S Gen Unact 2023 PDF e

ITUPublications International Telecommunication Union
United Nations Activities on

Artificial Intelligence (AI)
2023
United Nations Activities on
Artificial Intelligence (AI)
2023
DISCLAIMER
The designations employed and the presentation of the material in this publication do not imply the expression of any
opinion whatsoever on the part of ITU and of the Secretariat of the ITU, or CTBTO, FAO, IAEA, IFAD, ILO, IMF, IMO, IOM,
ITC, Office of the Secretary-General's Envoy on Technology, OHCHR, UN Global Pulse, UN Habitat, UN Tourism, UN
Women, UNAIDS, UNCTAD, UNDESA, UNDP, UNDPPA, UNECE, UNEP, UNESCO, UNFCCC, UNFPA, UNHCR, UNICC,
UNICEF, UNICRI, UNIDIR, UNIDO, UNITAR, UNJSPF, UNOCHA, UNOCT, UNODA, UNODC, UNOOSA, UNRISD, UNU,
UPU, WFP, WHO, WIPO, WMO and World Bank Group, concerning the legal status of any country, territory, city or area
or of its authorities, or concerning the delimitation of its frontiers or boundaries.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or
recommended by ITU or CTBTO, FAO, IAEA, IFAD, ILO, IMF, IMO, IOM, ITC, Office of the Secretary-General's Envoy
on Technology, OHCHR, UN Global Pulse, UN Habitat, UN Tourism, UN Women, UNAIDS, UNCTAD, UNDESA, UNDP,
UNDPPA, UNECE, UNEP, UNESCO, UNFCCC, UNFPA, UNHCR, UNICC, UNICEF, UNICRI, UNIDIR, UNIDO, UNITAR,
UNJSPF, UNOCHA, UNOCT, UNODA, UNODC, UNOOSA, UNRISD, UNU, UPU, WFP, WHO, WIPO, WMO and World
Bank Group in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names
of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by ITU to verify the information contained in this publication. However, the
published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for
the interpretation and use of the material lies with the reader.
The opinions, findings and conclusions expressed in this publication do not necessarily reflect the views of ITU or CTBTO,
FAO, IAEA, IFAD, ILO, IMF, IMO, IOM, ITC, Office of the Secretary-General's Envoy on Technology, OHCHR, UN Global
Pulse, UN Habitat, UN Tourism, UN Women, UNAIDS, UNCTAD, UNDESA, UNDP, UNDPPA, UNECE, UNEP, UNESCO,
UNFCCC, UNFPA, UNHCR, UNICC, UNICEF, UNICRI, UNIDIR, UNIDO, UNITAR, UNJSPF, UNOCHA, UNOCT, UNODA,
UNODC, UNOOSA, UNRISD, UNU, UPU, WFP, WHO, WIPO, WMO and World Bank Group or their membership.
ISBN
978-92-61-39001-3 (Electronic version)

978-92-61-39011-2 (EPUB version)
978-92-61-39021-1 (Mobi version)
Please consider the environment before printing this report.
© ITU 2024
Some rights reserved. This work is licensed to the public through a Creative Commons Attribution-Non-
Commercial-Share Alike 3.0 IGO license (CC BY-NC-SA 3.0 IGO).
Under the terms of this licence, you may copy, redistribute and adapt the work for non-commercial purposes,
provided the work is appropriately cited. In any use of this work, there should be no suggestion that ITU
endorse any specific organization, products or services. The unauthorized use of the ITU names or logos is
not permitted. If you adapt the work, then you must license your work under the same or equivalent Creative
Commons licence. If you create a translation of this work, you should add the following disclaimer along
with the suggested citation: “This translation was not created by the International Telecommunication Union
(ITU). ITU is not responsible for the content or accuracy of this translation. The original English edition shall
be the binding and authentic edition”. For more information, please visit https://creativecommons.org/
licenses/by-nc-sa/3.0/igo/
Foreword
2023 will go down in history as the year that generative artificial

intelligence — also known as “gen AI” — took the world by storm.
Public perceptions oscillated between shock, awe, and skepticism.
Companies began to rethink their business models amid projections

