y
P
P
Y
Y
11.0
China
Y
P
Y
N
P
P
N
Y
P
N
Y
N
6.0
Colombia
Y
P
Y
P
Y
P
Y
Y
N
P
Y
P
8.5
Denmark
Y
P
Y
Y
Y
Y
N
Y
P
P
Y
Y
9.5
Dominican
Republic
N
P
Y
P
Y
Y
P
N
N
N
Y
N
5.5
Egypt
Y
P
Y
N
N
Y
N
P
N
N
Y
N
5.0
Estonia
Y
P
Y
Y
Y
Y
P
N
Y
P
Y
N
8.5
Finland
Y
P
Y
Y
P
Y
N
P
Y
N
Y
N
7.5
France
Y
Y
Y
Y
P
P
Y
Y
Y
P
Y
P
10.0
Germany
Y
P
Y
Y
Y
Y
Y
Y
Y
P
Y
Y
11.0
Hong Kong
Y
P
Y
P
P
P
N
Y
P
P
Y
Y
8.5
India
Y
P
Y
P
Y
Y
N
P
N
N
Y
N
6.5
Indonesia
Y
P
Y
P
Y
Y
N
P
N
N
Y
N
6.5
Iran
N
N
Y
N
N
P
N
P
N
N
Y
N
3.0
Ireland
Y
P
Y
Y
Y
Y
P
P
Y
N
Y
P
9.0
Israel
Y
P
Y
Y
P
P
P
P
P
N
N
N
6.0
Italy
Y
P
Y
Y
Y
Y
Y
Y
Y
P
Y
Y
11.0
Japan
Y
Y
Y
Y
Y
Y
Y
Y
P
P
Y
N
10.0
Kazakhstan
N
N
Y
N
N
N
P
N
P
N
Y
N
3.0
Kenya
N
P
Y
P
P
Y
P
N
N
N
Y
N
5.0
490
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Korea
Y
Y
Y
Y
Y
Y
Y
Y
Y
P
Y
P
11.0
Malaysia
N
N
Y
P
N
P
P
Y
N
N
Y
N
4.5
Mexico
Y
P
Y
P
N
N
P
P
N
N
Y
Y
6.0
Netherlands
Y
P
Y
Y
N
Y
P
Y
Y
P
Y
N
8.5
Nigeria
N
P
Y
P
N
N
P
N
N
P
Y
N
4.0
Norway
Y
P
Y
Y
Y
P
Y
Y
Y
Y
Y
N
10.0
Philippines
N
P
Y
P
P
P
Y
P
P
P
Y
Y
7.5
Poland
Y
P
Y
Y
P
P
P
N
Y
N
Y
Y
8.0
Russia
Y
P
Y
N
N
Y
P
Y
P
N
Y
N
6.5
Rwanda
N
P
Y
N
P
N
P
N
N
N
Y
N
3.5
Saudi Arabia
Y
P
P
N
P
Y
P
P
N
N
Y
N
5.5
Singapore
N
P
Y
P
Y
Y
P
P
Y
N
Y
N
7.0
Slovenia
Y
P
Y
Y
Y
Y
N
P
N
N
Y
N
7.0
South Africa
N
P
Y
Y
Y
P
Y
N
Y
P
Y
N
7.5
Spain
Y
P
Y
Y
Y
Y
Y
Y
Y
P
Y
N
10.0
Sweden
Y
P
Y
Y
P
Y
Y
P
Y
N
Y
N
8.5
Switzerland
Y
P
Y
Y
Y
Y
P
Y
Y
N
Y
P
9.5
Taiwan
N
P
Y
Y
P
P
P
Y
N
N
N
N
5.0
Thailand
N
P
Y
N
N
P
P
P
N
N
Y
N
4.0
Turkey
Y
P
Y
N
Y
Y
Y
Y
P
N
Y
N
8.0
UAE
N
P
Y
N
Y
Y
P
Y
P
P
Y
N
7.0
U.K.
Y
P
Y
Y
N
Y
Y
Y
Y
N
Y
Y
9.5
U.S.
Y
P
Y
Y
P
Y
P
Y
P
P
N
N
7.5
Uruguay
N
P
Y
Y
P
P
Y
P
P
P
Y
N
7.0
491
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Country Rankings (2021)
Tier I
Germany (11.0)
Korea (11.0)
Canada (11.0)
Italy (11.0)
Tier II
Austria (10.0)
France (10.0)
Norway (10.0)
Japan (10.0)
Spain (10.0)
Denmark (9.5)
Switzerland (9.5)
UK (9.5)
Argentina (9.0)
Belgium (9.0)
Ireland (9.0)
Australia (9.0)
Tier III
Estonia (8.5)
Columbia (8.5)
Hong Kong (8.5)
Netherlands (8.5)
Sweden (8.5)
Poland (8.0)
Turkey (8.0)
Philippines (7.5)
US (7.5)
Finland (7.5)
South Africa (7.5)
Singapore (7.0)
Slovenia (7.0)
UAE (7.0)
Uruguay (7.0)
Brazil (6.5)
India (6.5)
Indonesia (6.5)
Russia (6.5)
492
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Tier IV
China (6.0)
Israel (6.0)
Mexico (6.0)
Dominican Republic (5.5)
Indonesia (5.5)
Saudi Arabia (5.5)
Taiwan (5.0)
Egypt (5.0)
Kenya (5.0)
Tier V
Bangladesh (4.5)
Malaysia (4.5)
Nigeria (4.5)
Thailand (4.0)
Rwanda (3.5)
Iran (3.0)
Kazakhstan 3.0)
493
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Country Distribution by Tier
TIER I
(High)
TIER 2
Canada
Germany
Italy
Korea
Argentina
Australia
Austria
Belgium
Denmark
France
Ireland
Japan
Norway
Spain
Switzerland
UK
TIER 3
(Middle)
Brazil
Columbia
Estonia
Finland
Hong Kong
India
Indonesia
Netherlands
Philippines
Poland
Russia
South Africa
Singapore
Slovenia
Sweden
Turkey
UAE
United States
Uruguay
494
TIER 4
China
Mexico
Dominican
Republic
Egypt
Indonesia
Israel
Kenya
Saudi
Arabia
Taiwan
TIER 5
(Low)
Bangladesh
Iran
Kazakhstan
Malaysia
Nigeria
Rwanda
Thailand
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Chart: AIDV Index by Country and Tier
495
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Chart: AIDV Index by Country, 2020 vs. 2021
496
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Chart: AIDV Country Scores by Individual Metrics
497
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Metrics
Q1. Has the country endorsed the OECD AI Principles?
Q2. Is the country implementing the OECD AI Principles?
Q3. Has the country endorsed the Universal Declaration of Human
Rights?
Q4. Is the country implementing the Universal Declaration for
Human Rights?
Q5. Has the country established a process for meaningful public
participation in the development of a national AI Policy?
Q6. Are materials about the country’s AI policies and practices
readily available to the public?
Q7. Does the country have an independent (agency/mechanism)
for AI oversight?
Q8. Do the following goals appear in the national AI policy:
“Fairness,” “Accountability,” “Transparency,” (“Rule of Law,”)
(“Fundamental Rights”)? [implementation? = legal force? =
enforcement?]
Q9. Has the country by law established a right to Algorithmic
Transparency? [GDPR? / COE+?]
Q10. Has the country supported the Universal Guidelines for AI?
Q11. Has the country endorsed the UNESCO Recommendation on
AI Ethics?
Q12: Has the country’s Data Protection Agency sponsored the
2018 GPA Resolution on AI and Ethics and the 2020 GPA
Resolution on AI and Accountability?
Response Codes
“Y” – Yes
“N” – No
“P” – Partly
498
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
METHODOLOGY
Scope
We assessed the AI policies and practices of the top 40 countries by
GDP. We also looked at 10 other countries we considered “high-impact.”
Our aim in this first survey was to examine those countries likely to have
the greatest policy impact in the AI field. We considered also influential
intergovernmental organizations, such as the institutions of the European
Union, the OECD and G20, but we did not attempt to evaluate their AI
policies.
Time Period
The research for the 2020 edition of the report was undertaken in
late 2020 for publication in mid-December 2020. For the current edition of
the report, published in mid-February 2022, we continued to gather
information throughout 2021 and into early 2022.
Factors
We identified 12 factors to assess national AI policies and practices.
The factors reflect well known frameworks for AI policy (the OECD/G20
AI Principles), human rights (the Universal Declaration for Human Rights),
and democratic decision-making (transparency, public participation, and
access to policy documents). We highlighted key themes for AI policy,
including algorithmic transparency and accountability. We also included
aspirational goals set out in the Universal Guidelines for AI as well as
support for the UNESCO Recommendation on AI Ethics, adopted in 2021.
On certain factors, we deferred to well established legal frameworks
and well-known international organizations. For example, countries within
the European Union are subject to the General Data Protection Regulation
which provides certain rights to those who are subject to automated
decision-making, including access to the underlying logic of an algorithm.
The Council of Europe Modernized Convention 108 provides similar legal
rights regarding AI. On general human rights assessments, we deferred to
the reports of Freedom House, Human Rights Watch, and Amnesty
International. We also recognized those countries that endorsed the
resolution on AI and Accountability, adopted by the Global Privacy
Assembly, the global association of leading privacy experts and officials.
On the issue of implementation, we recognize that it is difficult to
assess empirically progress toward AI policy goals, particularly when the
499
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
underlying objective is not measured in quantitative terms, such as would
be the case for research investment, papers published, or patents obtained.
Nonetheless we believe this must be a key component of the evaluation. We
turned first to the OECD, which has begun a process to track
implementation of the OECD AI Principles. The OECD published
substantial reports in 2020 and 2021 on implementation of the OECD AI
Principles.1947 The OECD has also encouraged member states to provide
overviews on national AI strategies to the OECD AI Group of Experts,
though at present these reports are not generally available to the public.
We looked next at national developments, both favorable and
controversial, concerning the implementation of AI policy. We consulted
official sources but also reviewed independent sources, such as news
sources, agencies, and thinks thanks not directly aligned with national
governments, for these assessments.
Finally, because AI policy is in the early days, there is far more
information about what governments intend to do than what they have done.
We encourage governments to establish independent agencies with annual
public reporting requirements to provide information about progress toward
national goals and compliances with international policy frameworks. Such
reports could provide the basis for future comparative evaluations.
The Metrics
Q1. Has the country endorsed the OECD/G20 AI Principles?
The OECD/G20 AI Principles are the first global framework for AI
policy. Endorsement of these principles provides a baseline to determine a
country’s compliance with international AI policy norms. Countries that
have endorsed the OECD/G20 AI Principles fall into three categories: (1)
OECD Member Countries, (2) Non-member OECD Countries that endorsed
the OECD AI Principles, and (3) G-20 Member countries that subsequently
endorsed the G20 AI Principles which follow closely the original OECD AI
Principles.1948
Determinations in this category are essentially binary: a country
has either endorsed the OECD/G20 AI Principles or it has not.
OECD, State of Implementation of the OECD AI Principles (June 18, 2021),
https://www.oecd.org/digital/state-of-implementation-of-the-oecd-ai-principles1cd40c44-en.htm
1948
The G20 AI Principles directly restate the value-based principles in Part I of the
OECD AI Principles
1947
500
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Q2. Is the country implementing the OECD AI Principles?
Endorsement alone of the AI Principles is not sufficient to determine
a country’s AI practices. The OECD itself has begun a process to track
implementation of the AI Principles, but the reporting to date is mostly
anecdotal and inconclusive. We begin our analysis of implementation with
the OECD reporting and then look to other sources, including government
documents, news articles and NGO reports, to assess implementation.
Determinations in this category are more nuanced: some countries
have called attention to their efforts to implement the OECD/G20 AI
principles. Others have done so in practice without explicit references to the
AI Principles. We have made reasonable efforts to identify national projects
that implement the OECD/G20 AI Principles, based on reporting from the
OECD, but information is often difficult to find. In some instances, we were
able to acknowledge partial implementation (P). In 2020, we concluded that
no country has fully implemented the OECD/G20 AI Principles and
therefore no country received a Y determination. In 2021, we have chosen
to recognize the leading role of four countries in the development and
implementation of the OECD AI Principles – Canada, France, Korea, and
Japan.
Q3. Has the country endorsed the Universal Declaration of Human Rights?
In the human rights field, the Universal Declaration of Human
Rights is the most well-known and widely adopted legal framework for the
protection of fundamental rights. Although the UDHR preceded the rise of
Artificial Intelligence, we anticipated that many of the significant policy
debates ahead will be grounded in principles set out in the Universal
Declaration. For this reason, we propose endorsement of the UDHR as a
second baseline to assess country AI policies and practices.
Determinations in this category are essentially binary: a country has
either endorsed the UDHR or it has not. The one notable exception is Saudi
Arabia which did not endorse the UDHR but is a member of the United
Nations and has recognized, according to human rights organizations,
certain human rights obligations.
Q4. Is the country implementing the Universal Declaration for Human
Rights?
Like the question regarding implementation of the OECD AI
Principles, measuring implementation of the UDHR is not a simple task.
Several well-established international organizations, such as Freedom
House and Human Rights Watch, have developed formal metrics to
evaluate compliance with human rights norms. We defer to these
organizations for the evaluation of general human rights practices, while
501
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
also noting that several of these factors may be useful in future evaluation
of AI practices.
Determinations in this category typically fell into two categories: Y,
a country widely recognized for its defense of human rights as generally
understood by reference to the UDHR, and P, a country in partial
compliance with human rights obligations. In 2021, we made this
determination more precise. Countries that Freedom House designated as
“Free” received Y. A country designated “Partly Free” was designated “P”
and countries designated “Note Free” were designated “N.”
Q5. Has the country established a process for meaningful public
participation in the development of a national AI Policy?
Almost every country in our report has set out a national AI strategy
or action plan. We have attempted to fairly summarize and present these
initiatives. But we are also interested in the development of these policies.
Was there an opportunity for public participation? Was there a formal
consultation process? Do the national AI policies reflect the views of those
who may be impacted by the deployment of AI techniques? And is there an
ongoing mechanism for public participation as national AI policies evolve?
Determinations in this category were based on our ability to identify
opportunities for meaningful public participation. The distinction between
a Y and P in this category reflected the quality of the opportunity for public
participation.
Q6. Are materials about the country’s AI policies and practices readily
available to the public?
Effective public participation requires public access to relevant
documents. Has the national government taken steps to ensure that
documents concerning AI policy are readily available, complete, and
accurate? Are the materials available on the website of a public agency or
are they maintained by a private company? Are there opportunities for
future comment?
The determinations in this category often aligned with the
determinations about public participation. We respect the practice of
countries to publish reports, and to seek public reports, in the national
language. We note however that the absence of an English translation may
make independent evaluation of a country’s AI policies and practices more
difficult. We discuss the issue of Language in more detail below.
502
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Q7. Does the country have an independent (agency/mechanism) for AI
oversight?
All governments understandably seek to advance national AI
priorities. And most governments have directed a science or industry
ministry to lead national efforts. But the deployment of AI techniques also
raises concerns about accountability, privacy and data protection, fairness,
transparency, and equity. For these reasons, we look to determine whether
countries have independent agencies, such as a data protection agency, a
human rights commission, or an AI ethics commission, to protect
fundamental rights.
Determinations in this category were based on the actual
establishment of mechanisms to oversee or guide AI practices. Again, the
difference between a Y and a P determination reflected the quality and
breadth of the oversight mechanisms.
Q8. Do the following goals appear in the national AI policy: “Fairness,”
“Accountability,” “Transparency,” “Rule of Law,” “Fundamental Rights”?
There are many themes in the AI policy realm. We identified these
five goals as the most significant. They appear frequently in AI policy
frameworks and they are grounded in law. We recognize that countries that
have endorsed the OECD/G20 AI Principles have, by implication, endorsed
these goals. But this question asks whether countries have explicitly
endorsed these goals in their national AI strategies.
Determinations in this category attempt to evaluate the extent to
which a country has prioritized these AI policy goals. Full endorsement
received a Y, partial endorsement a P.
Q9. Has the country by law established a right to Algorithmic Transparency?
One of the most significant AI policy issues today is Algorithmic
Transparency. We take the position that individuals should have the right to
access the logic, the factors, and the data that contributed to a decision
concerning them. This right is currently established in two legal
frameworks: The General Data Projection Regulation of the European
Union (Article 22) and the Council of Europe Convention 108+, the
modernized Privacy Convention (Article 9). Countries that are within the
EU and/or signatories to COE 108+ have therefore established this right.
We have also considered whether countries, by national law, have
established the right to algorithmic transparency.
For determinations in this category, we assigned a Y to those
countries that are subject to the GDPR and/or the Council of Europe
Convention. In a subsequent review, we will investigate whether countries
503
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
have implemented a right to algorithmic transparency. This will provide a
more detailed assessment of this key metric.
Q10. Has the country supported the Universal Guidelines for AI?
In 2018, more than 60 organizations, including leading scientific
societies, and 300 experts from over 40 countries endorsed the Universal
Guidelines for AI. The Universal Guidelines go beyond the OECD/G20 AI
Principles and establish “red lines” for certain AI practices, such as the
scoring of citizens, criminal sentencing, and facial recognition for mass
surveillance. Although there is no formal mechanism for countries to
endorse the UGAI, we are interested in whether countries have adopted
principles, and recognized red lines for AI, that go beyond the OECD/G20
AI Principles. Efforts to prohibit face surveillance or social scoring, for
example, reflect the spirit of the UGAI.
For determinations in this category, we could not assign a Y to any
country, but we did assign a P for countries that have specifically limited
certain AI applications. Countries that have done little to develop AI
policies likely received a N determination.
Q11. Has the country endorsed the UNESCO Recommendation on AI Ethics?
In November 2021, UNESCO member states adopted the first ever
global agreement on the Ethics of Artificial Intelligence. We consider this
a watershed moment in the development of AI policies and have
incorporated country support for the UNECO framework as a positive
indicator for national AI policies and practices. Determinations in this
category are similar to those regarding endorsement of the OECD AI
Principles and the UN Declaration of Human Rights, and not an indication
of implementation of the framework. We note that that Hong Kong, Israel,
Taiwan, and the United States were not among the signatories for reasons
unrelated to their views on AI policies and practices.
[Note on Methodology: in AIDV-2020 we asked in Q11 whether countries
supported the Social Contract for AI, which we described as “aspirational
goals for the Age of AI that go beyond the OECD/G20 AI Principles.” In
our assessment, country support for the UNESCO Recommendation on AI
Ethics constitutes a similar metric that is also more easily determined, at
least with regard to initial support. Nonetheless, this constitutes a change in
the methodology originally established, which we fully acknowledge.]
504
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Q12: Has the country’s Data Protection Agency endorsed the 2018 GPA
Resolution on AI and Ethics and the 2020 GPA Resolution on AI and
Accountability?
In the fall of 2018, the Global Privacy Assembly (then known as the
International Conference of Data Protection and Privacy Commissioners),
adopted a foundational Declaration on Ethics and Data Protection in
Artificial Intelligence.1949 The 2018 Declaration emphasized fairness,
vigilance, transparency and intelligibility, and measures to reduce unlawful
bias and discrimination. In 2020, the GPA adopted a resolution on AI and
Accountability.1950 That resolution sets out a dozen steps for AI
accountability, including the preparation of human rights impact
assessments.
We believe that support for these resolutions is an important
indicator of a country’s commitment to AI and data protection and effective
implementation of AI policy goals. We checked to see which countries
explicitly sponsored the resolutions. We will also consider other notable
initiatives in future global surveys of AI policies and practices.
For determinations in this category, we assigned a Y to countries
that sponsored both resolutions, an N to countries that sponsored neither (or
are not represented at the GPA), and P to the countries that sponsored only
resolution
As an aside to the Global Privacy Assembly, we would recommend
new mechanisms that would allow members to endorse resolutions
concerning AI in subsequent years. We will update country ratings
accordingly.