of significant market growth.1 And countries — especially in the
developing world — looked to the United Nations (UN), and to the
International Telecommunication Union (ITU), for guidance on how
to make this transformative technology work for their benefit. At the
same time, a third of humanity remains offline worldwide. As the
pace of AI innovation continues unabated, we must avoid creating
yet another digital divide. This requires striking a careful balance
between innovation, regulation, and inclusion.
This report — a joint effort between ITU and 46 UN agencies and bodies — features 408 projects
covering all 17 Sustainable Development Goals (SDGs). Seven UN entities are making their debut
this year. About a third involve collaboration with other UN entities, 24% with governments,
20% with academia, and 17% with the private sector.Every project within these pages shows
the commitment of the UN family to help AI reach its potential to deliver on the 2030 Agenda.
This UN-led innovation takes many different forms, from improving health outcomes using large
language models to mapping groundwater stores using satellite data and AI. nd it comes at
a time when AI governance efforts are ramping up inside and outside the UN. A recent white
paper developed through a collaborative effort led by ITU and UNESCO identifies over 50
existing policy instruments within the UN system that can inform and shape AI governance.
Read on to discover how we are working as one UN to ensure AI benefits everyone, accelerates
the SDGs, and ultimately advances our shared mission of peace, dignity, and prosperity on a
healthy planet.
Doreen Bogdan-Martin
ITU Secretary-General
1
The generative artificial intelligence market is expected to see a CAGR of over 24.4% from 2023 to 2030.
https://www.statista.com/outlook/tmo/artificial-intelligence/generative-ai/worldwide#analyst-opinion
iii
Executive Summary
Highlights
• 47 entities participated, 408 projects reported.
• An increase of 18% in participating entities and 45% in projects reported.
• Strong focus continues to be maintained by the reported projects on SDGs 9
(Industry, Innovation and Infrastructure), 10 (Reduced Inequalities), 16 (Peace,
justice and strong institutions) and 17 (Partnership for the Goals), with SDG 13
(Climate Action) returning to the top 5 list again, similar to the 2021 edition of
the Report. In 2022, SDG 3 (Good Health and Well-being) had featured among
the top five SDGs and while focus remains consistent on it this year as well, the
number of projects reporting work on SDG 13 (Climate Action) has increased,
bringing it back to the top 5 SDGs addressed.
• More focus continues to be needed on SDGs 6 (Clean Water and Sanitation), 7
(Affordable and Clean Energy), 12 (Responsible consumption and production), 14
(Life Below Water) and 15 (Life on Land), which is a finding that remains consistent
with the 2022 edition of the Report.
• Multi-stakeholder collaborations continue to remain a priority for the UN system.
Nearly 60% of the UN projects have reported collaborations with the UN
system, the private sector, governments, civil society, academia, or with another
international organization, demonstrating the UN’s focus on maintaining strong
partnerships with internal and external stakeholders.
• Consistent with the findings of the 2022 edition, the most common outputs of the
UN system’s work on AI remain reports and software tools like mobile applications,
web applications, searchable dashboards, and generative AI-powered chatbots.
Besides, about 13% indicated a focus on creating datasets.
Urgent action is needed to achieve the Sustainable Development Goals (SDGs) by 2030. With
the potential to drive progress across all 17 SDGs, the use of Artificial Intelligence (AI) can help
speed and scale interventions for this purpose.
Recognizing this, the different bodies, agencies, offices and departments of the UN system have
been exploring ways to leverage the potential of AI to drive change and impact across their issue
areas. In 2020, the UN System Chief Executives Board for Coordination (CEB) and its High-Level
Committee on Programmes (HLCP) established the Interagency Working Group on AI (IAWG-
AI), co-led by ITU and UNESCO, to bring together UN system expertise on AI in support of the
CEB and HLCP workstreams on the ethics of AI (led by UNESCO) and the strategic approach
and road map for supporting capacity development (led by ITU), and the related gap analysis
effort carried out by ITU, informed by the UN Activities on AI Report, to identify the gaps in UN
AI-related activities in order to help the UN system prioritize strategic actions.
Since 2021, the IAWG-AI has successfully galvanized expertise from across the UN system as well
as external stakeholder groups to advance the responsible development and use of AI in the UN,
underpinned by ethics and human rights, while driving forward the 2030 Agenda on Sustainable
Development. As part of the IAWG-AI, UNESCO and OICT have led the development of the
Principles for the Ethical Use of Artificial Intelligence in the United Nations System, which
were based on UNESCO’s Ethics of AI Recommendation and endorsed by the HLCP at its 43rd
iv
session and the CEB in 2022. In addition, the IAWG-AI United Nations System White Paper on
AI Governance, which analyses the UN system's institutional models, functions, and existing
international normative frameworks applicable to global AI governance, has been approved by
HLCP at its 47th session and was subsequently endorsed by the CEB in May 2024.
In 2023, the AI for Good platform, organized by ITU in partnership with more than 40 UN Sister
Agencies and co-convened with Switzerland, reached over 1 million online views and grew its
online community to over 100,000. The 2023 AI for Good Global Summit was joined by 2,500
in-person attendees and 10,000 online audience1. This multi-stakeholder community includes
representatives from 180+ countries and has consistently attracted broad based international
media coverage, making it the leading action-oriented, global and inclusive United Nations
platform on AI.
The AI for Good Neural Network continues to grow: an AI-powered community networking
and content platform designed to help users build connections with innovators and experts,
link innovative ideas with social impact opportunities, and bring the community together to
advance the SDGs using AI. AI for Good Partners showcase their work on the Neural Network
through weekly live sessions, virtual exhibitions, networking features and interactive content.
UN Partners have also created “poster boards” in the “UN SDG Zone” of the platform to virtually
exhibit their work on AI, viewable by all Neural Network users, and open for interaction with
users via the booth wall and smart matching system. The chapters of this Report are available
on the respective poster boards of the UN entities as well.
Complementary to these efforts, ITU, the UN’s specialized agency on digital technologies,
has been coordinating the compilation of an annual up-to-date directory since 2018 of all the
AI-related projects, initiatives, events and processes that are being carried out within the UN
system in the form of the UN Activities on AI Report.
1
This publication is being released during the 2024 AI for Good Summit – which is expected to have a
significantly larger audience.
v
Methodology:
• This edition’s chapters were updated by each of the participating UN bodies and
agencies respectively, based on the 2022 UN Activities on Artificial Intelligence
Report.
• New chapters were drafted for the 7 additional entities who have contributed to
this Report for the first time.
• Updated inputs were not provided by 7 entities for this edition. Their chapters
and projects have been retained from the 2022 edition in this Report.
• All submissions received this year have been harmonized and formatted for the
purpose of reproduction in the Report.
• For the Executive Summary, select data points were extracted from the
submissions received to develop an analysis along five specific indicators: SDGs
addressed, multi-stakeholder collaborations, types of projects, sectoral focus,
and project status.
• All submissions received by 30 April 2024 have been included in the Executive
Summary analysis.
The Report is a joint effort between ITU and 46 other UN agencies and bodies, all partners of
AI for Good and members of the UN Interagency Working Group on AI. It presents use cases,
projects and initiatives run by the UN system, in areas covering all 17 SDGs and ranging from
smart agriculture and food systems to transportation, financial services, and healthcare. This
Report is not intended to produce an exhaustive inventory of the UN system’s work on AI.
Rather, it is a tool to further collaboration and build common understanding around emerging
AI technologies and solutions.
In the 2023 edition, 47 UN entities were engaged, 7 entities for the first time, and 408 projects
have been presented. A brief analysis of the key tracks and trends arising from the submissions
is provided below to help provide an overview of the extensive work that is taking place within
the UN system as well as to assist UN bodies and agencies to identify the areas which could
benefit from increased interventions as they develop future projects.
With the increased number of projects submitted for the Report this year, along with the updates
made to the existing ones, the findings of this edition remain substantially consistent with those
of the 2022 edition of this Report. Some of the fluctuations in the analysis detailed below are
likely attributable to standard deviations occurring due to changes in personnel and the level
of reporting annually. Significantly, the substantial increase in project numbers demonstrates
the emphasis that is now being placed across the UN system on leveraging the potential of AI
across several issue areas through concrete projects.
Relative stability on the key thematic priorities and SDGs across the 2022 and 2023 editions
indicates a steadiness of efforts from entities in those areas. It could also be emblematic of the
demand that is emanating from Member States and stakeholders for UN action and intervention
across these topics. In particular, the renewed emphasis on SDG 13 (Climate Action), which
aligns the priority SDGs once again with those of the 2021 Report, is illustrative of the critical
timelines associated with achieving this SDG, the level of investment and effort needed to move
the needle in this respect, as well as a potential recognition of the capacity of the technology
to drive progress. Future editions of the Report will aim to continue analyzing these trends with
an aim to provide insight into the nature and status of the UN AI activities.
vi
This Report is being released at the 2024 edition of the annual AI for Good Global Summit.
Key tracks and trends

Key tracks and trends
1. SDG Mapping
1. SDG Mapping
Number of projects reporting work on a specific SDG