Hong Kong
Hong Kong constitutes a special case in our review of national AI
policies and practices. Although Hong Kong, an Administrative Region of
the People’s Republic of China (HKSAR) is not a country, it ranks number
37 in the world for GDP, placing it within the top 40 countries we reviewed
for the 2021 index. Hong Kong also has an active data protection agency
that has contributed to the formulation of the AI policies endorsed by the
Global Privacy Assembly. As Hong Kong is not a country it could not enter
ICPDPC, Declaration on Ethics and Data Protection in Artificial Intelligence
(including list of authors and co-sponsors) (Oct. 23, 2018),
http://globalprivacyassembly.org/wp-content/uploads/2018/10/20180922_ICDPPC40th_AI-Declaration_ADOPTED.pdf
1950
Global Privacy Assembly, Adopted Resolution on Accountability in the Development
and Use of Artificial Intelligence (including list of main sponsors and co-sponsors) (Oct.
2020), https://globalprivacyassembly.org/wp-content/uploads/2020/11/GPA-Resolutionon-Accountability-in-the-Development-and-Use-of-AI-EN.pdf
1949
505
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
into international agreements, such as the OECD AI Principles. So, we
treated commitments made by China to these polices as if they were made
by Hong Kong. At the same time, we recognized variances in AI practices
in the Region as well as the different ratings for compliance with the human
rights norms, as determined by Freedom House.
Exemplars
In developing the methodology, we also created a list of exemplar
countries for several metrics. For example, on Question 5, concerning
meaningful public participation, we were struck by the high level of public
engagement in Switzerland. On Question 6, concerning the availability of a
countries AI policies and practices, Germany’s Plattform Lernende Systeme
offers a map that shows, by region, AI developments across the country.
And the multiple agencies in France, the CNIL and the Defender of Rights,
provide a very good example of independent oversight for AI, highlighted
by Question 7.
Scoring
We assigned a numeric value of 1.0 to each “Y” answer, 0.5 to each
“P” answer, and 0.0 to each “N” or “U” answer. (We may revise scores
upward for U answers upon receipt of evidence regarding progress toward
the specific metric). We then tallied the numbers, weighing each metric
equally, and produced a total score. A top score would be 12, a bottom score
is 0. On the basis of total scores, we grouped countries by color gradation
and then into tiers. The groupings reflect a normalized distribution with
Yellow or Tier III as the median.
Search Strategy
To locate relevant policy materials, we conducted extensive online
searches. Key search terms, often used in combination with “AI” or
“Artificial Intelligence,” included: “Accountability,” “Algorithmic
Transparency,” “Data Protection,” “Digital,” “Ethical,” “Ethics,”
“Fairness,” “Governance,” “Law,” “Legislation,” “Policy,” “Poll,”
“Privacy,” “Regulation, “Strategy,” and “Technology.”
Descriptive Summary
Each country report includes a descriptive summary labelled
“Evaluation.” The evaluation does not precisely track the metrics. It is
506
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
intended to highlight the key findings in the country report and provide a
general overview for the reader.
Language
Our research team has language expertise in English, French,
German, Italian, Japanese, Korean, Mandarin, Russian, Spanish, Turkish,
Uyghur, and Vietnamese. However, we preferred English publications, as
they would be considered authoritative original sources or authoritative
translations for the international community from the original sources. In
some instances, we translated text from non-English to English with a
Machine Translation (“[MT]”) tool, such as DeepL Translate (“[DT]”) or
Google Translate (“[GT]”). We noted such instances in the citations.
Citation Format
We adopted a simplified citation format for the AI Social Index
2020. Each citation includes the author and title of the publication. Where
there are multiple authors, we provided the name of the institution if
available but not the names of the authors. We include also a date where
there was a final publication date. By way of contrast, cites to websites do
not include dates. And we included URLs, which we made transparent so
that the reader could quickly assess the source. In a paragraph where there
may be multiple references to the same source, we cited to the source in the
first instance, but not in subsequent instances unless there was an
intervening reference to a different source.
Gender Balance and Diversity
In the development of the AI and Democratic Values Index, the
selection of team members and reviewers, we strived to maintain gender
balance. We have also tried to promote diversity and regional
representation.
Bias
We did not explicitly examine the issue of bias in AI, although this
is a widely discussed topic and the focus of extensive research, including
the bias of data sets. Our view is that the most effective policy response to
the problem of bias is the explicit recognition of Fairness, Accuracy, and
Transparency in AI policy and the implementation of these principles in AI
practices. Several questions in the AI and Democratic Values Index (Q1,
507
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Q2, Q7, Q9, Q10, Q11, Q12) make these factors key metrics for the
evaluation of a nation’s AI policies.
We also recognize the inherent bias in the construction of all
surveys, including in the survey focus, the framing of questions, and the
research methodology.1951
Private Sector Practices
We did not attempt to review or evaluate the practices of private
firms or organizations. The AI and Democratic Values Index attempts only
to evaluate the policies and practices of national governments. We do
believe that private firms must act in compliance with law and through
democratic institutions, and that the evaluation of government policies must
ultimately be the measure of private sector practices.1952
Max Weber, Objectivity of Social Science and Science Policy (1904).
Further discussion of the methodology underlying the AI Index is presented in Marc
Rotenberg, Time to Assess National AI Policies, Blog@CACM (Nov. 24, 2020),
https://cacm.acm.org/blogs/blog-cacm/248921-time-to-assess-national-ai-policies/fulltext
1951
1952
508
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
REFERENCES
AI Policy Surveys - International and Regional
Access Now, Mapping Regulatory Proposals for Artificial Intelligence in
Europe (November 2018),
https://www.accessnow.org/cms/assets/uploads/2018/11/mapping_regulat
ory_proposals_for_AI_in_EU.pdf
AI Ethicist, National AI Strategies (Includes both National and
Multinational Strategies), https://www.aiethicist.org/national-strategies
Algorithm Watch, Automating Society 2020 (October 2020),
https://automatingsociety.algorithmwatch.org/wpcontent/uploads/2020/10/Automating-Society-Report-2020.pdf
Anna Jobin, Marcello Ienca and Effy Vayena, The global landscape of AI
ethics guidelines, Nature Machine Intelligence (Sept. 2019),
https://www.nature.com/articles/s42256-019-0088-2
Council of Europe, Graphical visualisation of the distribution of strategic
and ethical frameworks relating to artificial intelligence (Aug. 12, 2020),
https://www.coe.int/en/web/artificial-intelligence/-/graphicalvisualisation-of-the-distribution-of-strategic-and-ethical-frameworksrelating-to-artificial-intelligence
European Commission, AI Watch (2020),
https://ec.europa.eu/knowledge4policy/ai-watch_en
Future of Life, National and International AI Strategies,
https://futureoflife.org/national-international-ai-strategies/
G20 Digital Economy Task Force, Examples of National AI Policies
(2020), https://www.mcit.gov.sa/sites/default/files/examples-of-ainational-policies.pdf
Library of Congress, Regulation of Artificial Intelligence in Selected
Jurisdictions (Jan. 2019), https://www.loc.gov/item/2019668143/
Marc Rotenberg, AI Policy Sourcebook (2020),
https://epic.org/bookstore/ai2020/
509
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD, AI Policy Observatory (2022), https://oecd.ai
OECD, State of implementation of the OECD AI Principles, Insights from
national AI policies (2021), https://www.oecd.org/digital/state-ofimplementation-of-the-oecd-ai-principles-1cd40c44-en.htm
Oxford Insights, The Government AI Readiness Index 2021,
https://www.oxfordinsights.com/government-ai-readiness-index-2021
Roger Clark, Principles for AI: A Sourcebook (Aug. 26, 2020)
http://www.rogerclarke.com/EC/GAIP.html
AI Policy Surveys - Domestic
Australia, Department of Industry, Science, Energy and Resources, AI
Ethics Framework, https://www.industry.gov.au/strategies-for-thefuture/artificial intelligence
Canada, CIFAR, Pan-Canadian Artificial Intelligence Strategy,
https://www.cifar.ca/ai/pan-canadian-artificial-intelligence-strategy
China,《高等学校人工智能创新行动计划》
http://www.moe.gov.cn/jyb_xwfb/xw_fbh/moe_2069/xwfbh_2018n/xwfb
_20180608/201806/t20180608_338911.html
China,《机器人产业发展规划》（2016-2020年）
https://www.ndrc.gov.cn/xxgk/zcfb/ghwb/201604/t20160427_962181.htm
l
Belgium, AI4, AI4Belgium Strategy, https://www.ai4belgium.be/wpcontent/uploads/2019/04/report_en.pdf
Germany, The Federal Government of Germany, Artificial Intelligence
Strategy (Nov. 2018), https://www.ki-strategiedeutschland.de/home.html?file=files/downloads/Nationale_KIStrategie_engl.pdf
India, Department of Promotion of Industry and Internal Trade, Report of
Task Force on Artificial Intelligence (Mar. 2018),
https://dipp.gov.in/whats-new/report-task-force-artificial-intelligence
510
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Italia, Ministry of Economic Development, Strategia Nazionale per
l’Intelligenza Artificiale (2019),
https://www.mise.gov.it/images/stories/documenti/Strategia-NazionaleIntelligenza-Artificiale-Bozza-Consultazione.pdf
Japan, Prime Minister’s Office, AI Strategy 2019: AI for Everyone:
People, Industries, Regions and Governments (June 11, 2019),
https://www.kantei.go.jp/jp/singi/ai_senryaku/pdf/aistratagy2019en.pdf
Korea, HRST Policy Platform, AI R&D Strategy (May 2018),
https://hrstpolicy.re.kr/kistep/kr/policy/policyPlanKorDetail.html?board_s
eq=26570&board_class=BOARD01&rootId=2003000&menuId=2003102
Korea, Ministry of Science and ICT, Policies, National Strategy for
Artificial Intelligence (Mar. 23, 2020),
https://www.msit.go.kr/english/msipContents/contentsView.do?cateId=tst
60&artId=2771576
Mexico, IA2030Mx, Artificial Intelligence in Mexico: A National Agenda
(Nov. 2020) (English translation), https://www.ia2030.mx/
Netherlands, The Strategic Action Plan for Artificial Intelligence (2019),
https://www.government.nl/binaries/government/documents/reports/2019/
10/09/strategic-action-plan-for-artificial
intelligence/Strategic+Action+Plan+for+Artificial+Intelligence.pdf
Russia, GMIS, Artificial Intelligence: A Strategy for Russian start-up
(June 11, 2019), https://gmisummit.com/wp-content/uploads/2019/06/AiA-startegy-for-russian-startups.pdf
Rwanda, ID4Africa, Rwanda National ID Strategy,
https://www.id4africa.com/2019_event/presentations/PS1/5-JosephineMukesha-NIDA-Rwanda.pdf
Saudi Arabia, SDAIA, National Strategy for Data & AI: Realizing our
best tomorrow (Oct. 2020),
https://ai.sa/Brochure_NSDAI_Summit%20version_EN.pdf
Singapore, National Artificial Intelligence Strategy (2019):
https://www.smartnation.gov.sg/why-Smart-Nation/NationalAIStrategy
Spain, Government of Spain & Ministry of Science, Innovation, and
Universities, Spanish, RDI Strategy in Artificial Intelligence (2019),
511
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
https://www.ciencia.gob.es/stfls/MICINN/Ciencia/Ficheros/Estrategia_Int
eligencia_Artificial_EN.PDF
Sweden, Government Offices of Sweden, National Approach to Artificial
Intelligence (May 2018),
https://www.government.se/4a7451/contentassets/fe2ba005fb4943358757
4c513a837fac/national-approach-to-artificial-intelligence.pdf
Switherland, The Conference toward AI Network Society, Draft AI R&D
GUIDELINES for International Discussions (July 28, 2017),
https://www.soumu.go.jp/main_content/000507517.pdf
Turkey, Presidency of the Republic of Turkey, Digital Transformation
Office, Workshops on National AI Strategy and AI Institute Organized
(Feb. 19, 2020), https://cbddo.gov.tr/en/news/4701/ulusal-yapay-zekastratejisi-ve-yapay-zeka-enstitusu-calistaylari-duzenlendi
United Kingdom, HM Government, Industrial Strategy: Artificial
Intelligence Sector Deal
https://assets.publishing.service.gov.uk/government/uploads/system/uploa
ds/attachment_data/file/702810/180425_BEIS_AI_Sector_Deal__4_.pdf
United States, Congresswoman Robin Kelly, ICYMI: Kelly, Hurd Call for
Creation of National AI Strategy (Sept. 18, 2020),
https://robinkelly.house.gov/media-center/press-releases/icymi-hurd-kellycall-for-creation-of-national-ai-strategy
Related Country Surveys
Cryptography and Liberty: An International Survey of Encryption Policy
(EPIC, editions 1999-2004)
Privacy and Human Rights: An International Survey of Privacy Laws and
Practices (EPIC and Privacy International, editions 1998-2006)
David Banisar, National Comprehensive Data Protection/Privacy Laws
and Bills 2021 (Aug. 30, 2021), https://ssrn.com/abstract=1951416
GLOSSARY
ACM
AEPD
Association for Computing Machinery
Agencia Española de Protección de Datos (ESP)
512
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI
AIHLEG
AIDP
AIIA
AIIS
Artificial Intelligence
AI High Level Expert Group (EU)
Artificial Intelligence Development Plan (CHN)
Artificial Intelligence Industry Alliance (CHN)
Artificial Intelligence and Intelligent Systems Laboratory
(ITA)
AIRA
African Internet Rights Alliance
AIRC
AI Research Center (CHN)
AJL
Algorithmic Justice League
ANPD
Autoridade Nacional de Proteção de Dados (BRA)
AISCI
AI Social Contract Index
AIWS
AI World Society
APPI
Act on the Protection of Personal Information (JAP)
BAAI
Beijing Academy of Artificial Intelligence (CHN)
BEUC
European Consumer Organization
BGF
Boston Global Forum
BRI
Belt and Road Initiative (CHN)
C4AI
Artificial Intelligence Center (BRA)
CAHAI Ad Hoc Committee on Artificial Intelligence (COE)
CAIDP Center for AI and Digital Policy
CAS
Criminaliteits Anticipatie Systeem (NLD)
CCNE
National Consultative Committee on Bioethics (FRA)
CDEI
Center for Data Ethics and Innovation (GBR)
CEPEJ
European Commission for the Efficiency of Justice (COE)
CIFAR
Canadian Institute for Advanced Research (CAN)
CINI
Consortium for Informatics (ITA)
CJEU
Court of Justice of the European Union (EU
CLAIRE Confederation of Artificial Intelligence Laboratories in Europe
CNAM Council of the Caisse nationale d’assurance maladie (FRA)
CNIL
Commission Nationale de l’Informatique et des Libertés (FRA)
CNJ
Conselho Nacional de Justiça (BRA)
COE
Council of Europe
COMEST Commission on the Ethics of Scientific Knowledge and
Technology
CPSR
Computer Professionals for Social Responsibility
CSIRO
National Science Agency (AUS)
DFFT
Data Free Flows with Trust
DIGG
Agency for Data Administration (SWE)
DPIA
Data Protection Impact Assessments
DPA
Data Protection Agency
DT
DeepL Translate
513
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
DTO
EAD
EDPS
EDRi
FDPIC
FRA
FREMP
GDPR
GGE
GPA
GPAI
HDH
HLEG
IACI
ICCPR
IEEE
IJOP
IMDA
ICO
ITU
JSC
KIC
KKVK
LAWS
LIBE
LGPD
MCTIC
MDES
MDI
MEITY
MOST
NCAI
NCPO
NDMO
NHRI
NIC
NIDA
NIN
Digital Transformation Office (TUR)
Ethically Aligned Designed
European Data Protection Supervisor (EU)
European Digital Rights Initiative
Federal Data Protection and Information Commissioner (CHE)
Fundamental Rights Agency (COE / EU)
Working Party on Fundamental Rights, Citizens Rights and
Free Movement of Persons (EU)
General Data Protection Regulation (EU)
Group of Government Experts (GGE)
Global Privacy Assembly
Global Partnership on Artificial Intelligence
Health Data Hub (FRA)
High Level Expert Group
Innovation Center for AI (NLD)
International Convention on Civil and Political Rights
Institute of Electrical and Electronics Engineers
Integrated Joint Operations Platform (CHN)
Infocomm Media Development Authority (SGP)
Information Commission Office (GBR)
International Telecommunications Union
Jakarta Smart City
Kigali Innovation City (RWA)
Data Protection Authority (TUR)
Lethal Autonomous Weapon Systems
European Parliament Committee on Civil Liberties, Justice and
Home Affairs
Lei Geral de Proteção de Dados Pessoais (BRA)
Ministry of Science, Technology, Innovations and
Communications (BRA)
Ministry of Digital Economy and Society (THA)
Michael Dukakis Institute for Leadership and Innovation
Ministry of Electronics and Information Technology (IND)
Ministry of Science and Technology (TWN)
National Center for AI (KSA)
National Council for Peace and Order (THA)
National Data Management Office (KSA)
National Human Rights Institute
National Information Center (KSA)
National Identification Agency (RWA)
National Identity Number (RWA)
514
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
NPCDE
NSCAI
NXPO
National Pilot Committee for Digital Ethics (FRA)
National Security Commission on AI (USA)
Office of National Higher Education Science Research and
Innovation Policy Council (THA)
OAI
Office of Artificial Intelligence (GBR)
OGP
Open Government Partnership
OHCHR Office of the High Commissioner for Human Rights
PAI
Policies for AI (OECD)
PDPA
Personal Data Protection Act
PDPC
Personal Data Protection Commission
PIPC
Personal Information Protection Commission (KOR)
PIPEDA Personal Information Protection and Electronic Documents Act
(CAN)
PLA
People’s Liberation Army (CHN)
PPC
Personal Information Protection Commission (JAP)
RIPD
Red Iberoamericana de Protección de Datos
SCAAI Social Contract for the Age of AI
SDAIA Saudi Data and Artificial Intelligence Authority (KSA)
SDG
Sustainable Development Goals (UN)
SFLC
Software Freedom Law Center (IND)
UDHR
Universal Declaration of Human Rights
UGAI
Universal Guidelines for AI
UNICRI United Nations Interregional Crime and Justice Research
Institute
USRC
Unmanned Systems Research Center (CHN)
515
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
REFERENCE DOCUMENTS
OECD AI Principles
Recommendation of the Council on Artificial Intelligence
Adopted May 21, 2019
THE COUNCIL,
HAVING REGARD to Article 5 b) of the Convention on the Organisation
for Economic Co-operation and Development of 14 December 1960;
HAVING REGARD to the OECD Guidelines for Multinational Enterprises
[OECD/LEGAL/0144]; Recommendation of the Council concerning
Guidelines Governing the Protection of Privacy and Transborder Flows of
Personal Data [OECD/LEGAL/0188]; Recommendation of the Council
concerning Guidelines for Cryptography Policy [OECD/LEGAL/0289];
Recommendation of the Council for Enhanced Access and More Effective
Use
of
Public
Sector
Information
[OECD/LEGAL/0362];
Recommendation of the Council on Digital Security Risk Management for
Economic
and
Social
Prosperity
[OECD/LEGAL/0415];
Recommendation of the Council on Consumer Protection in E-commerce
[OECD/LEGAL/0422]; Declaration on the Digital Economy: Innovation,
Growth
and
Social
Prosperity
(Cancún
Declaration)
[OECD/LEGAL/0426]; Declaration on Strengthening SMEs and
Entrepreneurship
for
Productivity
and
Inclusive
Growth
[OECD/LEGAL/0439]; as well as the 2016 Ministerial Statement on
Building more Resilient and Inclusive Labour Markets, adopted at the
OECD Labour and Employment Ministerial Meeting;
HAVING REGARD to the Sustainable Development Goals set out in the
2030 Agenda for Sustainable Development adopted by the United Nations
General Assembly (A/RES/70/1) as well as the 1948 Universal Declaration
of Human Rights;
HAVING REGARD to the important work being carried out on artificial
intelligence (hereafter, “AI”) in other international governmental and nongovernmental fora;
RECOGNISING that AI has pervasive, far-reaching and global
implications that are transforming societies, economic sectors and the world
of work, and are likely to increasingly do so in the future;
RECOGNISING that AI has the potential to improve the welfare and wellbeing of people, to contribute to positive sustainable global economic
activity, to increase innovation and productivity, and to help respond to key
global challenges;
516
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
RECOGNISING that, at the same time, these transformations may have
disparate effects within, and between societies and economies, notably
regarding economic shifts, competition, transitions in the labour market,
inequalities, and implications for democracy and human rights, privacy and
data protection, and digital security;
RECOGNISING that trust is a key enabler of digital transformation; that,
although the nature of future AI applications and their implications may be
hard to foresee, the trustworthiness of AI systems is a key factor for the
diffusion and adoption of AI; and that a well-informed whole-of-society
public debate is necessary for capturing the beneficial potential of the
technology, while limiting the risks associated with it;
UNDERLINING that certain existing national and international legal,
regulatory and policy frameworks already have relevance to AI, including
those related to human rights, consumer and personal data protection,
intellectual property rights, responsible business conduct, and competition,
while noting that the appropriateness of some frameworks may need to be
assessed and new approaches developed;
RECOGNISING that given the rapid development and implementation of
AI, there is a need for a stable policy environment that promotes a humancentric approach to trustworthy AI, that fosters research, preserves
economic incentives to innovate, and that applies to all stakeholders
according to their role and the context;
CONSIDERING that embracing the opportunities offered, and addressing
the challenges raised, by AI applications, and empowering stakeholders to
engage is essential to fostering adoption of trustworthy AI in society, and to
turning AI trustworthiness into a competitive parameter in the global
marketplace;
On the proposal of the Committee on Digital Economy Policy:
I. AGREES that for the purpose of this Recommendation the following
terms should be understood as follows:
‒AI system: An AI system is a machine-based system that can, for a given
set of human-defined objectives, make predictions, recommendations, or
decisions influencing real or virtual environments. AI systems are designed
to operate with varying levels of autonomy.