200
150
100
50
0
SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Around 83% of the submissions have linked their projects with outcomes driving forward specific SDGs.
AmongAround
them,83% of the submissions
the overwhelming haveaddress
majority linked their
more projects withSDG,
than one outcomes driving
signaling forward
holistic, specific
multidimensional
SDGs.
projects. Among them, the overwhelming majority address more than one SDG, signaling holistic,
multidimensional projects.
SDGs 9 (Industry, Innovation and Infrastructure), 10 (Reduced Inequalities), 16 (Peace, Justice and Strong
Institutions),
SDGs 9 and 17 (Partnership
(Industry, Innovation for
andthe Goals) continue
Infrastructure), to remainInequalities),
10 (Reduced among the 16 top(Peace,
five most common
Justice
SDGs and
addressed
Strong by the UN AIand
Institutions), projects. However, for
17 (Partnership thisthe
year’s edition
Goals) reports
continue that while
to remain among focus
theontopSDG 3
(Goodfive
Health and Well-being) remains consistent, there is an increase in the number
most common SDGs addressed by the UN AI projects. However, this year’s edition reportsof projects reporting
on SDG 13 (Climate
that while focus Action)
on SDGbringing
3 (Good it up toand
Health theWell-being)
list of the top five consistent,
remains most common thereSDGs addressed by
is an increase
projects this year. Meanwhile, similar to the last edition, more targeted action could be taken across SDGs
in the number of projects reporting on SDG 13 (Climate Action) bringing it up to the list of the
6 (Clean Water and Sanitation), 7 (Affordable and Clean Energy), 12 (Responsible consumption and
top five most common SDGs addressed by projects this year. Meanwhile, similar to the last
production), 14 (Life Below Water) and 15 (Life on Land).
edition, more targeted action could be taken across SDGs 6 (Clean Water and Sanitation), 7
2. Project Subjectand
(Affordable Areas
Clean Energy), 12 (Responsible consumption and production), 14 (Life Below
Water) and 15 (Life on Land).
4
vii
2. Project Subject Areas
In addition to the SDG mapping, participants have also set out some of the issue areas within
which their AI projects are operating. Nearly 83% of the projects reported their specific issue
areas. In most cases, the projects are often reported as addressing multiple areas. Consistent with
the 2022 edition, human rights, ethics and justice, environment, agriculture, health, education,
gender, and telecommunications are tagged as priority subject areas in this edition. In particular,
projects focusing on agriculture and environment have increased since 2022 which correlates
to the overall increased focus on SDG 13 (Climate Action) in this year’s projects.
viii
In addition to the SDG mapping, participants have also set out some of the issue areas within which their
AI projects are operating. Nearly 83% of the projects reported their specific issue areas. In most cases, the
projects are often reported as addressing multiple areas. Consistent with the 2022 edition, human rights,
ethics and justice, environment, agriculture, health, education, gender, and telecommunications are
tagged as priority subject areas in this edition. In particular, projects focusing on agriculture and
environment have increased since 2022 which correlates to the overall increased focus on SDG 13 (Climate
Action) in this year’s projects.
3. Driving Multi-stakeholder Collaboration
3. Driving Multi-stakeholder Collaboration
Number of projects reporting collaborations with partner type
academia
civil society
government
other international organization
private sector
UN system
0 20 40 60 80 100 120 140
Proportion of project collaboration by SDG

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
SDG 1 SDG 2 SDG 3 SDG 4 SDG 5 SDG 6 SDG 7 SDG 8 SDG 9 SDG SDG SDG SDG SDG SDG SDG SDG
10 11 12 13 14 15 16 17
UN system private sector other international organization government civil society academia
Almost 60% of the UN projects have reported collaborations with the UN system, the private sector,
Almost 60% of
governments, the
civil UN projects
society, academia,have reported
or with anothercollaborations with the UN
international organization, system, thethe
demonstrating private
UN’s
focus ongovernments,
sector, maintaining strongcivilpartnerships with internal
society, academia, orand external
with anotherstakeholders. A breakdown
international of the
organization,
collaborations demonstrates
demonstrating the UN’s focus on maintaining strong partnerships with internal and external
stakeholders. A breakdown of the collaborations demonstrates
5
• Nearly 30% of the projects reporting collaborations with entities within the UN system;
and
• 24% of the projects reporting collaborations with government, 20% with academia, and
17% with the private sector.
In particular, the data shows a 33% increase in projects collaborating with governments in
comparison to the 2022 edition.
ix
4. Reports and software tools to address challenges
Project Outputs by SDGs
(Doughnut size = number of projects)
About 74% of the projects reported their project types or outputs this year. Among them,
consistent with the findings of the 2022 edition, a significant number have focused on outcome-
driven products and deliverables such as reports or software tools like mobile applications, web
applications, searchable dashboards, or generative AI-powered chatbots. Besides, about 13%
indicated a focus on creating datasets.
As of now, many of the current software tools produced on AI relate to SDGs 9 (Industry,
Innovation, and Infrastructure) and 13 (Climate action) as compared to 2022’s focus on SDGs
3 (Good health and well-being) and 9 (Industry, Innovation, and Infrastructure). Many reports
relate to SDG 10 (Reduced Inequalities), which is consistent with the last edition. Focus on SDGs
16 (Peace, justice and strong institutions) and 17 (Partnership for the goals) remains strong
across all the project outputs.
x
5. Looking forward
Project Status by SDG
While over 90% of the projects reported their current status, a number of projects did not report
specific end dates, or in some cases, the project duration.
Proportion of Projects by Reporting Status
other
9% completed
22%
development
4%
ongoing
65%
completed development ongoing other
Proportion of projects by SDGs by current status

100%
80%
60%
40%
20%
0%
SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG SDG
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
completed development ongoing other
In terms of status of the projects (in development, ongoing or completed), the majority of the reported
UN AIInprojects
terms ofare
status of the ongoing,
currently projects (in
i.e.development,
65% projects ongoing
comparedortocompleted), thewith
57% in 2022, majority
thoseofrelated
the to
SDG reported
8 (DecentUNworkAI and economic
projects growth)ongoing,
are currently and SDG i.e.
17 65%
(Partnership
projectsfor the goals)
compared to reporting the most
57% in 2022,
completed projects.
with those related to SDG 8 (Decent work and economic growth) and SDG 17 (Partnership for
the goals) reporting the most completed projects.
Of the projects in development, the majority feature software tools that are working on topics such as
intellectual
Of the property,
projects inenvironment,
development,human rights, ethics,
the majority featureand justice.
software tools that are working on topics
such as intellectual property, environment, human rights, ethics, and justice.
xi
7
Table of contents
Foreword�� iii
Executive Summary��iv
Key tracks and trends��vii
Comprehensive Nuclear-Test-Ban Treaty Organization�� 1
1. Description of Activities on AI�� 1

2. Related Sustainable Development Goals�� 3
3. Relevant links�� 3
The Food and Agriculture Organization�� 4

3. Relevant Links �� 11
International Atomic Energy Agency��12

3. Relevant Links�� 24
International Fund for Agricultural Development��25

International Labour Organisation��36

International Monetary Fund��44

International Maritime Organization��45
1. Description of Activities on AI �� 45
xii
International Organization for Migration�� 47

International Trade Centre�� 50

International Telecommunication Union�� 52

Office of the United Nations High Commissioner for Human Rights�� 71

Office of the Secretary-General's Envoy on Technology�� 80

United Nations Global Pulse�� 81

United Nations Human Settlements Programme�� 92

United Nations World Tourism Organization�� 98
xiii
3. Relevant Link�� 100
United Nations Women�� 101

The Joint United Nations Programme on HIV/AIDS�� 104

United Nations Conference on Trade and Development�� 106

2. Relevant Sustainable Development Goals�� 109
United Nations Department of Economic and Social Affairs�� 110

United Nations Development Programme�� 118

United Nations Department of Political and Peacebuilding Affairs and

Department of Peace Operations�� 149

United Nations Economic Commission for Europe�� 152

xiv
United Nations Environment Programme�� 163

2. Related Sustainable Development Goals �� 172
3. Relevant links �� 172
United Nations Educational, Scientific and Cultural Organization�� 173