‒AI system lifecycle: AI system lifecycle phases involve: i) ‘design, data
and models’; which is a context-dependent sequence encompassing
planning and design, data collection and processing, as well as model
building; ii) ‘verification
and
validation’; iii) ‘deployment’;
and iv) ‘operation and monitoring’. These phases often take place in an
iterative manner and are not necessarily sequential. The decision to retire
517
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
an AI system from operation may occur at any point during the operation
and monitoring phase.
‒AI knowledge: AI knowledge refers to the skills and resources, such as
data, code, algorithms, models, research, know-how, training programmes,
governance, processes and best practices, required to understand and
participate in the AI system lifecycle.
‒AI actors: AI actors are those who play an active role in the AI system
lifecycle, including organisations and individuals that deploy or operate AI.
‒Stakeholders: Stakeholders encompass all organisations and individuals
involved in, or affected by, AI systems, directly or indirectly. AI actors are
a subset of stakeholders.
Section 1:
Principles for responsible stewardship of trustworthy AI
II. RECOMMENDS that Members and non-Members adhering to this
Recommendation (hereafter the “Adherents”) promote and implement the
following principles for responsible stewardship of trustworthy AI, which
are relevant to all stakeholders.
III. CALLS ON all AI actors to promote and implement, according to their
respective roles, the following Principles for responsible stewardship of
trustworthy AI.
IV. UNDERLINES that the following principles are complementary and
should be considered as a whole.
1.1.Inclusive growth, sustainable development and well-being
Stakeholders should proactively engage in responsible stewardship of
trustworthy AI in pursuit of beneficial outcomes for people and the planet,
such as augmenting human capabilities and enhancing creativity, advancing
inclusion of underrepresented populations, reducing economic, social,
gender and other inequalities, and protecting natural environments, thus
invigorating inclusive growth, sustainable development and well-being.
1.2. Human-centred values and fairness
a) AI actors should respect the rule of law, human rights and
democratic values, throughout the AI system lifecycle. These
include freedom, dignity and autonomy, privacy and data protection,
non-discrimination and equality, diversity, fairness, social justice,
and internationally recognised labour rights.
b) To this end, AI actors should implement mechanisms and
safeguards, such as capacity for human determination, that are
appropriate to the context and consistent with the state of art.
1.3. Transparency and explainability
518
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI Actors should commit to transparency and responsible disclosure
regarding AI systems. To this end, they should provide meaningful
information, appropriate to the context, and consistent with the state of art:
i. to foster a general understanding of AI systems,
ii. to make stakeholders aware of their interactions with AI systems,
including in the workplace,
iii.to enable those affected by an AI system to understand the
outcome, and,
iv.to enable those adversely affected by an AI system to challenge
its outcome based on plain and easy-to-understand information on
the factors, and the logic that served as the basis for the prediction,
recommendation or decision.
1.4. Robustness, security and safety
a) AI systems should be robust, secure and safe throughout their
entire lifecycle so that, in conditions of normal use, foreseeable use
or misuse, or other adverse conditions, they function appropriately
and do not pose unreasonable safety risk.
b) To this end, AI actors should ensure traceability, including in
relation to datasets, processes and decisions made during the AI
system lifecycle, to enable analysis of the AI system’s outcomes and
responses to inquiry, appropriate to the context and consistent with
the state of art.
c) AI actors should, based on their roles, the context, and their ability
to act, apply a systematic risk management approach to each phase
of the AI system lifecycle on a continuous basis to address risks
related to AI systems, including privacy, digital security, safety and
bias.
1.5. Accountability
AI actors should be accountable for the proper functioning of AI systems
and for the respect of the above principles, based on their roles, the context,
and consistent with the state of art.
Section 2:
National policies and international co-operation
for trustworthy AI
V.RECOMMENDS that
Adherents
implement
the
following
recommendations, consistent with the principles in section 1, in their
national policies and international co-operation, with special attention to
small and medium-sized enterprises (SMEs).
2.1. Investing in AI research and development
a) Governments should consider long-term public investment, and
encourage private investment, in research and development,
519
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
including interdisciplinary efforts, to spur innovation in trustworthy
AI that focus on challenging technical issues and on AI-related
social, legal and ethical implications and policy issues.
b) Governments should also consider public investment and
encourage private investment in open datasets that are representative
and respect privacy and data protection to support an environment
for AI research and development that is free of inappropriate bias
and to improve interoperability and use of standards.
2.2. Fostering a digital ecosystem for AI
Governments should foster the development of, and access to, a digital
ecosystem for trustworthy AI. Such an ecosystem includes in particular
digital technologies and infrastructure, and mechanisms for sharing AI
knowledge, as appropriate. In this regard, governments should consider
promoting mechanisms, such as data trusts, to support the safe, fair, legal
and ethical sharing of data.
2.3. Shaping an enabling policy environment for AI
a) Governments should promote a policy environment that supports
an agile transition from the research and development stage to the
deployment and operation stage for trustworthy AI systems. To this
effect, they should consider using experimentation to provide a
controlled environment in which AI systems can be tested, and
scaled-up, as appropriate.
b) Governments should review and adapt, as appropriate, their
policy and regulatory frameworks and assessment mechanisms as
they apply to AI systems to encourage innovation and competition
for trustworthy AI.
2.4. Building human capacity and preparing for labour market
transformation
a) Governments should work closely with stakeholders to prepare
for the transformation of the world of work and of society. They
should empower people to effectively use and interact with AI
systems across the breadth of applications, including by equipping
them with the necessary skills.
b) Governments should take steps, including through social
dialogue, to ensure a fair transition for workers as AI is deployed,
such as through training programmes along the working life, support
for those affected by displacement, and access to new opportunities
in the labour market.
c) Governments should also work closely with stakeholders to
promote the responsible use of AI at work, to enhance the safety of
workers and the quality of jobs, to foster entrepreneurship and
520
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
productivity, and aim to ensure that the benefits from AI are broadly
and fairly shared.
2.5. International co-operation for trustworthy AI
a) Governments, including developing countries and with
stakeholders, should actively co-operate to advance these principles
and to progress on responsible stewardship of trustworthy AI.
b) Governments should work together in the OECD and other global
and regional fora to foster the sharing of AI knowledge, as
appropriate. They should encourage international, cross-sectoral
and open multi-stakeholder initiatives to garner long-term expertise
on AI.
c) Governments should promote the development of multistakeholder, consensus-driven global technical standards for
interoperable and trustworthy AI.
d) Governments should also encourage the development, and their
own use, of internationally comparable metrics to measure AI
research, development and deployment, and gather the evidence
base to assess progress in the implementation of these principles.
VI. INVITES the Secretary-General and Adherents to disseminate this
Recommendation.
VII. INVITES non-Adherents to take due account of, and adhere to, this
Recommendation.
VIII. INSTRUCTS the Committee on Digital Economy Policy:
a) to continue its important work on artificial intelligence building
on this Recommendation and taking into account work in other
international fora, and to further develop the measurement
framework for evidence-based AI policies;
b) to develop and iterate further practical guidance on the
implementation of this Recommendation, and to report to the
Council on progress made no later than end December 2019;
c) to provide a forum for exchanging information on AI policy and
activities including experience with the implementation of this
Recommendation, and to foster multi-stakeholder and
interdisciplinary dialogue to promote trust in and adoption of AI;
and
d) to monitor, in consultation with other relevant Committees, the
implementation of this Recommendation and report thereon to the
Council no later than five years following its adoption and regularly
thereafter.
521
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD AI Policy Adherents
The following countries have endorsed the OECD AI Principles
OECD Member Countries
Australia
Austria
Belgium
Canada
Chile
Colombia
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Israel
Italy
Japan
Korea
Latvia
Lithuania
Luxembourg
Mexico
Netherlands
New Zealand
Norway
Poland
Portugal
Slovak Republic
Slovenia
Spain
Sweden
Switzerland
Turkey
United Kingdom
United States
522
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
OECD Non-Member Countries
Argentina
Brazil
Costa Rica
Malta
Peru
Romania
Ukraine
G-20 Countries
China
India
Indonesia
Russia
Saudi Arabia
South Africa
As of December 1, 2020, 51 countries have endorsed the OECD/G20 AI
Principles.
523
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Universal Guidelines for AI
Universal Guidelines for Artificial Intelligence
23 October 2018
Brussels, Belgium
New developments in Artificial Intelligence are transforming the
world, from science and industry to government administration and finance.
The rise of AI decision-making also implicates fundamental rights of
fairness, accountability, and transparency. Modern data analysis produces
significant outcomes that have real life consequences for people in
employment, housing, credit, commerce, and criminal sentencing. Many of
these techniques are entirely opaque, leaving individuals unaware whether
the decisions were accurate, fair, or even about them.
We propose these Universal Guidelines to inform and improve the
design and use of AI. The Guidelines are intended to maximize the benefits
of AI, to minimize the risk, and to ensure the protection of human rights.
These Guidelines should be incorporated into ethical standards, adopted in
national law and international agreements, and built into the design of
systems. We state clearly that the primary responsibility for AI systems
must reside with those institutions that fund, develop, and deploy these
systems.
1. Right to Transparency. All individuals have the right to know the
basis of an AI decision that concerns them. This includes access to
the factors, the logic, and techniques that produced the outcome.
2. Right to Human Determination. All individuals have the right to
a final determination made by a person.
3. Identification Obligation. The institution responsible for an AI
system must be made known to the public.
4. Fairness Obligation. Institutions must ensure that AI systems do
not reflect unfair bias or make impermissible discriminatory
decisions.
5. Assessment and Accountability Obligation. An AI system should
be deployed only after an adequate evaluation of its purpose and
objectives, its benefits, as well as its risks. Institutions must be
responsible for decisions made by an AI system.
6. Accuracy, Reliability, and Validity Obligations. Institutions must
ensure the accuracy, reliability, and validity of decisions.
7. Data Quality Obligation. Institutions must establish data
provenance, and assure quality and relevance for the data input into
algorithms.
524
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
8. Public Safety Obligation. Institutions must assess the public safety
risks that arise from the deployment of AI systems that direct or
control physical devices, and implement safety controls.
9. Cybersecurity Obligation. Institutions must secure AI systems
against cybersecurity threats.
10. Prohibition on Secret Profiling. No institution shall establish or
maintain a secret profiling system.
11. Prohibition on Unitary Scoring. No national government shall
establish or maintain a general-purpose score on its citizens or
residents.
12. Termination Obligation. An institution that has established an AI
system has an affirmative obligation to terminate the system if
human control of the system is no longer possible.
525
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
UGAI Explanatory Memorandum
Context
The Universal Guidelines on Artificial Intelligence (UGAI) call
attention to the growing challenges of intelligent computational systems and
proposes concrete recommendations that can improve and inform their
design. At its core, the purpose of the UGAI is to promote transparency and
accountability for these systems and to ensure that people retain control over
the systems they create. Not all systems fall within the scope of these
Guidelines. Our concern is with those systems that impact the rights of
people. Above all else, these systems should do no harm.
The declaration is timely. Governments around the word are
developing policy proposals and institutions, both public and private, are
supporting research and development of “AI.” Invariably, there will be an
enormous impact on the public, regardless of their participation in the
design and development of these systems. And so, the UGAI reflects a
public perspective on these challenges.
The UGAI were announced at the 2018 International Data
Protection and Privacy Commissioners Conference, among the most
significant meetings of technology leaders and data protection experts in
history.
The UGAI builds on prior work by scientific societies, think tanks,
NGOs, and international organizations. The UGAI incorporates elements of
human rights doctrine, data protection law, and ethical guidelines. The
Guidelines include several well-established principles for AI governance,
and put forward new principles not previously found in similar policy
frameworks.
Terminology
The term “Artificial Intelligence” is both broad and imprecise. It
includes aspects of machine learning, rule-based decision-making, and
other computational techniques. There are also disputes regarding whether
Artificial Intelligence is possible. The UGAI simply acknowledges that this
term, in common use, covers a wide range of related issues and adopts the
term to engage the current debate. There is no attempt here to define its
boundaries, other than to assume that AI requires some degree of automated
decision-making. The term “Guidelines” follows the practice of policy
frameworks that speak primarily to governments and private companies.
The UGAI speaks to the obligations of “institutions” and the rights
of “individuals.” This follows from the articulation of fair information
practices in the data protection field. The UGAI takes the protection of the
526
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
individual as a fundamental goal. Institutions, public and private, are
understood to be those entities that develop and deploy AI systems. The
term “institution” was chosen rather than the more familiar “organization”
to underscore the permanent, ongoing nature of the obligations set out in
the Guidelines. There is one principle that is addressed to “national
governments.” The reason for this is discussed below.
Application
These Guidelines should be incorporated into ethical standards,
adopted in national law and international agreements, and built into the
design of systems.
The Principles
The elements of the Transparency Principle can be found in
several modern privacy laws, including the US Privacy Act, the EU Data
Protection Directive, the GDPR, and the Council of Europe Convention
108. The aim of this principle is to enable independent accountability for
automated decisions, with a primary emphasis on the right of the individual
to know the basis of an adverse determination. In practical terms, it may not
be possible for an individual to interpret the basis of a particular decision,
but this does not obviate the need to ensure that such an explanation is
possible.
The Right to a Human Determination reaffirms that individuals
and not machines are responsible for automated decision-making. In many
instances, such as the operation of an autonomous vehicle, it would not be
possible or practical to insert a human decision prior to an
automated decision. But the aim remains to ensure accountability. Thus
where an automated system fails, this principle should be understood as a
requirement that a human assessment of the outcome be made.
Identification Obligation. This principle seeks to address the
identification asymmetry that arises in the interaction between individuals
and AI systems. An AI system typically knows a great deal about an
individual; the individual may not even know the operator of the AI system.
The Identification Obligation establishes the foundation of AI
accountability which is to make clear the identity of an AI system and the
institution responsible.
The Fairness Obligation recognizes that all automated systems
make decisions that reflect bias and discrimination, but such decisions
should not be normatively unfair. There is no simple answer to the question
as to what is unfair or impermissible. The evaluation often depends on
context. But the Fairness Obligation makes clear that an assessment of
objective outcomes alone is not sufficient to evaluate an AI system.
527
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Normative consequences must be assessed, including those that preexist or
may be amplified by an AI system.
The Assessment and Accountability Obligation speaks to the
obligation to assess an AI system prior to and during deployment.
Regarding assessment, it should be understood that a central purpose of this
obligation is to determine whether an AI system should be established. If an
assessment reveals substantial risks, such as those suggested by principles
concerning Public Safety and Cybersecurity, then the project should not
move forward.
The Accuracy, Reliability, and Validity Obligations set out key
responsibilities associated with the outcome of automated decisions. The
terms are intended to be interpreted both independently and jointly.
The Data Quality Principle follows from the preceding obligation.
The Public Safety Obligation recognizes that AI systems control
devices in the physical world. For this reason, institutions must both assess
risks and take precautionary measures as appropriate.
The Cybersecurity Obligation follows from the Public Safety
Obligation and underscores the risk that even well-designed systems may
be the target of hostile actors. Those who develop and deploy AI systems
must take these risks into account.
The Prohibition on Secret Profiling follows from the earlier
Identification Obligation. The aim is to avoid the information asymmetry
that arises increasingly with AI systems and to ensure the possibility of
independent accountability.
The Prohibition on Unitary Scoring speaks directly to the risk of
a single, multi-purpose number assigned by a government to an individual.
In data protection law, universal identifiers that enable the profiling of
individuals across are disfavored. These identifiers are often regulated and
in some instances prohibited. The concern with universal scoring, described
here as “unitary scoring,” is even greater. A unitary score reflects not only
a unitary profile but also a predetermined outcome across multiple domains
of human activity. There is some risk that unitary scores will also emerge
in the private sector. Conceivably, such systems could be subject to market
competition and government regulations. But there is not even the
possibility of counterbalance with unitary scores assigned by government,
and therefore they should be prohibited.
The Termination Obligation is the ultimate statement of
accountability for an AI system. The obligation presumes that systems must
remain within human control. If that is no longer possible, the system should
be terminated.
528
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
UGAI References
Asilomar AI Principles (2017)
Aspen Institute Roundtable on Artificial Intelligence (2016)
Association for Computing Machinery, U.S. Public Policy
Counsel, Statement on Algorithmic Transparency and Accountability (Jan.
2017)
Council of Europe, Convention 108 (1981)
Council of Europe and Artificial Intelligence (2018)
Data and Society, Governing Artificial Intelligence (2018)
European Commission, High Level Expert Group on Artificial
Intelligence (2018)
EU General Data Protection Regulation (2018)
IEEE, Ethically Aligned Design (2016)
Japan, Ministry of Internal Affairs and Communications, AI R&D
Guidelines (2016)
Garry Kasparov, Deep Thinking: Where Machine Intelligence Ends and
Human Creativity Begins (2017)
Madrid Privacy Declaration (2009)
OECD, Artificial Intelligence (2018)
OECD, Privacy Guidelines (1980)
Cathy O’Neil, Weapons of Math Destruction (2016)
Frank Pasquale, The Black Box Society: The Secret Algorithms That
Control Money and Information (2015)
Privacy International, Artificial Intelligence (2018)
529
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
US Privacy Act (1974)
Toronto Declaration (2018)
Joseph Weizenbaum, Computer Power and Human Reason (1976)
Universal Declaration of Human Rights (1948)
530
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Universal Declaration of Human Rights
Preamble
Whereas recognition of the inherent dignity and of the equal and inalienable
rights of all members of the human family is the foundation of freedom,
justice and peace in the world,
Whereas disregard and contempt for human rights have resulted in
barbarous acts which have outraged the conscience of mankind, and the
advent of a world in which human beings shall enjoy freedom of speech and
belief and freedom from fear and want has been proclaimed as the highest
aspiration of the common people,
Whereas it is essential, if man is not to be compelled to have recourse, as a
last resort, to rebellion against tyranny and oppression, that human rights
should be protected by the rule of law,
Whereas it is essential to promote the development of friendly relations
between nations,
Whereas the peoples of the United Nations have in the Charter reaffirmed
their faith in fundamental human rights, in the dignity and worth of the
human person and in the equal rights of men and women and have
determined to promote social progress and better standards of life in larger
freedom,
Whereas Member States have pledged themselves to achieve, in cooperation with the United Nations, the promotion of universal respect for
and observance of human rights and fundamental freedoms,
Whereas a common understanding of these rights and freedoms is of the
greatest importance for the full realization of this pledge,
Now, Therefore THE GENERAL ASSEMBLY proclaims THIS
UNIVERSAL DECLARATION OF HUMAN RIGHTS as a common
standard of achievement for all peoples and all nations, to the end that every
individual and every organ of society, keeping this Declaration constantly
in mind, shall strive by teaching and education to promote respect for these
rights and freedoms and by progressive measures, national and
international, to secure their universal and effective recognition and
observance, both among the peoples of Member States themselves and
among the peoples of territories under their jurisdiction.
531
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Article 1
All human beings are born free and equal in dignity and rights. They
are endowed with reason and conscience and should act towards one another
in a spirit of brotherhood.
Article 2
Everyone is entitled to all the rights and freedoms set forth in this
Declaration, without distinction of any kind, such as race, colour, sex,
language, religion, political or other opinion, national or social origin,
property, birth or other status. Furthermore, no distinction shall be made on
the basis of the political, jurisdictional or international status of the country
or territory to which a person belongs, whether it be independent, trust, nonself-governing or under any other limitation of sovereignty.
Article 3
Everyone has the right to life, liberty and security of person.
Article 4
No one shall be held in slavery or servitude; slavery and the slave
trade shall be prohibited in all their forms.
Article 5
No one shall be subjected to torture or to cruel, inhuman or
degrading treatment or punishment.
Article 6
Everyone has the right to recognition everywhere as a person before
the law.
Article 7
All are equal before the law and are entitled without any
discrimination to equal protection of the law. All are entitled to equal
protection against any discrimination in violation of this Declaration and
against any incitement to such discrimination.
Article 8
Everyone has the right to an effective remedy by the competent
national tribunals for acts violating the fundamental rights granted him by
the constitution or by law.
532
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Article 9
No one shall be subjected to arbitrary arrest, detention or exile.
Article 10
Everyone is entitled in full equality to a fair and public hearing by
an independent and impartial tribunal, in the determination of his rights and
obligations and of any criminal charge against him.
Article 11
(1) Everyone charged with a penal offence has the right to be presumed
innocent until proved guilty according to law in a public trial at which he
has had all the guarantees necessary for his defence.
(2) No one shall be held guilty of any penal offence on account of any act
or omission which did not constitute a penal offence, under national or
international law, at the time when it was committed. Nor shall a heavier
penalty be imposed than the one that was applicable at the time the penal
offence was committed.
Article 12
No one shall be subjected to arbitrary interference with his privacy,
family, home or correspondence, nor to attacks upon his honour and
reputation. Everyone has the right to the protection of the law against such
interference or attacks.
Article 13
(1) Everyone has the right to freedom of movement and residence within
the borders of each state.
(2) Everyone has the right to leave any country, including his own, and to
return to his country.
Article 14
(1) Everyone has the right to seek and to enjoy in other countries asylum
from persecution.
(2) This right may not be invoked in the case of prosecutions genuinely
arising from non-political crimes or from acts contrary to the purposes and
principles of the United Nations.
533
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Article 15
(1) Everyone has the right to a nationality.
(2) No one shall be arbitrarily deprived of his nationality nor denied the
right to change his nationality.
Article 16
(1) Men and women of full age, without any limitation due to race,
nationality or religion, have the right to marry and to found a family. They
are entitled to equal rights as to marriage, during marriage and at its
dissolution.
(2) Marriage shall be entered into only with the free and full consent of the
intending
spouses.
(3) The family is the natural and fundamental group unit of society and is
entitled to protection by society and the State.
Article 17
(1) Everyone has the right to own property alone as well as in association
with others.
(2) No one shall be arbitrarily deprived of his property.
Article 18
Everyone has the right to freedom of thought, conscience and
religion; this right includes freedom to change his religion or belief, and
freedom, either alone or in community with others and in public or private,
to manifest his religion or belief in teaching, practice, worship and
observance.
Article 19
Everyone has the right to freedom of opinion and expression; this
right includes freedom to hold opinions without interference and to seek,
receive and impart information and ideas through any media and regardless
of frontiers.
Article 20
(1) Everyone has the right to freedom of peaceful assembly and
association.
(2) No one may be compelled to belong to an association.
Article 21
(1) Everyone has the right to take part in the government of his country,
directly
or
through
freely
chosen
representatives.
534
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
(2) Everyone has the right of equal access to public service in his country.
(3) The will of the people shall be the basis of the authority of government;
this will shall be expressed in periodic and genuine elections which shall be
by universal and equal suffrage and shall be held by secret vote or by
equivalent free voting procedures.
Article 22
Everyone, as a member of society, has the right to social security
and is entitled to realization, through national effort and international cooperation and in accordance with the organization and resources of each
State, of the economic, social and cultural rights indispensable for his
dignity and the free development of his personality.
Article 23
(1) Everyone has the right to work, to free choice of employment, to just
and favourable conditions of work and to protection against unemployment.
(2) Everyone, without any discrimination, has the right to equal pay for
equal work.
(3) Everyone who works has the right to just and favourable remuneration
ensuring for himself and his family an existence worthy of human dignity,
and supplemented, if necessary, by other means of social protection.
(4) Everyone has the right to form and to join trade unions for the protection
of his interests.
Article 24
Everyone has the right to rest and leisure, including reasonable
limitation of working hours and periodic holidays with pay.
Article 25
(1) Everyone has the right to a standard of living adequate for the health and
well-being of himself and of his family, including food, clothing, housing
and medical care and necessary social services, and the right to security in
the event of unemployment, sickness, disability, widowhood, old age or
other lack of livelihood in circumstances beyond his control.
(2) Motherhood and childhood are entitled to special care and assistance.
All children, whether born in or out of wedlock, shall enjoy the same social
protection.
Article 26
(1) Everyone has the right to education. Education shall be free, at least in
the elementary and fundamental stages. Elementary education shall be
compulsory. Technical and professional education shall be made generally
available and higher education shall be equally accessible to all on the basis
535
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
of
merit.
(2) Education shall be directed to the full development of the human
personality and to the strengthening of respect for human rights and
fundamental freedoms. It shall promote understanding, tolerance and
friendship among all nations, racial or religious groups, and shall further the
activities of the United Nations for the maintenance of peace.
(3) Parents have a prior right to choose the kind of education that shall be
given to their children.
Article 27
(1) Everyone has the right freely to participate in the cultural life of the
community, to enjoy the arts and to share in scientific advancement and its
benefits.
(2) Everyone has the right to the protection of the moral and material
interests resulting from any scientific, literary or artistic production of
which he is the author.
Article 28
Everyone is entitled to a social and international order in which the rights
and freedoms set forth in this Declaration can be fully realized.
Article 29
(1) Everyone has duties to the community in which alone the free and full
development of his personality is possible.
(2) In the exercise of his rights and freedoms, everyone shall be subject only
to such limitations as are determined by law solely for the purpose of
securing due recognition and respect for the rights and freedoms of others
and of meeting the just requirements of morality, public order and the
general welfare in a democratic society.
(3) These rights and freedoms may in no case be exercised contrary to the
purposes and principles of the United Nations.
Article 30
Nothing in this Declaration may be interpreted as implying for any
State, group or person any right to engage in any activity or to perform any
act aimed at the destruction of any of the rights and freedoms set forth
herein.
536
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
GPA Declaration on Ethics and Data Protection in AI
DECLARATION ON ETHICS AND DATA PROTECTION IN
ARTIFICAL INTELLIGENCE
40th International Conference of Data Protection and Privacy
Commissioners
23rd October 2018, Brussels
[Note: The International Conference of Data Protection and Privacy
Commissioners was later renamed the Global Privacy Assembly]
AUTHORS
Commission Nationale de l’Informatique et des Libertés (CNIL),
France
European Data Protection Supervisor (EDPS), European Union
Garante per la protezione dei dati personali, Italy
CO-SPONSORS:
Agencia de Acceso a la Información Pública, Argentina
Commission d’accès à l’information, Québec, Canada
Datatilsynet (Data Inspectorate), Norway
Information Commissioner’s Office (ICO), United Kingdom
Préposé fédéral à la protection des données et à la transparence,
Switzerland
Data protection Authority, Belgium
Privacy Commissioner for Personal Data, Hong-Kong
Data protection Commission, Ireland
Data Protection Office, Poland
Instituto Nacional de Transparencia, Acceso a la Información y
Protección de Datos Personales (INAI), Mexico
National Authority for Data Protection and Freedom of Information,
Hungary
Federal Commissioner for Data Protection and Freedom of
Information, Germany
Office of the Privacy Commissioner (OPC), Canada
National Privacy Commission, Philippines
The 40th International Conference of Data Protection and Privacy
Commissioners:
537
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Considering the initial discussion at the 38th International Conference of
Data Protection and Privacy Commissioners in Marrakesh on Artificial
intelligence, Robotics, Privacy and Data Protection;
Recognizing that artificial intelligence systems may bring significant
benefits for users and society, including by: increasing the rapidity of
processes and supporting decision-making; creating new ways to participate
in democratic processes; improving efficiency in public sector and industry;
achieving more equitable distribution of resources and opportunities;
offering new methods and solutions in various fields such as public health,
medical care, security, sustainable development, agriculture and transport;
bringing new opportunities in scientific research and education and;
providing individuals with more personalized services;
Taking into account the significant progress in certain areas of artificial
intelligence, in particular regarding the processing of large amounts of
information, the analysis and prediction of human behavior and
characteristics, and in related fields such as robotics, computer vision and
autonomous systems, likely to make significant progress in the near future;
Highlighting the rapid advancement of big data and artificial intelligence,
notably machine learning, in particular with the development of deep
learning technologies, allowing algorithms to solve complex operations
leading to potential decisions, making however such processes more
opaque;
Affirming that the respect of the rights to privacy and data protection are
increasingly challenged by the development of artificial intelligence and
that this development should be complemented by ethical and human rights
considerations;
Considering that machine learning technologies in particular, and artificial
intelligence systems in general, may rely on the processing of large sets of
personal data for their development, potentially impacting data protection
and privacy; also taking into account the potential risks induced by the
current trend of market concentration in the field of artificial intelligence;
Recognizing the link between collections, uses and disclosures of personal
information – the traditional sphere of privacy and data protection – on the
one hand, and the direct impacts on human rights more broadly, most
notably regarding discrimination and freedom of expression and
information, and thus acknowledging the need for data protection and
538
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
privacy authorities to think about human rights more broadly, and for data
protection and privacy authorities to work with other authorities addressing
human rights;
Pointing out that some data sets used to train machine learning-based and
artificial intelligence systems have been found to contain inherent bias
resulting in decisions which can unfairly discriminate against certain
individuals or groups, potentially restricting the availability of certain
services or content, and thus interfering with individuals’ rights such as
freedom of expression and information or resulting in the exclusion of
people from certain aspects of personal, social, professional life;
Stressing that artificial intelligence powered systems whose decisions
cannot be explained raise fundamental questions of accountability not only
for privacy and data protection law but also liability in the event of errors
and harm;
Noting that many stakeholders in the field of artificial intelligence have
expressed their concerns about the risks of malicious use of artificial
intelligence, as well as the risks related to privacy, data protection and
human dignity, pointing out for example that the development of artificial
intelligence in combination with mass surveillance raises concerns about
their possible use to curtail fundamental rights and freedoms;
Highlighting that those risks and challenges may affect individuals and
society, and that the extent and nature of potential consequences are
currently uncertain;
Emphasising the importance of trust, since strong data protection and
privacy safeguards help to build individuals’ trust in how their data is
processed, which encourages data sharing and thereby promotes innovation;
Taking the view that the current challenges triggered by the development of
artificial intelligence and machine learning systems reinforce the need for
the adoption of an international approach and standards, in order to ensure
the promotion and protection of human rights in all digital developments at
international level;
Reaffirming the commitment of data protection authorities and the
Conference of Data Protection and Privacy Commissioners to uphold data
protection and privacy principles in adapting to this evolving environment,
539
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
notably by engaging resources and developing new skills in order to be
prepared for future changes.
The 40th International Conference of Data Protection and Privacy
Commissioners considers that any creation, development and use of
artificial intelligence systems shall fully respect human rights, particularly
the rights to the protection of personal data and to privacy, as well as human
dignity, non-discrimination and fundamental values, and shall provide
solutions to allow individuals to maintain control and understanding of
artificial intelligence systems.
The Conference therefore endorses the following guiding principles, as
its core values to preserve human rights in the development of artificial
intelligence:
1. Artificial intelligence and machine learning technologies should be
designed, developed and used in respect of fundamental human rights and
in accordance with the fairness principle, in particular by:
a. Considering individuals’ reasonable expectations by ensuring that
the use of artificial intelligence systems remains consistent with
their original purposes, and that the data are used in a way that is not
incompatible with the original purpose of their collection,
b. taking into consideration not only the impact that the use of artificial
intelligence may have on the individual, but also the collective
impact on groups and on society at large,
c. ensuring that artificial intelligence systems are developed in a way
that facilitates human development and does not obstruct or
endanger it, thus recognizing the need for delineation and
boundaries on certain uses,
2. Continued attention and vigilance, as well as accountability, for the
potential effects and consequences of, artificial intelligence systems should
be ensured, in particular by:
a. promoting accountability of all relevant stakeholders to individuals,
supervisory authorities and other third parties as appropriate,
including through the realization of audit, continuous monitoring
and impact assessment of artificial intelligence systems, and
periodic review of oversight mechanisms;
b. fostering collective and joint responsibility, involving the whole
chain of actors and stakeholders, for example with the development
of collaborative standards and the sharing of best practices,
540
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
c. investing in awareness raising, education, research and training in
order to ensure a good level of information on and understanding of
artificial intelligence and its potential effects in society, and
d. establishing demonstrable governance processes for all relevant
actors, such as relying on trusted third parties or the setting up of
independent ethics committees,
3. Artificial intelligence systems transparency and intelligibility should
be improved, with the objective of effective implementation, in particular
by:
a. investing in public and private scientific research on explainable
artificial intelligence,
b. promoting transparency, intelligibility and reachability, for instance
through the development of innovative ways of communication,
taking into account the different levels of transparency and
information required for each relevant audience,
c. making organizations’ practices more transparent, notably by
promoting algorithmic transparency and the auditability of systems,
while ensuring meaningfulness of the information provided, and
d. guaranteeing the right to informational self-determination, notably
by ensuring that individuals are always informed appropriately
when they are interacting directly with an artificial intelligence
system or when they provide personal data to be processed by such
systems,
e. providing adequate information on the purpose and effects of
artificial intelligence systems in order to verify continuous
alignment with expectation of individuals and to enable overall
human control on such systems.
4. As part of an overall “ethics by design” approach, artificial intelligence
systems should be designed and developed responsibly, by applying the
principles of privacy by default and privacy by design, in particular by:
a. implementing technical and organizational measures and
procedures – proportional to the type of system that is developed –
to ensure that data subjects’ privacy and personal data are respected,
both when determining the means of the processing and at the
moment of data processing,
b. assessing and documenting the expected impacts on individuals and
society at the beginning of an artificial intelligence project and for
relevant developments during its entire life cycle, and
541
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
c. identifying specific requirements for ethical and fair use of the
systems and for respecting human rights as part of the development
and operations of any artificial intelligence system,
5. Empowerment of every individual should be promoted, and the
exercise of individuals’ rights should be encouraged, as well as the creation
of opportunities for public engagement, in particular by:
a. respecting data protection and privacy rights, including where
applicable the right to information, the right to access, the right to
object to processing and the right to erasure, and promoting those
rights through education and awareness campaigns,
b. respecting related rights including freedom of expression and
information, as well as non- discrimination,
c. recognizing that the right to object or appeal applies to technologies
that influence personal development or opinions and guaranteeing,
where applicable, individuals’ right not to be subject to a decision
based solely on automated processing if it significantly affects them
and, where not applicable, guaranteeing individuals’ right to
challenge such decision,
d. using the capabilities of artificial intelligence systems to foster an
equal empowerment and enhance public engagement, for example
through adaptable interfaces and accessible tools.
6. Unlawful biases or discriminations that may result from the use of data
in artificial intelligence should be reduced and mitigated, including by:
a. ensuring the respect of international legal instruments on human
rights and non-discrimination,
b. investing in research into technical ways to identify, address and
mitigate biases,
c. taking reasonable steps to ensure the personal data and information
used in automated decision making is accurate, up-to-date and as
complete as possible, and
d. elaborating specific guidance and principles in addressing biases
and discrimination, and promoting individuals’ and stakeholders’
awareness.
Taking into consideration the principles above, the 40th International
Conference of Data Protection and Privacy Commissioners calls for
common governance principles on artificial intelligence to be
established, fostering concerted international efforts in this field, in order to
ensure that its development and use take place in accordance with ethics
542
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and human values, and respect human dignity. These common governance
principles must be able to tackle the challenges raised by the rapid
evolutions of artificial intelligence technologies, on the basis of a multistakeholder approach in order to address all cross-sectoral issues at stake.
They must take place at an international level since the development of
artificial intelligence is a trans- border phenomenon and may affect all
humanity. The Conference should be involved in this international effort,
working with and supporting general and sectoral authorities in other fields
such as competition, market and consumer regulation.
The 40th International Conference of Data Protection and Privacy
Commissioners therefore establishes, as a contribution to a future common
governance at the international level, and in order to further elaborate
guidance to accompany the principles on Ethics and Data Protection in
Artificial Intelligence, a permanent working group addressing the
challenges of artificial intelligence development. This working group on
Ethics and Data Protection in Artificial Intelligence will be in charge of
promoting understanding of and respect for the principles of the present
resolution, by all relevant parties involved in the development of artificial
intelligence systems, including governments and public authorities,
standardization bodies, artificial intelligence systems designers, providers
and researchers, companies, citizens and end users of artificial intelligence
systems. The working group on Ethics and Data Protection in Artificial
Intelligence shall take into account the work carried out by other working
groups of the Conference and shall report regularly on its activities to the
Conference. The Conference thus endeavors to proactively support an
active public debate on digital ethics aiming at the creation of a strong
ethical culture and personal awareness in this field.