3. Relevant links �� 199
United Nations Framework Convention on Climate Change�� 200

United Nations Population Fund�� 202

United Nations High Commissioner for Refugees�� 218

United Nations International Computing Centre�� 227

United Nations Children’s Fund�� 234

United Nations Interregional Crime and Justice Research Institute�� 251

xv
United Nations Institute for Disarmament Research�� 263

United Nations Industrial Development Organization�� 266

3. Related Links�� 276
United Nations Institute for Training and Research�� 278

United Nations Joint Staff Pension Fund�� 283

United Nations Office for the Coordination of Humanitarian Affairs�� 286

United Nations Office for Counter Terrorism�� 291

United Nations Office for Disarmament Affairs�� 294

3. Related Links �� 297
United Nations Office on Drugs and Crime�� 298

xvi
United Nations Office for Outer Space Affairs�� 302

United Nations Research Institute for Social Development�� 306

2. Challenges and Opportunities�� 307
United Nations University�� 308

Universal Postal Union�� 321

World Bank Group�� 323

World Food Programme�� 335

World Health Organization�� 351

World Intellectual Property Organization�� 359

xvii
World Meteorological Organization�� 363

3. Related Links �� 366
xviii
United Nations Activities on Artificial Intelligence (AI)
Comprehensive Nuclear-Test-Ban Treaty Organization
CTBTO
1. Description of Activities on AI
The Comprehensive Nuclear-Test-Ban Treaty (CTBT) bans nuclear explosions on the Earth's
surface, in the atmosphere, underwater and underground. The Treaty has a unique and
comprehensive verification regime consisting of three pillars:
• The International Monitoring System (IMS) will, when complete, consist of 337 facilities
worldwide to monitor the planet for signs of nuclear explosions. Around 90 percent of
the facilities are already up and running.
• The International Data Centre (IDC) at the CTBTO's headquarters in Vienna acquires data
from the IMS monitoring stations. The data are processed automatically, reviewed by
human analysts and distributed to the CTBTO's Member States in both raw and analyzed
form. On-site inspections (OSI) can be dispatched to the area of a suspected nuclear
explosion if data from the IMS indicate that a nuclear test has taken place there. Inspectors
collect evidence at the suspected site.
Artificial Intelligence (AI) is applied in all three pillars of the verification regime as outlined
below.
Project 1: To detect fall army worm damage using a mobile application
Classifying signals from seismic stations to determine their seismic phase based on features
measured automatically. The features include amplitude, frequency content, particle motion
parameters, etc. Manual data processing of signals from seismic stations is cumbersome thus
the need to automate data processing at ICTBTO's International data center.
• Project Type (Status): Software project (Proof of concept)

• Project Domain: Global nuclear explosion monitoring
• AI Approach: Artificial Neural Networks (ANN) and Bayesian Classifiers
• Datasets: Automatic signals from the International Monitoring System (IMS), reviewed and
corrected by human analysts.
• Related SDGs: SDG 16 – Peace, Justice and Strong Institutions
• Resources/Skills: Human experts to review and correct the signals from seismic stations
of IMS
• Technology: Deep Learning
• Challenges: Improvement of the current system by retraining the existing ANNs on
a per station basis and replacing the ensemble of ANN and Bayes Classifiers with a
deeper ANN. Methods are being explored for seismic phase identification directly from
the waveform signal. Further studies are being undertaken to determine if the use of
additional information, such as the raw waveform data, during classification can further
improve performance
1
• Reported as part of 2022 Compendium on UN AI Activities? Yes
Project 2: Network Processing of detected signals to determine the events that have
triggered them
Detection of events by on-site inspections for every signal detected is time consuming and
expensive hence the need for network processing of signals detected at seismic, infrasound
and hydro-acoustic stations in determining the events that have caused these signals to be
observed.
• Project Type (Status): Software project (Deployment)

• AI Approach: Rule-based
• Datasets: Signals detected at seismic, infrasound and hydro-acoustic stations of IMS
• Challenges: Further research is being undertaken on the classification of radionuclide
spectra by ANNs
Project 3: NET-VISA (NETwork processing Vertically Integrated Seismic Analysis)
Improvement of the current rule-based system.
• Project Type (Status): Software project (Deployed)

• AI Approach: Machine Learning + physics model. The theoretical underpinnings are
based on the “Open Universe Probability Model”
• Project Partners: University of California (developing NET-VISA software)
• Resource: Bayesian approaches. Knowledge of seismic, infrasound, and hydro data
• Challenges: Extending to stations without detailed history from which to derive priors.
Project 4: Automatic triage
Distribute certain trouble tickets based on their content.

Project 5: Predicting failure at IMS stations
Predicting failure at IMS stations based on extensive State Of Health (SOH) parameters that
are continuously collected and store.

2
• AI Approach: Statistical methods and rule-based system; Next approach: ANNs and
Support Vector Machines (SVM).
• Datasets: IMS data and noble gas monitoring system SOH data.
CTBTO
• Project Partners: Pacific National Northwest Laboratory (PNNL)
Project 6: Seismic aftershock monitoring
Monitoring changes in the geological structures caused by a possible nuclear explosion and
classifying “weak” detections produced to enable separation of noise from signals of interest
(aftershocks).
• Project Type (Status): Software project (Testing)

• AI Approach: AI-based technique and Self Organizing Map (SOM)
• Datasets: IMS raw waveform data
• Project Partners: University of Stuttgart (developed AI-based technique)
Project 7: Satellite monitoring for On Site Inspection (OSI)
The use of air-and-spaceborne multispectral imagery (MSIR) for classification and change
detection in the inspection area, with the ultimate goal of limiting the search area and detecting
features of interest.
• Project Type (Status): Software project (Ideation)

• AI Approach: Pixel-based classification (unsupervised and supervised Machine learning),
object-based classification (decision rules and fuzzy-logic) and Change detection
techniques using Multivariate Alteration Detection (MAD)
• Datasets: Air and Space-borne multispectral imagery (MSIR)
• Resource: GIS (Geographic Information Systems) operations
• Challenges: Timeframe during an ‘On Sight Inspection (OSI)’ (e.g. availability of imagery)
2. Related Sustainable Development Goals

SDG 16 – Peace, Justice and Strong Institutions
3. Relevant links
www.ctbto.org
Contact information
Megan Slinkard, Chief, Software Applications, International Data Center Division (Megan.
Slinkard@ctbto.org)
3
The Food and Agriculture Organization
Project 1: AI-based interview on true cost accounting (TCA) case studies for SOFA
2024
• Project Description: AI-based interviewer application to implement qualitative surveys
with a wide range of stakeholders. The AI-based survey aims at collecting relevant
experiences and case studies that assess the hidden costs and benefits of various agrifood
system components to steer them towards sustainability. The objective of this initiative
was to showcase the feasibility of using AI to implement qualitative surveys as well as
create content for The State of Food and Agriculture (SOFA) 2024. With this survey the
SOFA team is inviting stakeholders to share case studies of assessments of hidden costs
and benefits of agrifood systems and how such assessments have been used to inform
decision-makers and other stakeholders in implementing transformative actions towards
sustainable agrifood systems.
• Project Type/Output: Software tool
• Project Status: Completed
• Project Start Year: 2023
• Project End Year: 2023
• Project Domain: Agrifood systems
• Data Source: Survey questions drafted by the team. The survey collects data on true cost
of food and related case studies submitted by participants.
• Publicly Available Data: Yes
• Technology/Platform: Azure OpenAI Service, Google Cloud Platform.
• Reported as part of 2022 Compendium on UN AI Activities? No
• Related Sustainable Development Goals (SDGs): SDG 1 (No Poverty), SDG 2 (Zero
Hunger) and SDG 3 (Good Health and Well-being)
• Links and Multimedia:
o Sofa 2024 AI Survey (fao.org)

o https://www.fao.org/fsnforum/call-submissions/hidden-costs-and-benefits-agrifood
-systems
• Lessons Learned: Close collaboration between the technical team and the domain experts
is essential. Testing the survey app using plausible and unplausible input is important
to improve survey quality and data privacy/security. Questions posed by the app need
to be more interactive to better approximate human interaction, improve interviewee
experience and preference of AI app over human interviewer. Many people logged in out
of curiosity and logged out. This may be better prevented with an interactive / dynamic
starting page to engage the participant.
4
• Contact information: Martha García Acevedo (Martha.GarciaAcevedo@fao.org); Andrea

Cattaneo (Andrea.Cattaneo@fao.org); Aslihan Arslan (Aslihan.Arslan@fao.org); Miguel
Benitez Humanes (Miguel.BenitezHumanes@fao.org).
FAO
Project 2: GEO-AI Challenge 2023
• Project Description: Cropland extent maps are the basic products for the practical
agricultural applications. This project aims to hold a challenge focusing on identifying
cropland extent with satellite images by collaboration with ITU and UNODC. Both satellite
imagery and ground truth data provided, and participants have tried to develop AI
solutions to generate cropland extent in the pilot regions with these materials. 71 teams
have submitted solutions, and five teams were prized based on the overall performance
of classification accuracy, technical report and presentation.
• Project Type/Output: Dataset and open-source script
• Project Domain: Agriculture
• Data Source: Satellite data, ground truth data
• Technology/Platform: Google Cloud Platform; Python
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty, SDG 2 – Zero
Hunger
• Contact information: Pengyu Hao (pengyu.hao@fao.org); Zhongxin Chen (zhongxin.
chen@fao.org).
Project 3: ATIO AI
• Project Description: Design the ATIO knowledge base and AI and web scraping
technologies for the automated collection of entries for the ATIO Knowledge base. Advise
on the strategic discussion on the roadmap, position and value of the ATIO Knowledge
base for academia, civil society, policymakers and governments.
• Project Type/Output: Software, Dataset
• Project Status: Ongoing
• Data Source: WIPO, Digital Agri Hub, Digital Impact Exchange and Green Technology
Selector
• Technology/Platform: Still to be selected.
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger
• Links and Multimedia: https://www.fao.org/documents/card/en/c/cc2506en
• Contact information: Fabrizio Brescani (Fabrizio.Brescani@fao.org), Erik VanIngen (Erik.
VanIngen@fao.org), Sonia Dias (Sonia.Dias@fao.org).
5
Project 4: Large Large Language Models (LLMs) for FAO’s Knowledge Enhancement
and Practical Application
• Project Description: Exploration and Implementation of Large Language Models (LLMs)
for FAO’s Knowledge Enhancement and Practical Application. With a primary focus on
leveraging LLMs, the initiative aims to elevate FAO’s understanding and utilization of
advanced language technologies.
• Project Type/Output: Software
• Data Source: FAO publications
• Technology/Platform: Still to be selected.
• Contact information: Fabrizio Brescani (Fabrizio.Brescani@fao.org), Erik VanIngen (Erik.
VanIngen@fao.org), Martha García Acevedo (Martha.GarciaAcevedo@fao.org).
Project 5: Crop phenology and crop calendar with remote sensing and GEO-AI
• Project Description: Crops phenology and crop calendars are essential to many
agricultural applications. This project uses time-series satellite remote sensing data and
auxiliary data to generate crop phenology data and crop calendar, employing machine
learning and GEO-AI. There are 2 phases of the project: First, algorithm development is
committed in several pilot regions, and then a global dataset will be produced.
• Project Type/Output: Dataset
• Project Status: Ongoing (Phase 1 - Development and Validation)
• Data Source: Satellite data, agricultural statistical data, land cover, land use data and in-
situ data.
• Technology/Platform: Google Cloud Platform; Python
Hunger, SDG 13 – Climate Action
• Contact information: Pengyu Hao (pengyu.hao@fao.org); Zhongxin Chen (zhongxin.
chen@fao.org).
Project 6: Creating Key Messages for the DSP mobile application

• Project Description: Digital Services Portfolio (DSP) is the FAO mobile application for
digital advisory services. It has been implemented in over 11 countries. One of the
biggest challenges during any implementation is creating content digestible by extension
services and farmers. While there is a lot of technical information, manually creating key
messages is very cumbersome and time consuming. The project is building a prototype
using Open AI chatbot to create key messages from FAO/Country based technical
documentation.
6
• Project Type/Output: Software tool (Proof of Concept)

FAO
• Project Domain: Agriculture Digital Advisory
• Data Source: Technical documentation.
• Technology/Platform: Open AI
• Lessons Learned: The improvement of the current system by automating summary of the
technical documentation and converting it to the appropriate language levels has been
a challenge. It is essential to include human experts to review and correct key messages.
• Contact information: Dioguen Zaridze (Dioguen.Zaridze@fao.org), Misael Sabbadini
(Misael.Sabbadini@fao.org).
Project 7: Animal Feed Balance calculation

• Project Description: Animal Feed Calculator is an existing FAO application that helps
livestock keepers mixing cow feed based on agricultural byproducts. Milk production
levels and quality heavily depend on the nutritional intake of the animals. The application
contains Feed database, products available in a country with their nutritional values and
nutritional requirements of the milking cows by weight category and levels of production.
The project aims to use the power of Large Language Models (LLM) and advanced ML
techniques to assist farmers to create best mix proportions based on available ingredients.
• Project Type/Output: Software tool (Proof of Concept)
• Project Domain: Agriculture Digital Advisory
• Data Source: Technical country-based data.
• Technology/Platform: Open AI, Python.
• Lessons Learned: Teaching the Open AI algorithms for the calculation of the nutritional
value of a mix is challenging.
• Contact information: Dioguen Zaridze (Dioguen.Zaridze@fao.org), Misael Sabbadini
(Misael.Sabbadini@fao.org).
Project 8: Strengthening global access to agricultural information and knowledge

(Hand-In-Hand Geospatial Platform)
• Project Description: Earth observation and geospatial IT play a critical role in the
agricultural and related sectors. FAO has created the Hand-in-Hand Geospatial Platform
that hosts data sourced from FAO, FAO partners in the public and private sectors including
from across the UN, NGOs, government institutions, academia and space agencies.
The platform has significantly increased the interoperability of FAO geospatial data as
well as the cost-effective maintenance and sustainability of different FAO geospatial
applications. Machine learning and AI are used in cutting edge quantitative remote
sensing in agriculture; world class cloud computing capabilities; enabling unprecedented
7
cross-sectoral knowledge discovery by integrating data on Soil, Land, Water, Climate,

Fisheries, Livestock, Crops, Forestry, Trade, Social and Economics and much more.
• Project Type/Output: Dataset, Software tool
• Data Source: FAO projects data and relevant data from external data providers, covering
all sub-disciplines of agriculture from animal health to trade/markets.
• Technology/Platform: Google Cloud Platform; TerriaJS; GeoNetwork; CKAN
• Project Updates: more datasets added or updated, and new functions such as support for
5 languages, user profiles, single sign-on, now support predefined shapes, user styling,
and GeoJSON annotations, and raster export.
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 2 – Zero
Hunger; SDG 13 – Climate Action.
o https://www.fao.org/agroinformatics/en
o https://youtu.be/KLqgcpQUuHo
o https://data.apps.fao.org/
• Lessons Learned: Data federation supported by standardization makes a huge difference

in data sharing. Open data sharing enables various users to contribute and get good
results to solve identified problems.
• Contact information: Karl Morteo (karl.morteo@fao.org); Zhongxin Chen (zhongxin.chen@
fao.org).
Project 9: FAO Data Lab