543
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
GPA Resolution on AI and Accountability
RESOLUTION ON ACCOUNTABILITY
IN THE DEVELOPMENT AND USE OF ARTIFICIAL
INTELLIGENCE
Global Privacy Assembly
October 2020
Sponsors
Privacy Commissioner for Personal Data, Hong Kong, China
Superintendence of Industry and Commerce, Colombia
Federal Commissioner for Data Protection and Freedom of
Information, Germany
Information Commissioner’s Office, United Kingdom
Co-Sponsors
Agencia de Acceso a la Información Pública, Argentina
Office of the Privacy Commissioner of Canada
Information Access Commission, Quebec, Canada
European Data Protection Supervisor, European Union
Data Protection Commission, Italy
National Institute for Transparency, Access to Information and
Personal Data Protection, Mexico
Office of the Privacy Commissioner, New Zealand
National Privacy Commission, Philippines
Personal Data Protection Office, Poland
National Data Protection Commission, Portugal
Data Protection Authority, Republic of San Marino
National Commission for Informatics and Liberties, Burkina Faso
Office of the Information and Privacy Commissioner, Ontario, Canada
The 2020 GLOBAL PRIVACY ASSEMBLY:
Recalling the Declaration on Ethics and Data Protection in Artificial
Intelligence made by the 40th International Conference of the Data
Protection and Privacy Commissioners on 23 October 2018, which
endorsed inter alia the principle of accountability of all relevant
stakeholders to individuals, supervisory authorities and other third parties,
and which established a permanent Working Group (AI WG) to address the
challenges of development of artificial intelligence (AI), and promote
understanding of and respect for the principles of the Declaration,
Highlighting that the Work Programme of the AI WG includes an action to
prepare a statement on the essential need for accountability and liability of
human actors for AI systems,
544
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Taking into account the results of a survey conducted by the AI WG in
May and June 2020, to gather the views of the members of the Global
Privacy Assembly on accountability for AI systems, as detailed in the
Explanatory Note,
Noting that international organisations (including the United Nations, the
Organisation for Economic Co-operation and Development, the Council of
Europe and the European Commission), governments, civil society bodies,
and technology companies have produced and continue to produce
guidelines and recommendation on the legal and ethical development of AI,
and that the need for accountability and a human-centric approach are
common themes within these guidelines,
Noting that accountability is to be understood as the compliance and
demonstration of compliance with personal data protection and privacy
regulations, in particular through the adoption and implementation of
appropriate, practicable, systematic and effective measures,
Affirming that the responsibility for the operation and effects of AI systems
remains with human actors,
Taking the view that in order to be effective, accountability obligations
should be assessed against clearly defined principles and frameworks, and
extend to both organisations that develop AI systems and organisations that
use them,
Emphasising that the principle of accountability encompasses
accountability to the people affected by the decisions made by or with AI
systems, as well as to supervisory authorities and, where appropriate, to
other third parties, and that beyond the compliance element, accountability
should also be demonstrated in order to build trust with the stakeholders,
Recognising that AI systems may affect human rights in different ways, the
application of specific obligations should take into account the risks for
human rights as well as the importance of the principle of human
accountability,
Asserting that in order to support the trustworthiness of organisations
developing and using AI systems, these organisations should work closely
with policy-makers, individuals and other stakeholders (e.g. nongovernment organisations, public authorities and academia) to resolve
concerns and rectify adverse impacts on human rights.
The 2020 GLOBAL PRIVACY ASSEMBLY therefore resolves to:
1. Urge organisations that develop or use AI systems to consider
implementing the following accountability measures:
545
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
(1) Assess the potential impact to human rights (including data
protection and privacy rights) before the development and/or use of
AI;
(2) Test the robustness, reliability, accuracy and data security of AI
before putting it into use, including identifying and addressing bias in
the systems and the data they use that may lead to unfair outcomes;
(3) Keep records of impact assessment, design, development, testing
and use of AI;
(4) Disclose the results of the data protection, privacy and human
rights impact assessment of AI;
(5) Ensure transparency and openness by disclosing the use of AI, the
data being used and the logic involved in the AI;
(6) Ensure an accountable human actor is identified (a) with whom
concerns related to automated decisions can be raised and rights can be
exercised, and (b) who can trigger evaluation of the decision process
and human intervention;
(7) Provide explanations in clear and understandable language for the
automated decisions made by AI upon request;
(8) Make human intervention on the automated decision made by AI
upon request;
(9) Continuously monitor and evaluate the performance and impacts of
AI by human beings, and act promptly and firmly to address identified
issues;
(10) Implement whistleblowing / reporting mechanisms about noncompliance or significant risk in the use of AI;
(11) Ensure the auditability of AI systems and be prepared to
demonstrate accountability to data protection authorities on request;
and
(12) Engage in multi-stakeholder discussions (including with nongovernmental organisations, public authorities and academia) to
identify and address the wider socio- economic impact of AI and to
ensure algorithmic vigilance.
2. Urge organisations that develop or use AI systems to implement
accountability measures which are appropriate regarding the risks of
interference with human rights.
3. Call upon all members of the Global Privacy Assembly to work with
organisations that develop or use AI systems in their jurisdictions and
globally to promote the principles adopted in its 2018 resolution, and
accountability in the development and use of AI, and the adoption of
accountability measures;
546
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
4. Encourage governments to consider the need to make legislative
changes in personal data protection laws, to make clear the legal
obligations regarding accountability in the development and use of AI,
where such provisions are not already in place; and
5. Encourage governments, public authorities, standardisation bodies,
organisations developing or using AI systems and all other relevant
stakeholders to work with data protection authorities in establishing
principles, standards, and accountability mechanisms, such as
certification, for the purpose of demonstrating legal compliance,
accountability and ethics in the development and use of AI systems.
[An Explanatory Note accompanies the Resolution. The Explanatory Note
summarizes the opinions of the members of the Global Privacy Assembly
on the measures for demonstrating accountability in the development and
use of AI.]
547
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
UNESCO Recommendation on AI Ethics
The UNESCO Recommendation on the Ethics of Artificial
Intelligence
Adopted November 24, 2021
Preamble
The General Conference of the United Nations Educational, Scientific and
Cultural Organization (UNESCO), meeting in Paris from 9 to 24, at its 41st
session,
Recognizing the profound and dynamic positive and negative impacts of
artificial intelligence (AI) on societies, environment, ecosystems and
human lives, including the human mind, in part because of the new ways in
which its use influences human thinking, interaction and decision-making
and affects education, human, social and natural sciences, culture, and
communication and information,
Recalling that, by the terms of its Constitution, UNESCO seeks to
contribute to peace and security by promoting collaboration among nations
through education, the sciences, culture, and communication and
information, in order to further universal respect for justice, for the rule of
law and for the human rights and fundamental freedoms which are affirmed
for the peoples of the world,
Convinced that the Recommendation presented here, as a standard-setting
instrument developed through a global approach, based on international
law, focusing on human dignity and human rights, as well as gender
equality, social and economic justice and development, physical and mental
wellbeing, diversity, interconnectedness, inclusiveness, and environmental
and ecosystem protection can guide AI technologies in a responsible
direction,
Guided by the purposes and principles of the Charter of the United Nations,
Considering that AI technologies can be of great service to humanity and
all countries can benefit from them, but also raise fundamental ethical
concerns, for instance regarding the biases they can embed and exacerbate,
potentially resulting in discrimination, inequality, digital divides, exclusion
and a threat to cultural, social and biological diversity and social or
548
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
economic divides; the need for transparency and understandability of the
workings of algorithms and the data with which they have been trained; and
their potential impact on, including but not limited to, human dignity,
human rights and fundamental freedoms, gender equality, democracy,
social, economic, political and cultural processes, scientific and
engineering practices, animal welfare, and the environment and
ecosystems,
Also recognizing that AI technologies can deepen existing divides and
inequalities in the world, within and between countries, and that justice,
trust and fairness must be upheld so that no country and no one should be
left behind, either by having fair access to AI technologies and enjoying
their benefits or in the protection against their negative implications, while
recognizing the different circumstances of different countries and
respecting the desire of some people not to take part in all technological
developments,
Conscious of the fact that all countries are facing an acceleration in the use
of information and communication technologies and AI technologies, as
well as an increasing need for media and information literacy, and that the
digital economy presents important societal, economic and environmental
challenges and opportunities of benefit-sharing, especially for low- and
middleincome countries (LMICs), including but not limited to least
developed countries (LDCs), landlocked developing countries (LLDCs)
and small island developing States (SIDS), requiring the recognition,
protection and promotion of endogenous cultures, values and knowledge in
order to develop sustainable digital economies,
Further recognizing that AI technologies have the potential to be beneficial
to the environment and ecosystems, and in order for those benefits to be
realized, potential harms to and negative impacts on the environment and
ecosystems should not be ignored but instead addressed, Noting that
addressing risks and ethical concerns should not hamper innovation and
development but rather provide new opportunities and stimulate ethicallyconducted research and innovation that anchor AI technologies in human
rights and fundamental freedoms, values and principles, and moral and
ethical reflection,
Also recalling that in November 2019, the General Conference of
UNESCO, at its 40th session, adopted 40 C/Resolution 37, by which it
mandated the Director-General “to prepare an international standard-setting
549
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
instrument on the ethics of artificial intelligence (AI) in the form of a
recommendation”, which is to be submitted to the General Conference at its
41st session in 2021,
Recognizing that the development of AI technologies necessitates a
commensurate increase in data, media and information literacy as well as
access to independent, pluralistic, trusted sources of information, including
as part of efforts to mitigate risks of misinformation, disinformation and
hate speech, and harm caused through the misuse of personal data,
Observing that a normative framework for AI technologies and its social
implications finds its basis in international and national legal frameworks,
human rights and fundamental freedoms, ethics, need for access to data,
information and knowledge, the freedom of research and innovation,
human and environmental and ecosystem well-being, and connects ethical
values and principles to the challenges and opportunities linked to AI
technologies, based on common understanding and shared aims,
Also recognizing that ethical values and principles can help develop and
implement rights-based policy measures and legal norms, by providing
guidance with a view to the fast pace of technological development,
Also convinced that globally accepted ethical standards for AI
technologies, in full respect of international law, in particular human rights
law, can play a key role in developing AI-related norms across the globe,
Bearing in mind the Universal Declaration of Human Rights (1948), the
instruments of the international human rights framework, including the
Convention Relating to the Status of Refugees (1951), the Discrimination
(Employment and Occupation) Convention (1958), the International
Convention on the Elimination of All Forms of Racial Discrimination
(1965), the International Covenant on Civil and Political Rights (1966), the
International Covenant on Economic, Social and Cultural Rights (1966),
the Convention on the Elimination of All Forms of Discrimination against
Women (1979), the Convention on the Rights of the Child (1989), and the
Convention on the Rights of Persons with Disabilities (2006), the
Convention against Discrimination in Education (1960), the Convention on
the Protection and Promotion of the Diversity of Cultural Expressions
(2005), as well as any other relevant international instruments,
recommendations and declarations,
550
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Also noting the United Nations Declaration on the Right to Development
(1986); the Declaration on the Responsibilities of the Present Generations
Towards Future Generations (1997); the Universal Declaration on
Bioethics and Human Rights (2005); the United Nations Declaration on the
Rights of Indigenous Peoples (2007); the United Nations General Assembly
resolution on the review of the World Summit on the Information Society
(A/RES/70/125) (2015); the United Nations General Assembly Resolution
on Transforming our world: the 2030 Agenda for Sustainable Development
(A/RES/70/1) (2015); the Recommendation Concerning the Preservation
of, and Access to, Documentary Heritage Including in Digital Form (2015);
the Declaration of Ethical Principles in relation to Climate Change (2017);
the Recommendation on Science and Scientific Researchers (2017); the
Internet Universality Indicators (endorsed by UNESCO’s International
Programme for the Development of Communication in 2018), including the
ROAM principles (endorsed by UNESCO’s General Conference in 2015);
the Human Rights Council’s resolution on “The right to privacy in the
digital age” (A/HRC/RES/42/15) (2019); and the Human Rights Council’s
resolution on “New and emerging digital technologies and human rights”
(A/HRC/RES/41/11) (2019),
Emphasizing that specific attention must be paid to LMICs, including but
not limited to LDCs, LLDCs and SIDS, as they have their own capacity but
have been underrepresented in the AI ethics debate, which raises concerns
about neglecting local knowledge, cultural pluralism, value systems and the
demands of global fairness to deal with the positive and negative impacts
of AI technologies,
Also conscious of the many existing national policies, other frameworks
and initiatives elaborated by relevant United Nations entities,
intergovernmental organizations, including regional organizations, as well
as those by the private sector, professional organizations, nongovernmental organizations, and the scientific community, related to the
ethics and regulation of AI technologies,
Further convinced that AI technologies can bring important benefits, but
that achieving them can also amplify tension around innovation,
asymmetric access to knowledge and technologies, including the digital and
civic literacy deficit that limits the public’s ability to engage in topics
related to AI, as well as barriers to access to information and gaps in
capacity, human and institutional capacities, barriers to access to
technological innovation, and a lack of adequate physical and digital
551
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
infrastructure and regulatory frameworks, including those related to data,
all of which need to be addressed,
Underlining that the strengthening of global cooperation and solidarity,
including through multilateralism, is needed to facilitate fair access to AI
technologies and address the challenges that they bring to diversity and
interconnectivity of cultures and ethical systems, to mitigate potential
misuse, to realize the full potential that AI can bring, especially in the area
of development, and to ensure that national AI strategies are guided by
ethical principles,
Taking fully into account that the rapid development of AI technologies
challenges their ethical implementation and governance, as well as the
respect for and protection of cultural diversity, and has the potential to
disrupt local and regional ethical standards and values,
1. Adopts the present Recommendation on the Ethics of Artificial
Intelligence;
2. Recommends that Member States apply on a voluntary basis the
provisions of this Recommendation by taking appropriate steps,
including whatever legislative or other measures may be required, in
conformity with the constitutional practice and governing structures
of each State, to give effect within their jurisdictions to the principles
and norms of the Recommendation in conformity with international
law, including international human rights law;
3. Also recommends that Member States engage all stakeholders,
including business enterprises, to ensure that they play their
respective roles in the implementation of this Recommendation; and
bring the Recommendation to the attention of the authorities, bodies,
research and academic organizations, institutions and organizations
in public, private and civil society sectors involved in AI
technologies, so that the development and use of AI technologies are
guided by both sound scientific research as well as ethical analysis
and evaluation.
I. Scope of Application
1.
This Recommendation addresses ethical issues related to the
domain of Artificial Intelligence to the extent that they are within
UNESCO’s mandate. It approaches AI ethics as a systematic
normative reflection, based on a holistic, comprehensive,
multicultural and evolving framework of interdependent values,
principles and actions that can guide societies in dealing responsibly
552
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
with the known and unknown impacts of AI technologies on human
beings, societies and the environment and ecosystems, and offers
them a basis to accept or reject AI technologies. It considers ethics as
a dynamic basis for the normative evaluation and guidance of AI
technologies, referring to human dignity, well-being and the
prevention of harm as a compass and as rooted in the ethics of science
and technology.
2.
This Recommendation does not have the ambition to
provide one single definition of AI, since such a definition would
need to change over time, in accordance with technological
developments. Rather, its ambition is to address those features of AI
systems that are of central ethical relevance. Therefore, this
Recommendation approaches AI systems as systems which have the
capacity to process data and information in a way that resembles
intelligent behaviour, and typically includes aspects of reasoning,
learning, perception, prediction, planning or control. Three elements
have a central place in this approach:
(a) AI systems are information-processing technologies that
integrate models and algorithms that produce a capacity to learn
and to perform cognitive tasks leading to outcomes such as
prediction and decision-making in material and virtual
environments. AI systems are designed to operate with varying
degrees of autonomy by means of knowledge modelling and
representation and by exploiting data and calculating
correlations. AI systems may include several methods, such as
but not limited to:
(i) machine learning, including deep learning and
reinforcement learning;
(ii) machine reasoning, including planning,
scheduling, knowledge representation and reasoning,
search, and optimization.
AI systems can be used in cyber-physical systems, including
the Internet of things, robotic systems, social robotics, and
human-computer interfaces, which involve control, perception,
the processing of data collected by sensors, and the operation
of actuators in the environment in which AI systems work.
(b) Ethical questions regarding AI systems pertain to all stages
of the AI system life cycle, understood here to range from
research, design and development to deployment and use,
including maintenance, operation, trade, financing, monitoring
553
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
and evaluation, validation, end-of-use, disassembly and
termination. In addition, AI actors can be defined as any actor
involved in at least one stage of the AI system life cycle, and
can refer both to natural and legal persons, such as researchers,
programmers, engineers, data scientists, end-users, business
enterprises, universities and public and private entities, among
others.
(c) AI systems raise new types of ethical issues that include, but
are not limited to, their impact on decision-making,
employment and labour, social interaction, health care,
education, media, access to information, digital divide,
personal data and consumer protection, environment,
democracy, rule of law, security and policing, dual use, and
human rights and fundamental freedoms, including freedom of
expression, privacy and non-discrimination. Furthermore, new
ethical challenges are created by the potential of AI algorithms
to reproduce and reinforce existing biases, and thus to
exacerbate already existing forms of discrimination, prejudice
and stereotyping. Some of these issues are related to the
capacity of AI systems to perform tasks which previously only
living beings could do, and which were in some cases even
limited to human beings only. These characteristics give AI
systems a profound, new role in human practices and society,
as well as in their relationship with the environment and
ecosystems, creating a new context for children and young
people to grow up in, develop an understanding of the world
and themselves, critically understand media and information,
and learn to make decisions. In the long term, AI systems could
challenge humans’ special sense of experience and agency,
raising additional concerns about, inter alia, human selfunderstanding, social, cultural and environmental interaction,
autonomy, agency, worth and dignity.
3.
This Recommendation pays specific attention to the broader
ethical implications of AI systems in relation to the central domains
of UNESCO: education, science, culture, and communication and
information, as explored in the 2019 Preliminary Study on the Ethics
of Artificial Intelligence by the UNESCO World Commission on
Ethics of Scientific Knowledge and Technology (COMEST):
(a) Education, because living in digitalizing societies requires
new educational practices, ethical reflection, critical thinking,
responsible design practices and new skills, given the
554
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
implications for the labour market, employability and civic
participation.
(b) Science, in the broadest sense and including all academic
fields from the natural sciences and medical sciences to the
social sciences and humanities, as AI technologies bring new
research capacities and approaches, have implications for our
concepts of scientific understanding and explanation, and
create a new basis for decision-making.
(c) Cultural identity and diversity, as AI technologies can
enrich cultural and creative industries, but can also lead to an
increased concentration of supply of cultural content, data,
markets and income in the hands of only a few actors, with
potential negative implications for the diversity and pluralism
of languages, media, cultural expressions, participation and
equality.
(d) Communication and information, as AI technologies play an
increasingly important role in the processing, structuring and
provision of information; the issues of automated journalism
and the algorithmic provision of news and moderation and
curation of content on social media and search engines are just
a few examples raising issues related to access to information,
disinformation, misinformation, hate speech, the emergence of
new forms of societal narratives, discrimination, freedom of
expression, privacy and media and information literacy, among
others.
4.
This Recommendation is addressed to Member States, both
as AI actors and as authorities responsible for developing legal and
regulatory frameworks throughout the entire AI system life cycle, and
for promoting business responsibility. It also provides ethical
guidance to all AI actors, including the public and private sectors, by
providing a basis for an ethical impact assessment of AI systems
throughout their life cycle.
II. Aims and Objectives
5.
This Recommendation aims to provide a basis to make AI
systems work for the good of humanity, individuals, societies and the
environment and ecosystems, and to prevent harm. It also aims at
stimulating the peaceful use of AI systems.
6.
In addition to the existing ethical frameworks regarding AI
around the world, this Recommendation aims to bring a globally
accepted normative instrument that focuses not only on the
articulation of values and principles, but also on their practical
555
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
realization, via concrete policy recommendations, with a strong
emphasis on inclusion issues of gender equality and protection of the
environment and ecosystems.
7.
Because the complexity of the ethical issues surrounding AI
necessitates the cooperation of multiple stakeholders across the
various levels and sectors of international, regional and national
communities, this Recommendation aims to enable stakeholders to
take shared responsibility based on a global and intercultural
dialogue.
8.