• Project Description: The Data Lab employs cutting-edge methods, including web scraping
and geospatial analysis, to source timely data from non-traditional sources. By using AI
algorithms, it recognizes patterns in unstructured data, generating insights from news
sources, summarizing policy documents, and enhancing FAO databases. Additionally,
it pioneers geospatial methods to fill data gaps, collaborating across FAO divisions.
By integrating AI, the Data Lab boosts processing efficiency, aiding evidence-based
policymaking for agri-food systems transformations.
• Project Type/Output: Software, Dataset
• Project Status: Ongoing Programme of work
• Project Domain: Agriculture, Trade, Food and Agricultural Statistics
• Data Source: Internet web scraping, literature mining, sentiment analysis, social media,
Earth Observation satellite imagery, Government policies, unstructured documents
• Technology/Platform: R, Jupyter, Python, running on the Google Cloud Platform, OpenAI
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG 12 –
Responsible Consumption and Production
• Links and Multimedia: https://www.fao.org/datalab/en
8
• Contact information: Carola Fabi (Carola.Fabi@fao.org); Christian Mongeau (Christian.

Mongeau@fao.org).
FAO
Project 10: Detecting Fall armyworm through user submitted photos (FAMEWS)
• Project Description: Combines an online monitoring platform for mapping data collected
by the FAMEWS mobile app whenever fields are scouted, or pheromone traps are
checked for FAW. The platform provides a real-time situation overview with maps and
analytics of FAW infestations at global, country and sub-country levels. The FAMEWS
mobile app enables data collection of scouting data, which can be collected manually
or through an image recognition model which provides immediate advice on FAW
infestation. The global monitoring platform and the mobile app are designed to expand
with the evolving needs of farmers, analysts and decision-makers. Both are accessible for
free and are helping to reduce crop yield losses and minimize risk of further introduction
and spread of FAW.
• Department/Division: Plant Production and Protection Division (NSP)
• Project Type/Output: Dataset
• Data Sources: Field scouting geo-referenced data, Pheromone traps data, Picture of FAW
damage from the field.
• Link to Data:
o https://a pp. powerbi. com/v iew? r= e yJrI joiMGQ3MzE wMTctMTlhO C00YzM3LT
kxNTgtOWIxY jIwN2YyNjF lIiwidCI6I jE2M2FjNDY 4LWFiYjgtN DRkMC04MW
ZkLWQ5ZGIxNWUzYWY5NiIsImMiOjh9
o https://data.apps.fao.org/

• Technology/Platform: Google AI, TensorFlow
• Project Updates: A new version of the FAW global platform was published. The project
was extended up to the end of October 2024
• Partnership(s)/Collaborator(s):
o Academia: Penn state University, USA (PlantVillage platform)
o www.fao.org/fall-armyworm/monitoring-tools/en/
o http://www.fao.org/3/CA1089EN/ca1089en.pdf
o FAO-FAMEWS V3 - Apps on Google Play
• Lessons Learned:
o The most important challenge was promoting the adoption of the application and
convincing FAO members to share their data.
o The second challenge was the sustainability of the project as it is difficult to maintain
and promote the system without any financial support.
o The accuracy of the collected data was also another challenge as it is crowd sourced
data.
9
• Contact information: Maged Elkahky (maged.elkahky@fao.org).
Project 11: WaPOR (Water Productivity through Open access of Remotely sensed
derived data)
• Project Description: WaPOR is FAO’s portal to monitor Water Productivity through
Open-access of Remotely sensed derived data. It assists countries in monitoring water
productivity, identifying water productivity gaps, proposing solutions to reduce these gaps
and contributing to a sustainable increase of agricultural production. At the same time, it
considers ecosystems and the equitable use of water resources, which should eventually
lead to an overall reduction of water stress. The WaPOR portal provides open access to
key land and water variables (including reference and actual evapotranspiration, biomass,
land cover, precipitation) in near – real time at global level with a spatial resolution of 300
m. and, for Africa and the Near East and project countries at 100 m resolutions with pilot
areas at 20 m.
• Project Type/Output: Dataset, Software tool
• Data Source: Geospatial Database with remote sensing data input. The database is
publicly accessible, developed with open access data and open-source algorithms. It
provides near real time information from 2009 to date.
• Technology/Platform: Google Cloud services, Python, Jupyter
• Project updates: the WaPOR data is now global.
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG 6 – Clean
Water and Sanitation; SDG 13 – Climate Action
o Government: Ministry of Foreign Affairs of the Netherlands.

o Private Sector: eLEAF and private sector entities in selected project countries.
o Civil Society: not for profit organizations in several project countries.
o Academia: IHE Delft.
o Research: IWMI (CGIAR centre).
o https://www.fao.org/in-action/remote-sensing-for-water-productivity/en/
o https://wapor.apps.fao.org/and https://data.apps.fao.org/wapor/
o https://www.youtube.com/watch?v=ZX7SOhk97hA
o https://www.youtube.com/watch?v=gA_t4HuFNhM
• Contact information: Livia Peiser (Livia.Peiser@fao.org), Jippe Hoogeveen (Jippe.

Hoogeveen@fao.org) or wapor@fao.org.

SDG 1, 2, 3, 6, 12, 13, 14
10
3. Relevant Links
https://www.fao.org/home/en
FAO
Contact Information
CSI-Director (CSI-Director@fao.org)
11
International Atomic Energy Agency
Project 1: AI for Atoms – the IAEA's knowledge-sharing platform for partnership on

AI applications in the nuclear field
• Project Description: The IAEA provides interdisciplinary fora for professionals to discuss
and foster collaboration on the use of AI in nuclear applications, science and technology
and is committed to sharing knowledge and forging partnerships through its AI for
Atoms platform. As part of this initiative, the IAEA cooperates with the International
Telecommunication Union, the UN Interagency Working Group on AI and almost 40 other
UN organizations to provide a solid foundation for accelerated sustainable development
with AI. The AI for Atoms platform provides information on the IAEA’s activities on AI,
featuring all related initiatives, news, publications, and events.
• Project Domain: Agriculture, Environment, Energy, Health, and Nuclear Science,
Technology and Applications
• Related Sustainable Development Goals (SDGs): SDG 2 – No Hunger; SDG 3 – Good
Health and Well-Being; SDG 6 – Clean Water and Sanitation; SDG 7 – Affordable and
Clean Energy; SDG 9 – Industry, Innovation, and Infrastructure; SDG 13 – Climate
Action; SDG 15 – Life on Land; SDG 17 – Partnerships for the Goals
o AI for Atoms: https://nucleus-new.iaea.org/sites/ai4atoms/
• Contact information: ai4atoms@iaea.org
Project 2: Crowdsourcing AI Solutions for Climate Change

• Project Description: A new contest for start-up companies was launched by the IAEA, the
ITU, the FAO and UNESCO focusing on jumpstarting AI-powered solutions for food and
climate-smart agriculture and water resource management. The five problems in need of
solutions included:
o How can AI help estimate soil properties across large areas? (based on data collected
from infrared spectroscopy, gamma spectrometry techniques and satellite imagery)
o How can AI help map and monitor soil moisture levels? (using data from cosmic ray
neutron sensors and gamma spectrometry techniques, in combination with satellite
imagery)
o How can AI complement the Earth Map to help users identify water-related challenges
in specific areas and provide useful insights on and analyses of water resources?
o How can AI help identify water resources vulnerable to climate change and support
flood and drought forecasting and early warning?
12
o How can AI be integrated into isotope hydrology studies to support water management
and mitigate the world’s water problems?
The project culminated in a COP28 high-level event showcasing the winning projects and
IAEA
solutions. The IAEA is working with the winning start-ups to identify synergies with its data
collected using various nuclear technologies.
• Project Domain: Agriculture, Environment, Water and Nuclear Science, Technology and
Applications
• Related Sustainable Development Goals (SDGs): SDG 2 – No Hunger; SDG 6 – Clean
Water and Sanitation; SDG 9 – Industry, Innovation, and Infrastructure; SDG 13 – Climate
Action; SDG 17 – Partnerships for the Goals
o AI Solutions for Climate Change – Innovation Factory, https://aiforgood.itu.int/about

-ai-for-good/aiml-solutions-for-climate-change/
• Contact information: Noura El-Haj (N.El-Haj@iaea.org)
Project 3: Publication on Artificial Intelligence in Medical Physics

• Project Description: The deployment of technologies based on AI in the medical use of
radiation (i.e., radiation oncology, diagnostic imaging, nuclear medicine) and medical
physics is expected to grow over the next 5–10 years. Clinically Qualified Medical Physicists
(CQMP) are expected to play an important role in ensuring safe and effective clinical
implementation of AI based tools. This publication launched in November 2023 frames
the roles and responsibilities of CQMPs in the field of implementation and utilization
of AI in the medical uses of radiation. The publication provides definitions, principles,
examples, and associated risks of AI-based tools in medical physics. Furthermore, it
provides principles and needs on academic education and continuing professional
development for CQMPs.
• Project Domain: Health, Medical Physics
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Well-Being
o INTERNATIONAL ATOMIC ENERGY AGENCY, Artificial Intelligence in Medical

Physics, Training Course Series No. 83, IAEA, Vienna (2023), https://www.iaea.org/
publications/15450/artificial-intelligence-in-medical-physics
• Contact information: Mauro Carrara (M.Carrara@iaea.org), Egor Titovich (E.Titovich@iaea

.org)
Project 4: AI-assisted Contouring Skills in Radiotherapy

• Project Description: The IAEA is conducting a Coordinated Research Project on The
Potential of E-learning Interventions for AI-assisted Contouring Skills in Radiotherapy
(CRP E33046). The objective of the project is to investigate changes in inter-observer
variation and bias after e-learning in delineation guidelines and the use of deep learning-
based auto-segmentation of organs-at-risk in head-and-neck cancer. In recent years, AI-
algorithms, namely deep learning-based algorithms, have improved auto-segmentation
drastically. It is believed that AI-tools lower variation and increase the accuracy and
compliance of plans, which improves the chance of cure. AI-tools may also make care
more cost effective by reducing the human capacity required, which is important in our
overstretched health systems. A wide palette of commercial deep learning-based auto-
segmentation solutions are emerging with the promise of leveraging the aforementioned
13
benefits. The selection and contouring of target volumes and organs-at-risk has become a
key step in modern radiation oncology. Concepts and terms for definition of gross tumour
volume, clinical target volume and organs at risk have been continuously evolving (e.g.,
through International Commission on Radiation Units and Measurements (ICRU) reports
50, 62, 78, 83) and have become widely disseminated and accepted by the European
and international radiation oncology community. Currently, 102 participants from 23 IAEA
Member States from 24 radiotherapy centers are participating in the project.
• Project Domain: Health
• Contact information: Egor Titovich (E.Titovich@iaea.org)
Project 5: AI to Assess Climate Impact on Global Lakes

• Project Description: Global warming is considered a major threat to Earth’s lakes water
budgets and quality. However, flow regulation, over-exploitation, lack of hydrological
data, and disparate evaluation methods hamper comparative global estimates of lake
vulnerability to evaporation. The stable isotope composition of 1,257 global lakes
was analyzed using Artificial Intelligence techniques. It was found that in most of the
lakes, this depends on precipitation and groundwater recharge subsequently altered
by catchment and lake evaporation processes. Isotope mass-balance modelling shows
that ca. 20 % of water inflow in global lakes is lost through evaporation and ca. 10 %
of lakes in arid and temperate zones experience extreme evaporative losses >40 %
of the total inflow. Precipitation amount, limpidity, wind speed, relative humidity, and
solar radiation are predominant controls on lake isotope composition and evaporation,
regardless of the climatic zone. The promotion of systematic global isotopic monitoring
of Earth’s lakes provides a direct and comparative approach to detect the impacts of
climatic and catchment-scale changes on water-balance and evaporation trends. This
project aims to establish a global network of isotopes in lakes (GNIL) to improve our AI-
based assessments. Current activities include the design of the database, the selection of
variables and the assessment of global coverage. Stable water isotope assays provide a
low-cost effective tool to study lake-catchment changes with regards to sample collection
and analysis. Additionally, stable isotope data from lakes is fully comparative globally,
thereby providing a competitive advantage under the current scenario of different
international methods and approaches that are not easy to compare in time and scale
and which result in the current lack of the comparable data for lakes and catchments.
• Project Domain: Environment, Water Resource Management
• Data Source: Satellite open-source data on climate, hydrology and lakes parameters and
stable water isotopes data in global lakes.
• Link to data: The data can be obtained by request at https://nucleus-new.iaea.org/sites/
ihn/Pages/GNIR.aspx.
• Technology/Platform: R programming, Isotope Hydrology Collaboration Platform
• Related Sustainable Development Goals (SDGs): SDG 6 – Clean Water and Sanitation,
SDG 13 - Climate Action
• Contact information: Yuliya Vystavna (Y.Vystavna@iaea.org), Astrid Harjung (A.Harjung@
iaea.org)
Project 6: AI for Water and Environment

• Project Description: This project aims to promote and enable the use of isotopic
techniques with AI tools for better management of water and environmental resources,
as well as adaptation to climate change worldwide. Recognizing that with the increasing
availability of data from satellites, unmanned airborne vehicles and sensor networks, there
14
is a myriad of data available to couple and explore in conjunction with the IAEA’s global
isotope databases.
The project is a platform for scientists working with AI tools at the interface of isotope
hydrology, water resources protection and management. This facilitates sharing of
IAEA
experiences in the use of machine and deep learning for hydrological and environmental
modelling, challenges, and research opportunities to move forward. The project aims to
find synergies between isotope techniques, high-frequency or remote sensing, open-
source resources, and AI to show how these can help inform policies to mitigate the
world's water problems. Besides exchanging knowledge on current developments in
the area that feed into research activities of the isotope hydrology section, the project
foresees submission of a review paper to a hydrological journal that showcases different
applications of AI in isotope hydrology. This paper intends to explore what role machine
learning and big data can play in the advancement of isotope hydrology.
• Project Domain: Environment
• Technology/Platform: R programming
• Related Sustainable Development Goals (SDGs) SDG 6 – Clean Water and Sanitation,
SDG 13 - Climate Action
• Contact information: Astrid Harjung (A.Harjung@iaea.org), Yuliya Vystavna (Y.Vystavna@
iaea.org)
Project 7: Zoonotic Disease Integrated Action (ZODIAC) Respiratory Disease