The objectives of this Recommendation are:
(a) to provide a universal framework of values, principles and
actions to guide States in the formulation of their legislation,
policies or other instruments regarding AI, consistent with
international law;
(b) to guide the actions of individuals, groups, communities,
institutions and private sector companies to ensure the
embedding of ethics in all stages of the AI system life cycle;
(c) to protect, promote and respect human rights and
fundamental freedoms, human dignity and equality, including
gender equality; to safeguard the interests of present and future
generations; to preserve the environment, biodiversity and
ecosystems; and to respect cultural diversity in all stages of the
AI system life cycle;
(d) to foster multi-stakeholder, multidisciplinary and pluralistic
dialogue and consensus building about ethical issues relating to
AI systems;
(e) to promote equitable access to developments and
knowledge in the field of AI and the sharing of benefits, with
particular attention to the needs and contributions of LMICs,
including LDCs, LLDCs and SIDS.
III. Values and Principles
9.
The values and principles included below should be
respected by all actors in the AI system life cycle, in the first place
and, where needed and appropriate, be promoted through
amendments to the existing and elaboration of new legislation,
regulations and business guidelines. This must comply with
international law, including the United Nations Charter and Member
States’ human rights obligations, and should be in line with
internationally agreed social, political, environmental, educational,
556
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
scientific and economic sustainability objectives, such as the United
Nations Sustainable Development Goals (SDGs).
10.
Values play a powerful role as motivating ideals in shaping
policy measures and legal norms. While the set of values outlined
below thus inspires desirable behaviour and represents the
foundations of principles, the principles unpack the values underlying
them more concretely so that the values can be more easily
operationalized in policy statements and actions.
11.
While all the values and principles outlined below are
desirable per se, in any practical contexts, there may be tensions
between these values and principles. In any given situation, a
contextual assessment will be necessary to manage potential tensions,
taking into account the principle of proportionality and in compliance
with human rights and fundamental freedoms. In all cases, any
possible limitations on human rights and fundamental freedoms must
have a lawful basis, and be reasonable, necessary and proportionate,
and consistent with States’ obligations under international law. To
navigate such scenarios judiciously will typically require engagement
with a broad range of appropriate stakeholders, making use of social
dialogue, as well as ethical deliberation, due diligence and impact
assessment.
12.
The trustworthiness and integrity of the life cycle of AI
systems is essential to ensure that AI technologies will work for the
good of humanity, individuals, societies and the environment and
ecosystems, and embody the values and principles set out in this
Recommendation. People should have good reason to trust that AI
systems can bring individual and shared benefits, while adequate
measures are taken to mitigate risks. An essential requirement for
trustworthiness is that, throughout their life cycle, AI systems are
subject to thorough monitoring by the relevant stakeholders as
appropriate. As trustworthiness is an outcome of the
operationalization of the principles in this document, the policy
actions proposed in this Recommendation are all directed at
promoting trustworthiness in all stages of the AI system life cycle.
1) VALUES
Respect, protection and promotion of human rights and fundamental
freedoms and human dignity
13.
The inviolable and inherent dignity of every human
constitutes the foundation for the universal, indivisible, inalienable,
interdependent and interrelated system of human rights and
557
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
fundamental freedoms. Therefore, respect, protection and promotion
of human dignity and rights as established by international law,
including international human rights law, is essential throughout the
life cycle of AI systems. Human dignity relates to the recognition of
the intrinsic and equal worth of each individual human being,
regardless of race, colour, descent, gender, age, language, religion,
political opinion, national origin, ethnic origin, social origin,
economic or social condition of birth, or disability and any other
grounds.
14.
No human being or human community should be harmed or
subordinated, whether physically, economically, socially, politically,
culturally or mentally during any phase of the life cycle of AI
systems. Throughout the life cycle of AI systems, the quality of life
of human beings should be enhanced, while the definition of “quality
of life” should be left open to individuals or groups, as long as there
is no violation or abuse of human rights and fundamental freedoms,
or the dignity of humans in terms of this definition.
15.
Persons may interact with AI systems throughout their life
cycle and receive assistance from them, such as care for vulnerable
people or people in vulnerable situations, including but not limited to
children, older persons, persons with disabilities or the ill. Within
such interactions, persons should never be objectified, nor should
their dignity be otherwise undermined, or human rights and
fundamental freedoms violated or abused.
16.
Human rights and fundamental freedoms must be respected,
protected and promoted throughout the life cycle of AI systems.
Governments, private sector, civil society, international
organizations, technical communities and academia must respect
human rights instruments and frameworks in their interventions in
the processes surrounding the life cycle of AI systems. New
technologies need to provide new means to advocate, defend and
exercise human rights and not to infringe them.
Environment and ecosystem flourishing
17.
Environmental and ecosystem flourishing should be
recognized, protected and promoted through the life cycle of AI
systems. Furthermore, environment and ecosystems are the
existential necessity for humanity and other living beings to be able
to enjoy the benefits of advances in AI.
18.
All actors involved in the life cycle of AI systems must
comply with applicable international law and domestic legislation,
standards and practices, such as precaution, designed for
558
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
environmental and ecosystem protection and restoration, and
sustainable development. They should reduce the environmental
impact of AI systems, including but not limited to its carbon
footprint, to ensure the minimization of climate change and
environmental risk factors, and prevent the unsustainable
exploitation, use and transformation of natural resources contributing
to the deterioration of the environment and the degradation of
ecosystems.
Ensuring diversity and inclusiveness
19.
Respect, protection and promotion of diversity and
inclusiveness should be ensured throughout the life cycle of AI
systems, consistent with international law, including human rights
law. This may be done by promoting active participation of all
individuals or groups regardless of race, colour, descent, gender, age,
language, religion, political opinion, national origin, ethnic origin,
social origin, economic or social condition of birth, or disability and
any other grounds.
20.
The scope of lifestyle choices, beliefs, opinions, expressions
or personal experiences, including the optional use of AI systems and
the co-design of these architectures should not be restricted during
any phase of the life cycle of AI systems.
21.
Furthermore, efforts, including international cooperation,
should be made to overcome, and never take advantage of, the lack
of necessary technological infrastructure, education and skills, as
well as legal frameworks, particularly in LMICs, LDCs, LLDCs and
SIDS, affecting communities.
Living in peaceful, just and interconnected societies
22.
AI actors should play a participative and enabling role to
ensure peaceful and just societies, which is based on an
interconnected future for the benefit of all, consistent with human
rights and fundamental freedoms. The value of living in peaceful and
just societies points to the potential of AI systems to contribute
throughout their life cycle to the interconnectedness of all living
creatures with each other and with the natural environment.
23.
The notion of humans being interconnected is based on the
knowledge that every human belongs to a greater whole, which
thrives when all its constituent parts are enabled to thrive. Living in
peaceful, just and interconnected societies requires an organic,
immediate, uncalculated bond of solidarity, characterized by a
permanent search for peaceful relations, tending towards care for
others and the natural environment in the broadest sense of the term.
559
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
24.
This value demands that peace, inclusiveness and justice,
equity and interconnectedness should be promoted throughout the
life cycle of AI systems, in so far as the processes of the life cycle of
AI systems should not segregate, objectify or undermine freedom and
autonomous decision-making as well as the safety of human beings
and communities, divide and turn individuals and groups against each
other, or threaten the coexistence between humans, other living
beings and the natural environment.
2) PRINCIPLES
Proportionality and Do No Harm
25.
It should be recognized that AI technologies do not
necessarily, per se, ensure human and environmental and ecosystem
flourishing. Furthermore, none of the processes related to the AI
system life cycle shall exceed what is necessary to achieve legitimate
aims or objectives and should be appropriate to the context. In the
event of possible occurrence of any harm to human beings, human
rights and fundamental freedoms, communities and society at large
or the environment and ecosystems, the implementation of
procedures for risk assessment and the adoption of measures in order
to preclude the occurrence of such harm should be ensured.
26.
The choice to use AI systems and which AI method to use
should be justified in the following ways: (a) the AI method chosen
should be appropriate and proportional to achieve a given legitimate
aim; (b) the AI method chosen should not infringe upon the
foundational values captured in this document, in particular, its use
must not violate or abuse human rights; and (c) the AI method should
be appropriate to the context and should be based on rigorous
scientific foundations. In scenarios where decisions are understood to
have an impact that is irreversible or difficult to reverse or may
involve life and death decisions, final human determination should
apply. In particular, AI systems should not be used for social scoring
or mass surveillance purposes.
Safety and security
27.
Unwanted harms (safety risks), as well as vulnerabilities to
attack (security risks) should be avoided and should be addressed,
prevented and eliminated throughout the life cycle of AI systems to
ensure human, environmental and ecosystem safety and security.
Safe and secure AI will be enabled by the development of sustainable,
560
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
privacy-protective data access frameworks that foster better training
and validation of AI models utilizing quality data.
Fairness and non-discrimination
28.
AI actors should promote social justice and safeguard
fairness and non-discrimination of any kind in compliance with
international law. This implies an inclusive approach to ensuring that
the benefits of AI technologies are available and accessible to all,
taking into consideration the specific needs of different age groups,
cultural systems, different language groups, persons with disabilities,
girls and women, and disadvantaged, marginalized and vulnerable
people or people in vulnerable situations. Member States should work
to promote inclusive access for all, including local communities, to
AI systems with locally relevant content and services, and with
respect for multilingualism and cultural diversity. Member States
should work to tackle digital divides and ensure inclusive access to
and participation in the development of AI. At the national level,
Member States should promote equity between rural and urban areas,
and among all persons regardless of race, colour, descent, gender,
age, language, religion, political opinion, national origin, ethnic
origin, social origin, economic or social condition of birth, or
disability and any other grounds, in terms of access to and
participation in the AI system life cycle. At the international level,
the most technologically advanced countries have a responsibility of
solidarity with the least advanced to ensure that the benefits of AI
technologies are shared such that access to and participation in the AI
system life cycle for the latter contributes to a fairer world order with
regard to information, communication, culture, education, research
and socio-economic and political stability.
29.
AI actors should make all reasonable efforts to minimize
and avoid reinforcing or perpetuating discriminatory or biased
applications and outcomes throughout the life cycle of the AI system
to ensure fairness of such systems. Effective remedy should be
available against discrimination and biased algorithmic
determination.
30.
Furthermore, digital and knowledge divides within and
between countries need to be addressed throughout an AI system life
cycle, including in terms of access and quality of access to technology
and data, in accordance with relevant national, regional and
international legal frameworks, as well as in terms of connectivity,
knowledge and skills and meaningful participation of the affected
communities, such that every person is treated equitably.
561
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Sustainability
31.
The development of sustainable societies relies on the
achievement of a complex set of objectives on a continuum of human,
social, cultural, economic and environmental dimensions. The advent
of AI technologies can either benefit sustainability objectives or
hinder their realization, depending on how they are applied across
countries with varying levels of development. The continuous
assessment of the human, social, cultural, economic and
environmental impact of AI technologies should therefore be carried
out with full cognizance of the implications of AI technologies for
sustainability as a set of constantly evolving goals across a range of
dimensions, such as currently identified in the Sustainable
Development Goals (SDGs) of the United Nations.
Right to Privacy, and Data Protection
32.
Privacy, a right essential to the protection of human dignity,
human autonomy and human agency, must be respected, protected
and promoted throughout the life cycle of AI systems. It is important
that data for AI systems be collected, used, shared, archived and
deleted in ways that are consistent with international law and in line
with the values and principles set forth in this Recommendation,
while respecting relevant national, regional and international legal
frameworks.
33.
Adequate data protection frameworks and governance
mechanisms should be established in a multi-stakeholder approach at
the national or international level, protected by judicial systems, and
ensured throughout the life cycle of AI systems. Data protection
frameworks and any related mechanisms should take reference from
international data protection principles and standards concerning the
collection, use and disclosure of personal data and exercise of their
rights by data subjects while ensuring a legitimate aim and a valid
legal basis for the processing of personal data, including informed
consent.
34.
Algorithmic systems require adequate privacy impact
assessments, which also include societal and ethical considerations
of their use and an innovative use of the privacy by design approach.
AI actors need to ensure that they are accountable for the design and
implementation of AI systems in such a way as to ensure that personal
information is protected throughout the life cycle of the AI system.
562
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Human oversight and determination
35.
Member States should ensure that it is always possible to
attribute ethical and legal responsibility for any stage of the life cycle
of AI systems, as well as in cases of remedy related to AI systems, to
physical persons or to existing legal entities. Human oversight refers
thus not only to individual human oversight, but to inclusive public
oversight, as appropriate.
36.
It may be the case that sometimes humans would choose to
rely on AI systems for reasons of efficacy, but the decision to cede
control in limited contexts remains that of humans, as humans can
resort to AI systems in decision-making and acting, but an AI system
can never replace ultimate human responsibility and accountability.
As a rule, life and death decisions should not be ceded to AI systems.
Transparency and explainability
37.
The transparency and explainability of AI systems are often
essential preconditions to ensure the respect, protection and
promotion of human rights, fundamental freedoms and ethical
principles. Transparency is necessary for relevant national and
international liability regimes to work effectively. A lack of
transparency could also undermine the possibility of effectively
challenging decisions based on outcomes produced by AI systems
and may thereby infringe the right to a fair trial and effective remedy,
and limits the areas in which these systems can be legally used.
38.
While efforts need to be made to increase transparency and
explainability of AI systems, including those with extra-territorial
impact, throughout their life cycle to support democratic governance,
the level of transparency and explainability should always be
appropriate to the context and impact, as there may be a need to
balance between transparency and explainability and other principles
such as privacy, safety and security. People should be fully informed
when a decision is informed by or is made on the basis of AI
algorithms, including when it affects their safety or human rights, and
in those circumstances should have the opportunity to request
explanatory information from the relevant AI actor or public sector
institutions. In addition, individuals should be able to access the
reasons for a decision affecting their rights and freedoms, and have
the option of making submissions to a designated staff member of the
private sector company or public sector institution able to review and
correct the decision. AI actors should inform users when a product or
service is provided directly or with the assistance of AI systems in a
proper and timely manner.
563
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
39.
From a socio-technical lens, greater transparency
contributes to more peaceful, just, democratic and inclusive societies.
It allows for public scrutiny that can decrease corruption and
discrimination, and can also help detect and prevent negative impacts
on human rights. Transparency aims at providing appropriate
information to the respective addressees to enable their
understanding and foster trust. Specific to the AI system,
transparency can enable people to understand how each stage of an
AI system is put in place, appropriate to the context and sensitivity of
the AI system. It may also include insight into factors that affect a
specific prediction or decision, and whether or not appropriate
assurances (such as safety or fairness measures) are in place. In cases
of serious threats of adverse human rights impacts, transparency may
also require the sharing of code or datasets.
40.
Explainability refers to making intelligible and providing
insight into the outcome of AI systems. The explainability of AI
systems also refers to the understandability of the input, output and
the functioning of each algorithmic building block and how it
contributes to the outcome of the systems. Thus, explainability is
closely related to transparency, as outcomes and subprocesses
leading to outcomes should aim to be understandable and traceable,
appropriate to the context. AI actors should commit to ensuring that
the algorithms developed are explainable. In the case of AI
applications that impact the end user in a way that is not temporary,
easily reversible or otherwise low risk, it should be ensured that the
meaningful explanation is provided with any decision that resulted in
the action taken in order for the outcome to be considered transparent.
41. Transparency and explainability relate closely to adequate
responsibility and accountability measures, as well as to the
trustworthiness of AI systems.
Responsibility and accountability
42.
AI actors and Member States should respect, protect and
promote human rights and fundamental freedoms, and should also
promote the protection of the environment and ecosystems, assuming
their respective ethical and legal responsibility, in accordance with
national and international law, in particular Member States’ human
rights obligations, and ethical guidance throughout the life cycle of
AI systems, including with respect to AI actors within their effective
territory and control. The ethical responsibility and liability for the
decisions and actions based in any way on an AI system should
564
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
always ultimately be attributable to AI actors corresponding to their
role in the life cycle of the AI system.
43.
Appropriate oversight, impact assessment, audit and due
diligence mechanisms, including whistle-blowers’ protection, should
be developed to ensure accountability for AI systems and their impact
throughout their life cycle. Both technical and institutional designs
should ensure auditability and traceability of (the working of) AI
systems in particular to address any conflicts with human rights
norms and standards and threats to environmental and ecosystem
wellbeing.
Awareness and literacy
44.
Public awareness and understanding of AI technologies and
the value of data should be promoted through open and accessible
education, civic engagement, digital skills and AI ethics training,
media and information literacy and training led jointly by
governments, intergovernmental organizations, civil society,
academia, the media, community leaders and the private sector, and
considering the existing linguistic, social and cultural diversity, to
ensure effective public participation so that all members of society
can take informed decisions about their use of AI systems and be
protected from undue influence.
45.
Learning about the impact of AI systems should include
learning about, through and for human rights and fundamental
freedoms, meaning that the approach and understanding of AI
systems should be grounded by their impact on human rights and
access to rights, as well as on the environment and ecosystems.
Multi-stakeholder and adaptive governance and collaboration
46.
International law and national sovereignty must be
respected in the use of data. That means that States, complying with
international law, can regulate the data generated within or passing
through their territories, and take measures towards effective
regulation of data, including data protection, based on respect for the
right to privacy in accordance with international law and other human
rights norms and standards.
47.
Participation of different stakeholders throughout the AI
system life cycle is necessary for inclusive approaches to AI
governance, enabling the benefits to be shared by all, and to
contribute to sustainable development. Stakeholders include but are
not limited to governments, intergovernmental organizations, the
technical community, civil society, researchers and academia, media,
education, policy-makers, private sector companies, human rights
565
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
institutions and equality bodies, anti-discrimination monitoring
bodies, and groups for youth and children. The adoption of open
standards and interoperability to facilitate collaboration should be in
place. Measures should be adopted to take into account shifts in
technologies, the emergence of new groups of stakeholders, and to
allow for meaningful participation by marginalized groups,
communities and individuals and, where relevant, in the case of
Indigenous Peoples, respect for the self-governance of their data.
IV. Areas for Policy Action
48.
The policy actions described in the following policy areas
operationalize the values and principles set out in this
Recommendation. The main action is for Member States to put in
place effective measures, including, for example, policy frameworks
or mechanisms, and to ensure that other stakeholders, such as private
sector companies, academic and research institutions, and civil
society adhere to them by, among other actions, encouraging all
stakeholders to develop human rights, rule of law, democracy, and
ethical impact assessment and due diligence tools in line with
guidance including the United Nations Guiding Principles on
Business and Human Rights. The process for developing such
policies or mechanisms should be inclusive of all stakeholders and
should take into account the circumstances and priorities of each
Member State. UNESCO can be a partner and support Member States
in the development as well as monitoring and evaluation of policy
mechanisms.
49.
UNESCO recognizes that Member States will be at different
stages of readiness to implement this Recommendation, in terms of
scientific, technological, economic, educational, legal, regulatory,
infrastructural, societal, cultural and other dimensions. It is noted that
“readiness” here is a dynamic status. In order to enable the effective
implementation of this Recommendation, UNESCO will therefore:
(1) develop a readiness assessment methodology to assist interested
Member States in identifying their status at specific moments of their
readiness trajectory along a continuum of dimensions; and (2) ensure
support for interested Member States in terms of developing a
UNESCO methodology for Ethical Impact Assessment (EIA) of AI
technologies, sharing of best practices, assessment guidelines and
other mechanisms and analytical work.
566
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Policy Area 1: Ethical Impact Assessment
50.
Member States should introduce frameworks for impact
assessments, such as ethical impact assessment, to identify and assess
benefits, concerns and risks of AI systems, as well as appropriate risk
prevention, mitigation and monitoring measures, among other
assurance mechanisms. Such impact assessments should identify
impacts on human rights and fundamental freedoms, in particular but
not limited to the rights of marginalized and vulnerable people or
people in vulnerable situations, labour rights, the environment and
ecosystems and ethical and social implications, and facilitate citizen
participation in line with the values and principles set forth in this
Recommendation.
51.
Member States and private sector companies should
develop due diligence and oversight mechanisms to identify, prevent,
mitigate and account for how they address the impact of AI systems
on the respect for human rights, rule of law and inclusive societies.
Member States should also be able to assess the socio-economic
impact of AI systems on poverty and ensure that the gap between
people living in wealth and poverty, as well as the digital divide
among and within countries, are not increased with the massive
adoption of AI technologies at present and in the future. In order to
do this, in particular, enforceable transparency protocols should be
implemented, corresponding to the access to information, including
information of public interest held by private entities. Member States,
private sector companies and civil society should investigate the
sociological and psychological effects of AI-based recommendations
on humans in their decision-making autonomy. AI systems identified
as potential risks to human rights should be broadly tested by AI
actors, including in real-world conditions if needed, as part of the
Ethical Impact Assessment, before releasing them in the market.
52.
Member States and business enterprises should implement
appropriate measures to monitor all phases of an AI system life cycle,
including the functioning of algorithms used for decision making, the
data, as well as AI actors involved in the process, especially in public
services and where direct end-user interaction is needed, as part of
ethical impact assessment. Member States’ human rights law
obligations should form part of the ethical aspects of AI system
assessments.
53.
Governments should adopt a regulatory framework that sets
out a procedure, particularly for public authorities, to carry out ethical
impact assessments on AI systems to predict consequences, mitigate
567
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
risks, avoid harmful consequences, facilitate citizen participation and
address societal challenges. The assessment should also establish
appropriate oversight mechanisms, including auditability,
traceability and explainability, which enable the assessment of
algorithms, data and design processes, as well as include external
review of AI systems. Ethical impact assessments should be
transparent and open to the public, where appropriate. Such
assessments should also be multidisciplinary, multi-stakeholder,
multicultural, pluralistic and inclusive. The public authorities should
be required to monitor the AI systems implemented and/or deployed
by those authorities by introducing appropriate mechanisms and
tools.
Policy Area 2: Ethical Governance and Stewardship
54.
Member States should ensure that AI governance
mechanisms are inclusive, transparent, multidisciplinary, multilateral
(this includes the possibility of mitigation and redress of harm across
borders) and multi-stakeholder. In particular, governance should
include aspects of anticipation, and effective protection, monitoring
of impact, enforcement and redress.
55.
Member States should ensure that harms caused through AI
systems are investigated and redressed, by enacting strong
enforcement mechanisms and remedial actions, to make certain that
human rights and fundamental freedoms and the rule of law are
respected in the digital world and in the physical world. Such
mechanisms and actions should include remediation mechanisms
provided by private and public sector companies. The auditability and
traceability of AI systems should be promoted to this end. In addition,
Member States should strengthen their institutional capacities to
deliver on this commitment and should collaborate with researchers
and other stakeholders to investigate, prevent and mitigate any
potentially malicious uses of AI systems.
56.
Member States are encouraged to develop national and
regional AI strategies and to consider forms of soft governance such
as a certification mechanism for AI systems and the mutual
recognition of their certification, according to the sensitivity of the
application domain and expected impact on human rights, the
environment and ecosystems, and other ethical considerations set
forth in this Recommendation. Such a mechanism might include
different levels of audit of systems, data, and adherence to ethical
guidelines and to procedural requirements in view of ethical aspects.
568
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
At the same time, such a mechanism should not hinder innovation or
disadvantage small and medium enterprises or start-ups, civil society
as well as research and science organizations, as a result of an
excessive administrative burden. These mechanisms should also
include a regular monitoring component to ensure system robustness
and continued integrity and adherence to ethical guidelines over the
entire life cycle of the AI system, requiring re-certification if
necessary.
57.
Member States and public authorities should carry out
transparent self-assessment of existing and proposed AI systems,
which, in particular, should include the assessment of whether the
adoption of AI is appropriate and, if so, should include further
assessment to determine what the appropriate method is, as well as
assessment as to whether such adoption would result in violations or
abuses of Member States’ human rights law obligations, and if that is
the case, prohibit its use.
58.
Member States should encourage public entities, private
sector companies and civil society organizations to involve different
stakeholders in their AI governance and to consider adding the role
of an independent AI Ethics Officer or some other mechanism to
oversee ethical impact assessment, auditing and continuous
monitoring efforts and ensure ethical guidance of AI systems.
Member States, private sector companies and civil society
organizations, with the support of UNESCO, are encouraged to create
a network of independent AI Ethics Officers to give support to this
process at national, regional and international levels.
59.
Member States should foster the development of, and access
to, a digital ecosystem for ethical and inclusive development of AI
systems at the national level, including to address gaps in access to
the AI system life cycle, while contributing to international
collaboration. Such an ecosystem includes, in particular, digital
technologies and infrastructure, and mechanisms for sharing AI
knowledge, as appropriate.
60.
Member States should establish mechanisms, in
collaboration with international organizations, transnational
corporations, academic institutions and civil society, to ensure the
active participation of all Member States, especially LMICs, in
particular LDCs, LLDCs and SIDS, in international discussions
concerning AI governance. This can be through the provision of
funds, ensuring equal regional participation, or any other
mechanisms. Furthermore, in order to ensure the inclusiveness of AI
569
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
fora, Member States should facilitate the travel of AI actors in and
out of their territory, especially from LMICs, in particular LDCs,
LLDCs and SIDS, for the purpose of participating in these fora.
61.
Amendments to the existing or elaboration of new national
legislation addressing AI systems must comply with Member States’
human rights law obligations and promote human rights and
fundamental freedoms throughout the AI system life cycle.
Promotion thereof should also take the form of governance
initiatives, good exemplars of collaborative practices regarding AI
systems, and national and international technical and methodological
guidelines as AI technologies advance. Diverse sectors, including the
private sector, in their practices regarding AI systems must respect,
protect and promote human rights and fundamental freedoms using
existing and new instruments in combination with this
Recommendation.
62.
Member States that acquire Al systems for human rightssensitive use cases, such as law enforcement, welfare, employment,
media and information providers, health care and the independent
judiciary system should provide mechanisms to monitor the social
and economic impact of such systems by appropriate oversight
authorities, including independent data protection authorities,
sectoral oversight and public bodies responsible for oversight.
63.
Member States should enhance the capacity of the judiciary
to make decisions related to AI systems as per the rule of law and in
line with international law and standards, including in the use of AI
systems in their deliberations, while ensuring that the principle of
human oversight is upheld. In case AI systems are used by the
judiciary, sufficient safeguards are needed to guarantee inter alia the
protection of fundamental human rights, the rule of law, judicial
independence as well as the principle of human oversight, and to
ensure a trustworthy, public interest-oriented and human-centric
development and use of AI systems in the judiciary.
64.
Member States should ensure that governments and
multilateral organizations play a leading role in ensuring the safety
and security of AI systems, with multi-stakeholder participation.
Specifically, Member States, international organizations and other
relevant bodies should develop international standards that describe
measurable, testable levels of safety and transparency, so that
systems can be objectively assessed and levels of compliance
determined. Furthermore, Member States and business enterprises
should continuously support strategic research on potential safety and
570
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
security risks of AI technologies and should encourage research into
transparency and explainability, inclusion and literacy by putting
additional funding into those areas for different domains and at
different levels, such as technical and natural language.
65.
Member States should implement policies to ensure that the
actions of AI actors are consistent with international human rights
law, standards and principles throughout the life cycle of AI systems,
while taking into full consideration the current cultural and social
diversities, including local customs and religious traditions, with due
regard to the precedence and universality of human rights.
66.
Member States should put in place mechanisms to require
AI actors to disclose and combat any kind of stereotyping in the
outcomes of AI systems and data, whether by design or by
negligence, and to ensure that training data sets for AI systems do not
foster cultural, economic or social inequalities, prejudice, the
spreading of disinformation and misinformation, and disruption of
freedom of expression and access to information. Particular attention
should be given to regions where the data are scarce.
67.
Member States should implement policies to promote and
increase diversity and inclusiveness that reflect their populations in
AI development teams and training datasets, and to ensure equal
access to AI technologies and their benefits, particularly for
marginalized groups, both from rural and urban zones.
68.
Member States should develop, review and adapt, as
appropriate, regulatory frameworks to achieve accountability and
responsibility for the content and outcomes of AI systems at the
different phases of their life cycle. Member States should, where
necessary, introduce liability frameworks or clarify the interpretation
of existing frameworks to ensure the attribution of accountability for
the outcomes and the functioning of AI systems. Furthermore, when
developing regulatory frameworks, Member States should, in
particular, take into account that ultimate responsibility and
accountability must always lie with natural or legal persons and that
AI systems should not be given legal personality themselves. To
ensure this, such regulatory frameworks should be consistent with the
principle of human oversight and establish a comprehensive
approach focused on AI actors and the technological processes
involved across the different stages of the AI system life cycle.
69.
In order to establish norms where these do not exist, or to
adapt the existing legal frameworks, Member States should involve
all AI actors (including, but not limited to, researchers,
571
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
representatives of civil society and law enforcement, insurers,
investors, manufacturers, engineers, lawyers and users). The norms
can mature into best practices, laws and regulations. Member States
are further encouraged to use mechanisms such as policy prototypes
and regulatory sandboxes to accelerate the development of laws,
regulations and policies, including regular reviews thereof, in line
with the rapid development of new technologies and ensure that laws
and regulations can be tested in a safe environment before being
officially adopted. Member States should support local governments
in the development of local policies, regulations and laws in line with
national and international legal frameworks.
70.
Member States should set clear requirements for AI system
transparency and explainability so as to help ensure the
trustworthiness of the full AI system life cycle. Such requirements
should involve the design and implementation of impact mechanisms
that take into consideration the nature of application domain,
intended use, target audience and feasibility of each particular AI
system.
Policy Area 3: Data Policy
71.
Member States should work to develop data governance
strategies that ensure the continual evaluation of the quality of
training data for AI systems including the adequacy of the data
collection and selection processes, proper data security and
protection measures, as well as feedback mechanisms to learn from
mistakes and share best practices among all AI actors.
72.
Member States should put in place appropriate safeguards
to protect the right to privacy in accordance with international law,
including addressing concerns such as surveillance. Member States
should, among others, adopt or enforce legislative frameworks that
provide appropriate protection, compliant with international law.
Member States should strongly encourage all AI actors, including
business enterprises, to follow existing international standards and,
in particular, to carry out adequate privacy impact assessments, as
part of ethical impact assessments, which take into account the wider
socio-economic impact of the intended data processing, and to apply
privacy by design in their systems. Privacy should be respected,
protected and promoted throughout the life cycle of AI systems.
73.
Member States should ensure that individuals retain rights
over their personal data and are protected by a framework, which
notably foresees: transparency; appropriate safeguards for the
572
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
processing of sensitive data; an appropriate level of data protection;
effective and meaningful accountability schemes and mechanisms;
the full enjoyment of the data subjects’ rights and the ability to access
and erase their personal data in AI systems, except for certain
circumstances in compliance with international law; an appropriate
level of protection in full compliance with data protection legislation
where data are being used for commercial purposes such as enabling
micro-targeted advertising, transferred cross-border; and an effective
independent oversight as part of a data governance mechanism which
keeps individuals in control of their personal data and fosters the
benefits of a free flow of information internationally, including
access to data.
74.
Member States should establish their data policies or
equivalent frameworks, or reinforce existing ones, to ensure full
security for personal data and sensitive data, which, if disclosed, may
cause exceptional damage, injury or hardship to individuals.
Examples include data relating to offences, criminal proceedings and
convictions, and related security measures; biometric, genetic and
health data; and -personal data such as that relating to race, colour,
descent, gender, age, language, religion, political opinion, national
origin, ethnic origin, social origin, economic or social condition of
birth, or disability and any other characteristics.
75.
Member States should promote open data. In this regard,
Member States should consider reviewing their policies and
regulatory frameworks, including on access to information and open
government to reflect AI-specific requirements and promoting
mechanisms, such as open repositories for publicly funded or
publicly held data and source code and data trusts, to support the safe,
fair, legal and ethical sharing of data, among others.
76.
Member States should promote and facilitate the use of
quality and robust datasets for training, development and use of AI
systems, and exercise vigilance in overseeing their collection and use.
This could, if possible and feasible, include investing in the creation
of gold standard datasets, including open and trustworthy datasets,
which are diverse, constructed on a valid legal basis, including
consent of data subjects, when required by law. Standards for
annotating datasets should be encouraged, including disaggregating
data on gender and other bases, so it can easily be determined how a
dataset is gathered and what properties it has.
77.
Member States, as also suggested in the report of the United
Nations Secretary-General’s High-level Panel on Digital
573
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Cooperation, with the support of the United Nations and UNESCO,
should adopt a digital commons approach to data where appropriate,
increase interoperability of tools and datasets and interfaces of
systems hosting data, and encourage private sector companies to
share the data they collect with all stakeholders, as appropriate, for
research, innovation or public benefits. They should also promote
public and private efforts to create collaborative platforms to share
quality data in trusted and secured data spaces.
Policy Area 4: Development and International Cooperation
78.
Member States and transnational corporations should
prioritize AI ethics by including discussions of AI-related ethical
issues into relevant international, intergovernmental and
multistakeholder fora.
79.
Member States should ensure that the use of AI in areas of
development such as education, science, culture, communication and
information, health care, agriculture and food supply, environment,
natural resource and infrastructure management, economic planning
and growth, among others, adheres to the values and principles set
forth in this Recommendation.
80.
Member States should work through international
organizations to provide platforms for international cooperation on
AI for development, including by contributing expertise, funding,
data, domain knowledge, infrastructure, and facilitating multistakeholder collaboration to tackle challenging development
problems, especially for LMICs, in particular LDCs, LLDCs and
SIDS.
81.
Member States should work to promote international
collaboration on AI research and innovation, including research and
innovation centres and networks that promote greater participation
and leadership of researchers from LMICs and other countries,
including LDCs, LLDCs and SIDS.
82.
Member States should promote AI ethics research by
engaging international organizations and research institutions, as
well as transnational corporations, that can be a basis for the ethical
use of AI systems by public and private entities, including research
into the applicability of specific ethical frameworks in specific
cultures and contexts, and the possibilities to develop technologically
feasible solutions in line with these frameworks.
83.
Member States should encourage international cooperation
and collaboration in the field of AI to bridge geo-technological lines.
574
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Technological exchanges and consultations should take place
between Member States and their populations, between the public
and private sectors, and between and among the most and least
technologically advanced countries in full respect of international
law.
Policy Area 5: Environment and Ecosystems
84.
Member States and business enterprises should assess the
direct and indirect environmental impact throughout the AI system
life cycle, including, but not limited to, its carbon footprint, energy
consumption and the environmental impact of raw material
extraction for supporting the manufacturing of AI technologies, and
reduce the environmental impact of AI systems and data
infrastructures. Member States should ensure compliance of all AI
actors with environmental law, policies and practices.
85.
Member States should introduce incentives, when needed
and appropriate, to ensure the development and adoption of rightsbased and ethical AI-powered solutions for disaster risk resilience;
the monitoring, protection and regeneration of the environment and
ecosystems; and the preservation of the planet. These AI systems
should involve the participation of local and indigenous communities
throughout the life cycle of AI systems and should support circular
economy type approaches and sustainable consumption and
production patterns. Some examples include using AI systems, when
needed and appropriate, to:
(a) Support the protection, monitoring and management of natural
resources.
(b) Support the prediction, prevention, control and mitigation of
climate-related problems.
(c) Support a more efficient and sustainable food ecosystem.
(d) Support the acceleration of access to and mass adoption of
sustainable energy.
(e) Enable and promote the mainstreaming of sustainable
infrastructure, sustainable business models and sustainable
finance for sustainable development.
(f) Detect pollutants or predict levels of pollution and thus help
relevant stakeholders identify, plan and put in place targeted
interventions to prevent and reduce pollution and exposure.
86.
When choosing AI methods, given the potential dataintensive or resource-intensive character of some of them and the
respective impact on the environment, Member States should ensure
that AI actors, in line with the principle of proportionality, favour
575
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
data, energy and resource efficient AI methods. Requirements should
be developed to ensure that appropriate evidence is available to show
that an AI application will have the intended effect, or that safeguards
accompanying an AI application can support the justification for its
use. If this cannot be done, the precautionary principle must be
favoured, and in instances where there are disproportionate negative
impacts on the environment, AI should not be used.
Policy Area 6: Gender
87.
Member States should ensure that the potential for digital
technologies and artificial intelligence to contribute to achieving
gender equality is fully maximized, and must ensure that the human
rights and fundamental freedoms of girls and women, and their safety
and integrity are not violated at any stage of the AI system life cycle.
Moreover, Ethical Impact Assessment should include a transversal
gender perspective.
88.
Member States should have dedicated funds from their
public budgets linked to financing gender-responsive schemes,
ensure that national digital policies include a gender action plan, and
develop relevant policies, for example, on labour education, targeted
at supporting girls and women to make sure they are not left out of
the digital economy powered by AI. Special investment in providing
targeted programmes and gender-specific language, to increase the
opportunities of girls’ and women’s participation in science,
technology, engineering, and mathematics (STEM), including
information and communication technologies (ICT) disciplines,
preparedness, employability, equal career development and
professional growth of girls and women, should be considered and
implemented.
89.
Member States should ensure that the potential of AI
systems to advance the achievement of gender equality is realized.
They should ensure that these technologies do not exacerbate the
already wide gender gaps existing in several fields in the analogue
world, and instead eliminate those gaps. These gaps include: the
gender wage gap; the unequal representation in certain professions
and activities; the lack of representation at top management positions,
boards of directors, or research teams in the AI field; the education
gap; the digital and AI access, adoption, usage and affordability gap;
and the unequal distribution of unpaid work and of the caring
responsibilities in our societies.
576
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
90.
Member States should ensure that gender stereotyping and
discriminatory biases are not translated into AI systems, and instead
identify and proactively redress these. Efforts are necessary to avoid
the compounding negative effect of technological divides in
achieving gender equality and avoiding violence such as harassment,
bullying or trafficking of girls and women and under-represented
groups, including in the online domain.
91.
Member States should encourage female entrepreneurship,
participation and engagement in all stages of an AI system life cycle
by offering and promoting economic, regulatory incentives, among
other incentives and support schemes, as well as policies that aim at
a balanced gender participation in AI research in academia, gender
representation on digital and AI companies’ top management
positions, boards of directors and research teams. Member States
should ensure that public funds (for innovation, research and
technologies) are channelled to inclusive programmes and
companies, with clear gender representation, and that private funds
are similarly encouraged through affirmative action principles.
Policies on harassment-free environments should be developed and
enforced, together with the encouragement of the transfer of best
practices on how to promote diversity throughout the AI system life
cycle.
92.
Member States should promote gender diversity in AI
research in academia and industry by offering incentives to girls and
women to enter the field, putting in place mechanisms to fight gender
stereotyping and harassment within the AI research community, and
encouraging academic and private entities to share best practices on
how to enhance gender diversity.
93.
UNESCO can help form a repository of best practices for
incentivizing the participation of girls, women and under-represented
groups in all stages of the AI system life cycle.
Policy Area 7: Culture
94.
Member States are encouraged to incorporate AI systems,
where appropriate, in the preservation, enrichment, understanding,
promotion, management and accessibility of tangible, documentary
and intangible cultural heritage, including endangered languages as
well as indigenous languages and knowledges, for example by
introducing or updating educational programmes related to the
application of AI systems in these areas, where appropriate, and by
ensuring a participatory approach, targeted at institutions and the
public.
577
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
95.
Member States are encouraged to examine and address the
cultural impact of AI systems, especially natural language processing
(NLP) applications such as automated translation and voice
assistants, on the nuances of human language and expression. Such
assessments should provide input for the design and implementation
of strategies that maximize the benefits from these systems by
bridging cultural gaps and increasing human understanding, as well
as addressing the negative implications such as the reduction of use,
which could lead to the disappearance of endangered languages, local
dialects, and tonal and cultural variations associated with human
language and expression.
96.
Member States should promote AI education and digital
training for artists and creative professionals to assess the suitability
of AI technologies for use in their profession, and contribute to the
design and implementation of suitable AI technologies, as AI
technologies are being used to create, produce, distribute, broadcast
and consume a variety of cultural goods and services, bearing in mind
the importance of preserving cultural heritage, diversity and artistic
freedom.
97.
Member States should promote awareness and evaluation of
AI tools among local cultural industries and small and medium
enterprises working in the field of culture, to avoid the risk of
concentration in the cultural market.
98.
Member States should engage technology companies and
other stakeholders to promote a diverse supply of and plural access
to cultural expressions, and in particular to ensure that algorithmic
recommendation enhances the visibility and discoverability of local
content.
99.
Member States should foster new research at the
intersection between AI and intellectual property (IP), for example to
determine whether or how to protect with IP rights the works created
by means of Al technologies. Member States should also assess how
AI technologies are affecting the rights or interests of IP owners,
whose works are used to research, develop, train or implement AI
applications.
100.
Member States should encourage museums, galleries,
libraries and archives at the national level to use AI systems to
highlight their collections and enhance their libraries, databases and
knowledge base, while also providing access to their users.
578
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Policy Area 8: Education and Research
101.
Member States should work with international
organizations, educational institutions and private and nongovernmental entities to provide adequate AI literacy education to the
public on all levels in all countries in order to empower people and
reduce the digital divides and digital access inequalities resulting
from the wide adoption of AI systems.
102.
Member States should promote the acquisition of
“prerequisite skills” for AI education, such as basic literacy,
numeracy, coding and digital skills, and media and information
literacy, as well as critical and creative thinking, teamwork,
communication, socio-emotional and AI ethics skills, especially in
countries and in regions or areas within countries where there are
notable gaps in the education of these skills.
103.
Member States should promote general awareness
programmes about AI developments, including on data and the
opportunities and challenges brought about by AI technologies, the
impact of AI systems on human rights and their implications,
including children’s rights. These programmes should be accessible
to non-technical as well as technical groups.
104.
Member States should encourage research initiatives on the
responsible and ethical use of AI technologies in teaching, teacher
training and e-learning, among other issues, to enhance opportunities
and mitigate the challenges and risks involved in this area. The
initiatives should be accompanied by an adequate assessment of the
quality of education and impact on students and teachers of the use
of AI technologies. Member States should also ensure that AI
technologies empower students and teachers and enhance their
experience, bearing in mind that relational and social aspects and the
value of traditional forms of education are vital in teacher-student and
student-student relationships and should be considered when
discussing the adoption of AI technologies in education. AI systems
used in learning should be subject to strict requirements when it
comes to the monitoring, assessment of abilities, or prediction of the
learners’ behaviours. AI should support the learning process without
reducing cognitive abilities and without extracting sensitive
information, in compliance with relevant personal data protection
standards. The data handed over to acquire knowledge collected
during the learner’s interactions with the AI system must not be
subject to misuse, misappropriation or criminal exploitation,
including for commercial purposes.
579
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
105.
Member States should promote the participation and
leadership of girls and women, diverse ethnicities and cultures,
persons with disabilities, marginalized and vulnerable people or
people in vulnerable situations, minorities and all persons not
enjoying the full benefits of digital inclusion, in AI education
programmes at all levels, as well as the monitoring and sharing of
best practices in this regard with other Member States.
106.
Member States should develop, in accordance with their
national education programmes and traditions, AI ethics curricula for
all levels, and promote cross-collaboration between AI technical
skills education and humanistic, ethical and social aspects of AI
education. Online courses and digital resources of AI ethics education
should be developed in local languages, including indigenous
languages, and take into account the diversity of environments,
especially ensuring accessibility of formats for persons with
disabilities.
107.
Member States should promote and support AI research,
notably AI ethics research, including for example through investing
in such research or by creating incentives for the public and private
sectors to invest in this area, recognizing that research contributes
significantly to the further development and improvement of AI
technologies with a view to promoting international law and the
values and principles set forth in this Recommendation. Member
States should also publicly promote the best practices of, and
cooperation with, researchers and companies who develop AI in an
ethical manner.
108.
Member States should ensure that AI researchers are trained
in research ethics and require them to include ethical considerations
in their designs, products and publications, especially in the analyses
of the datasets they use, how they are annotated, and the quality and
scope of the results with possible applications.
109.
Member States should encourage private sector companies
to facilitate the access of the scientific community to their data for
research, especially in LMICs, in particular LDCs, LLDCs and SIDS.
This access should conform to relevant privacy and data protection
standards.
110.
To ensure a critical evaluation of AI research and proper
monitoring of potential misuses or adverse effects, Member States
should ensure that any future developments with regards to AI
technologies should be based on rigorous and independent scientific
research, and promote interdisciplinary AI research by including
580
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
disciplines other than science, technology, engineering and
mathematics (STEM), such as cultural studies, education, ethics,
international relations, law, linguistics, philosophy, political science,
sociology and psychology.
111.
Recognizing that AI technologies present great
opportunities to help advance scientific knowledge and practice,
especially in traditionally model-driven disciplines, Member States
should encourage scientific communities to be aware of the benefits,
limits and risks of their use; this includes attempting to ensure that
conclusions drawn from data-driven approaches, models and
treatments are robust and sound. Furthermore, Member States should
welcome and support the role of the scientific community in
contributing to policy and in cultivating awareness of the strengths
and weaknesses of AI technologies.
Policy Area 9: Communication and Information
112.
Member States should use AI systems to improve access to
information and knowledge. This can include support to researchers,
academia, journalists, the general public and developers, to enhance
freedom of expression, academic and scientific freedoms, access to
information, and increased proactive disclosure of official data and
information.
113.
Member States should ensure that AI actors respect and
promote freedom of expression as well as access to information with
regard to automated content generation, moderation and curation.
Appropriate frameworks, including regulation, should enable
transparency of online communication and information operators and
ensure users have access to a diversity of viewpoints, as well as
processes for prompt notification to the users on the reasons for
removal or other treatment of content, and appeal mechanisms that
allow users to seek redress.
114.
Member States should invest in and promote digital and
media and information literacy skills to strengthen critical thinking
and competencies needed to understand the use and implication of AI
systems, in order to mitigate and counter disinformation,
misinformation and hate speech. A better understanding and
evaluation of both the positive and potentially harmful effects of
recommender systems should be part of those efforts.
115.
Member States should create enabling environments for
media to have the rights and resources to effectively report on the
benefits and harms of AI systems, and also encourage media to make
ethical use of AI systems in their operations
581
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
Policy Area 10: Economy and Labour
116.
Member States should assess and address the impact of AI
systems on labour markets and its implications for education
requirements, in all countries and with special emphasis on countries
where the economy is labour-intensive. This can include the
introduction of a wider range of “core” and interdisciplinary skills at
all education levels to provide current workers and new generations
a fair chance of finding jobs in a rapidly changing market, and to
ensure their awareness of the ethical aspects of AI systems. Skills
such as “learning how to learn”, communication, critical thinking,
teamwork, empathy, and the ability to transfer one’s knowledge
across domains, should be taught alongside specialist, technical
skills, as well as low-skilled tasks. Being transparent about what
skills are in demand and updating curricula around these are key.
117.
Member States should support collaboration agreements
among governments, academic institutions, vocational education and
training institutions, industry, workers’ organizations and civil
society to bridge the gap of skillset requirements to align training
programmes and strategies with the implications of the future of work
and the needs of industry, including small and medium enterprises.
Project-based teaching and learning approaches for AI should be
promoted, allowing for partnerships between public institutions,
private sector companies, universities and research centres.
118.
Member States should work with private sector companies,
civil society organizations and other stakeholders, including workers
and unions to ensure a fair transition for at-risk employees. This
includes putting in place upskilling and reskilling programmes,
finding effective mechanisms of retaining employees during those
transition periods, and exploring “safety net” programmes for those
who cannot be retrained. Member States should develop and
implement programmes to research and address the challenges
identified that could include upskilling and reskilling, enhanced
social protection, proactive industry policies and interventions, tax
benefits, new taxation forms, among others. Member States should
ensure that there is sufficient public funding to support these
programmes. Relevant regulations, such as tax regimes, should be
carefully examined and changed if needed to counteract the
consequences of unemployment caused by AI-based automation.
119.
Member States should encourage and support researchers to
analyse the impact of AI systems on the local labour environment in
order to anticipate future trends and challenges. These studies should
582
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
have an interdisciplinary approach and investigate the impact of AI
systems on economic, social and geographic sectors, as well as on
human-robot interactions and human-human relationships, in order
to advise on reskilling and redeployment best practices.
120.
Member States should take appropriate steps to ensure
competitive markets and consumer protection, considering possible
measures and mechanisms at national, regional and international
levels, to prevent abuse of dominant market positions, including by
monopolies, in relation to AI systems throughout their life cycle,
whether these are data, research, technology, or market. Member
States should prevent the resulting inequalities, assess relevant
markets and promote competitive markets. Due consideration should
be given to LMICs, in particular LDCs, LLDCs and SIDS, which are
more exposed and vulnerable to the possibility of abuses of market
dominance as a result of a lack of infrastructure, human capacity and
regulations, among other factors. AI actors developing AI systems in
countries which have established or adopted ethical standards on AI
should respect these standards when exporting these products,
developing or applying their AI systems in countries where such
standards may not exist, while respecting applicable international law
and domestic legislation, standards and practices of these countries.
Policy Area 11: Health and Social Wellbeing
121.
Member States should endeavour to employ effective AI
systems for improving human health and protecting the right to life,
including mitigating disease outbreaks, while building and
maintaining international solidarity to tackle global health risks and
uncertainties, and ensure that their deployment of AI systems in
health care be consistent with international law and their human
rights law obligations. Member States should ensure that actors
involved in health care AI systems take into consideration the
importance of a patient’s relationships with their family and with
health care staff.
122.
Member States should ensure that the development and
deployment of AI systems related to health in general and mental
health in particular, paying due attention to children and youth, is
regulated to the effect that they are safe, effective, efficient,
scientifically and medically proven and enable evidence-based
innovation and medical progress. Moreover, in the related area of
digital health interventions, Member States are strongly encouraged
to actively involve patients and their representatives in all relevant
steps of the development of the system.
583
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
123.
Member States should pay particular attention in regulating
prediction, detection and treatment solutions for health care in AI
applications by:
(a) ensuring oversight to minimize and mitigate bias;
(b) ensuring that the professional, the patient, caregiver or service
user is included as a “domain expert” in the team in all relevant
steps when developing the algorithms;
(c) paying due attention to privacy because of the potential need
for being medically monitored and ensuring that all relevant
national and international data protection requirements are met;
(d) ensuring effective mechanisms so that those whose personal
data is being analysed are aware of and provide informed
consent for the use and analysis of their data, without
preventing access to health care;
(e) ensuring the human care and final decision of diagnosis and
treatment are taken always by humans while acknowledging
that AI systems can also assist in their work;
(f) ensuring, where necessary, the review of AI systems by an
ethical research committee prior to clinical use.
124.
Member States should establish research on the effects and
regulation of potential harms to mental health related to AI systems,
such as higher degrees of depression, anxiety, social isolation,
developing addiction, trafficking, radicalization and misinformation,
among others.
125.
Member States should develop guidelines for human-robot
interactions and their impact on human-human relationships, based
on research and directed at the future development of robots, and with
special attention to the mental and physical health of human beings.
Particular attention should be given to the use of robots in health care
and the care for older persons and persons with disabilities, in
education, and robots for use by children, toy robots, chatbots and
companion robots for children and adults. Furthermore, assistance of
AI technologies should be applied to increase the safety and
ergonomic use of robots, including in a human-robot working
environment. Special attention should be paid to the possibility of
using AI to manipulate and abuse human cognitive biases.
126.
Member States should ensure that human-robot interactions
comply with the same values and principles that apply to any other
584
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
AI systems, including human rights and fundamental freedoms, the
promotion of diversity, and the protection of vulnerable people or
people in vulnerable situations. Ethical questions related to AIpowered systems for neuro technologies and brain-computer
interfaces should be considered in order to preserve human dignity
and autonomy.
127.
Member States should ensure that users can easily identify
whether they are interacting with a living being, or with an AI system
imitating human or animal characteristics, and can effectively refuse
such interaction and request human intervention.
128.
Member States should implement policies to raise
awareness about the anthropomorphization of AI technologies and
technologies that recognize and mimic human emotions, including in
the language used to mention them, and assess the manifestations,
ethical implications and possible limitations of such
anthropomorphization, in particular in the context of robot-human
interaction and especially when children are involved.
129.
Member States should encourage and promote collaborative
research into the effects of longterm interaction of people with AI
systems, paying particular attention to the psychological and
cognitive impact that these systems can have on children and young
people. This should be done using multiple norms, principles,
protocols, disciplinary approaches, and assessment of the
modification of behaviours and habits, as well as careful evaluation
of the downstream cultural and societal impacts. Furthermore,
Member States should encourage research on the effect of AI
technologies on health system performance and health outcomes.
130.
Member States, as well as all stakeholders, should put in
place mechanisms to meaningfully engage children and young people
in conversations, debates and decision-making with regard to the
impact of AI systems on their lives and futures.
V. Monitoring and Evaluation
131.
Member States should, according to their specific
conditions, governing structures and constitutional provisions,
credibly and transparently monitor and evaluate policies,
programmes and mechanisms related to ethics of AI, using a
combination of quantitative and qualitative approaches. To support
Member States, UNESCO can contribute by:
(a) developing a UNESCO methodology for Ethical Impact
Assessment (EIA) of AI technologies based on rigorous
scientific research and grounded in international human rights
585
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
law, guidance for its implementation in all stages of the AI
system life cycle, and capacity-building materials to support
Member States’ efforts to train government officials, policymakers and other relevant AI actors on EIA methodology;
(b) developing a UNESCO readiness assessment methodology
to assist Member States in identifying their status at specific
moments of their readiness trajectory along a continuum of
dimensions;
(c) developing a UNESCO methodology to evaluate ex ante
and ex post the effectiveness and efficiency of the policies for
AI ethics and incentives against defined objectives;
(d) strengthening the research- and evidence-based analysis of
and reporting on policies regarding AI ethics;
(e) collecting and disseminating progress, innovations, research
reports, scientific publications, data and statistics regarding
policies for AI ethics, including through existing initiatives, to
support sharing best practices and mutual learning, and to
advance the implementation of this Recommendation.
132.
Processes for monitoring and evaluation should ensure
broad participation of all stakeholders, including, but not limited to,
vulnerable people or people in vulnerable situations. Social, cultural
and gender diversity should be ensured, with a view to improving
learning processes and strengthening the connections between
findings, decision-making, transparency and accountability for
results.
133.
In the interests of promoting best policies and practices
related to ethics of AI, appropriate tools and indicators should be
developed for assessing the effectiveness and efficiency thereof
against agreed standards, priorities and targets, including specific
targets for persons belonging to disadvantaged, marginalized
populations, and vulnerable people or people in vulnerable situations,
as well as the impact of AI systems at individual and societal levels.
The monitoring and assessment of the impact of AI systems and
related AI ethics policies and practices should be carried out
continuously in a systematic way proportionate to the relevant risks.
This should be based on internationally agreed frameworks and
involve evaluations of private and public institutions, providers and
programmes, including self-evaluations, as well as tracer studies and
the development of sets of indicators. Data collection and processing
should be conducted in accordance with international law, national
586
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
legislation on data protection and data privacy, and the values and
principles outlined in this Recommendation.
134.
In particular, Member States may wish to consider possible
mechanisms for monitoring and evaluation, such as an ethics
commission, AI ethics observatory, repository covering human
rights-compliant and ethical development of AI systems, or
contributions to existing initiatives by addressing adherence to
ethical principles across UNESCO’s areas of competence, an
experience-sharing mechanism, AI regulatory sandboxes, and an
assessment guide for all AI actors to evaluate their adherence to
policy recommendations mentioned in this document.
VI. Utilization and Exploitation of the Present Recommendation
135.
Member States and all other stakeholders as identified in
this Recommendation should respect, promote and protect the ethical
values, principles and standards regarding AI that are identified in
this Recommendation, and should take all feasible steps to give effect
to its policy recommendations.
136.
Member States should strive to extend and complement
their own action in respect of this Recommendation, by cooperating
with all relevant national and international governmental and nongovernmental organizations, as well as transnational corporations
and scientific organizations, whose activities fall within the scope and
objectives of this Recommendation. The development of a UNESCO
Ethical Impact Assessment methodology and the establishment of
national commissions for the ethics of AI can be important
instruments for this.
VII. Promotion of the Present Recommendation
137.
UNESCO has the vocation to be the principal United
Nations agency to promote and disseminate this Recommendation,
and accordingly will work in collaboration with other relevant United
Nations entities, while respecting their mandate and avoiding
duplication of work.
138.
UNESCO, including its bodies, such as the World
Commission on the Ethics of Scientific Knowledge and Technology
(COMEST), the International Bioethics Committee (IBC) and the
Intergovernmental Bioethics Committee (IGBC), will also work in
collaboration with other international, regional and sub-regional
governmental and non-governmental organizations.
139.
Even though, within UNESCO, the mandate to promote and
protect falls within the authority of governments and
587
Artificial Intelligence and Democratic Values 2021
Center for AI and Digital Policy
intergovernmental bodies, civil society will be an important actor to
advocate for the public sector’s interests and therefore UNESCO
needs to ensure and promote its legitimacy.
VIII. Final Provisions
140.
This Recommendation needs to be understood as a whole,
and the foundational values and principles are to be understood as
complementary and interrelated.
141.
Nothing in this Recommendation may be interpreted as
replacing, altering or otherwise prejudicing States’ obligations or
rights under international law, or as approval for any State, other
political, economic or social actor, group or person to engage in any
activity or perform any act contrary to human rights, fundamental
freedoms, human dignity and concern for the environment and
ecosystems, both living and non-living.
588
The Center for AI and Digital Policy
The Center for AI and Digital Policy aims to promote a better
society, more fair, more just —a world where technology promotes broad
social inclusion based on fundamental rights, democratic institutions, and
the rule of law. The Center advises national governments and international
organizations on AI policies and practices; publishes commentaries on AI
policy; publishes annually Artificial Intelligence and Democratic Values;
provides training and certification for future leaders in the AI policy field;
organizes educational events with AI policy experts; and supports AI policy
initiatives that promote human-centric and trustworthy AI, including the
Universal Guidelines for AI. The Center also monitors implementation of
the OECD AI Principles and other AI policy frameworks; and supports the
establishment of new legal frameworks for AI that safeguard the rule of law,
democratic institutions, and fundamental rights. More information about the
Center is available at CAIDP.ORG.
______________________________
About this Report
Artificial Intelligence and Democratic Values – the essential companion
to AI policy practitioners, academics, and civil society (CAIDP 2022)
Reflecting the collaborative work of more than 100 AI policy
experts around the world, AI and Democratic Values is the first
comprehensive survey of national AI policies and practices, with country
ratings and rankings. The report sets out a state-of-the-art analysis of 50
counties worldwide, based on a rigorous and transparent methodology. AI
and Democratic Values provides an up-to-date review of the AI global
policy landscape, describing recent developments at the European Union
and the Council of Europe, as well as changes to national AI strategies, and
current controversies surrounding social scoring, facial recognition, and
lethal autonomous weapons. AI and Democratic Values also features the
primary legal instruments for AI policy, including the Universal Guidelines
for AI, the OECD AI Principles, the resolutions of the Global Privacy
Assembly concerning AI, and the new UNESCO Recommendation on AI
Ethics. Published in the EPUB format, AI and Democratic Values offers
free and easy access to more than 1,000 policy documents in the field of AI
policy.