Phenotype Observatory
• Project Description: Medical imaging plays a crucial role in diagnosing and monitoring
infectious diseases, and the COVID-19 pandemic has highlighted the critical role of
chest imaging, including computed tomography (CT) scans and baseline X ray analysis,
in the early stages of the disease. Medical imaging technologies, such as computed
tomography (CT), can detect lesions smaller than 0.5 mm and assess activity without
the need for a biopsy. This enables the identification of specific disease characteristics
in each patient. However, the analysis of the complex images generated by these
technologies can be challenging to the naked eye. Radiomics, a method for extracting
large-scale imaging data from medical imaging studies like CT and X ray scans, utilizes
data-characterization algorithms to identify disease findings that are not visible to the
naked eye. This approach can be complemented by AI, ML, and deep learning to
identify patterns of lung involvement in COVID-19 patients. The IAEA has initiated a
project with the aim of identifying specific characteristics associated with different virus
variants and determining if there are any epidemiological and clinical differences in the
development of disease complications and specific medical imaging manifestations in
several respiratory infectious diseases.
By 2026 The ZODIAC Respiratory Disease Phenotype Observatory will be collecting a
continuous stream of imaging- and associated clinical data of patients with respiratory
disease from around the world.
• Project Domain: Human Health, Zoonotic Diseases, Pandemic Prevention
o ZODIAC webpage: https://nucleus.iaea.org/sites/zodiac/SitePages/Home.aspx
• Contact information: Enrique Estrada Lobato (E.Estrada-Lobato@iaea.org)
15
Project 8: AI to Support Remediation of Radioactive Contamination in Agriculture

• Project Description: Remediation of radioactive contamination of farmland requires
accurate soil data and exchangeable potassium (Kex.) plays a major role. As potassium
competes with radiocaesium in soil-to-plant transfer, it can help reduce the crop uptake of
this major fallout radionuclide. Information on potassium content in the soil is essential for
optimizing remediation of radioactive contamination. Through the international research
network under the IAEA funded Coordinated Research Project D1.50.19 on ‘“Remediation
of Radioactive Contaminated Agricultural Land”, the Joint FAO/IAEA Centre of Nuclear
Techniques in Food and Agriculture investigated how to predict exchangeable potassium
in soil through Mid-Infrared Spectroscopy and Artificial Intelligence. In our study, which
was finalized this year and is expected to be published soon, we show how this can be
now achieved using a Convolutional Neural Network (CNN) model trained on a large
Mid-Infrared (MIR) soil spectral library (40000 samples Kex determined with 1M NH4OAc,
pH 7), compiled by the National Soil Survey Center of the United States Department
of Agriculture. Using Partial Least Squares Regression as a baseline, we found that our
implemented CNN leads to a significantly higher prediction performance of Kex. when
a large amount of data is available (10000), increasing the coefficient of determination
from 0.64 to 0.79, and reducing the Mean Absolute Percentage Error from 135% to
31%. Furthermore, in order to provide end-users with required interpretive keys, we
implemented the GradientShap algorithm to identify the spectral regions considered
important by the model for predicting Kex. Used in the context of the implemented CNN
on various Soil Taxonomy Orders, it allowed (i) to relate the important spectral features
to domain knowledge and (ii) to demonstrate that including all Soil Taxonomy Orders
in CNN-based modeling is beneficial as spectral features learned can be reused across
different, sometimes underrepresented soil orders.
In early 2023, KU Leuven and Curtin universities in collaboration with the Joint FAO/IAEA
Centre, launched a PhD project to explore the integration of AI, soil spectroscopy, and
remote sensing in Kex characterization at both local and global scales. This research, which
seeks to bridge soil science and deep learning, focuses on three key questions:
o The feasibility of using Mid-Infrared spectroscopy (MIRS) and deep learning with large
soil spectral libraries to accurately predict Kex for local-scale remediation.
o Assessing Near-Infrared spectroscopy’s (NIRS) potential for high-throughput Kex
prediction, considering its lower cost and portability compared to MIRS.
o Exploring global-scale Kex prediction using remote sensing data and geographical
soil-forming factors in advanced modeling approaches.

• Project Domain: Food and Agriculture
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG 3 – Good
Health and Well-Being
• Contact information: Simon Kelly (S.Kelly@iaea.org), Gerd Dercon (G.Dercon@iaea.org)
Project 9: AI to Assess and Use Marine Environmental Radioactivity Data

• Project Description: Levels of radionuclides in the marine environment can be influenced
by a range of manmade inputs (e.g., nuclear accidents and authorised discharges
from nuclear fuel cycle facilities), underlying biogeochemical conditions, and other
dynamic such as climate change and sea level rise. On the other hand, the presence of
radionuclides – both natural and artificial – in the marine environment offers a wealth of
possibilities for the application of radiotracers for quantification of climate and ocean
change. The IAEA Marine Radioactivity Information System (MARIS) has been maintained
for many years to provide data underpinning related monitoring and research activities.
MARIS is currently being redeveloped in order to offer improved support and possibilities
to its user community, including for AI and ML applications. The planned improvements
16
include adoption of Open Science principles, data management according to FAIR and/or
GEO principles, improved API access and adoption of netCDF and related technologies,
with all associated improvements in metadata provision and opening of access to a rich
range of tools for accessing, analysing and visualizing data. The overall aim is to promote
IAEA
data re-use and to facilitate Open Science. Specifically for AI a range of possibilities are
foreseen to incorporate marine modelling for prediction of inter alia transport, dispersion
and settling rates for future and sparsely observed scenarios; visualisation of radionuclide
levels in four dimensions; time series analysis; improved radiotracer applications and
reanalyses using combined datasets.
• Project Domain: Marine Environment, Climate and Ocean Change, Emergency
Preparedness and Response, Radiological Environmental Impact Assessment
• Data Source: IAEA Marine Radioactivity Information System https://maris.iaea.org/
• Related Sustainable Development Goals (SDGs): SDG 14 – Life Below Water; SDG 13–
Climate Action
• Contact information: Paul McGinnity (P.Mc-ginnity@iaea.org)
Project 10: AI for Nuclear Data Development and Evaluation

• Project Description: Explore, test, and deploy AI methods to improve and speed-up
nuclear data development and evaluation. The IAEA coordinates and contributes to
nuclear data development projects, e.g., within the International Nuclear Data Evaluation
Network (INDEN) and the Neutron Data Standards project, and provides databases with
atomic, molecular and nuclear data to support developments in various nuclear-related
domains, such as fusion energy research, nuclear medicine, nuclear astrophysics and
fundamental nuclear science.
Concerning nuclear data development, progress on AI-related activities is shared
among collaborators within the various projects and networks. Additionally, open-source
evaluation packages in R and Python are developed at IAEA that leverage machine
learning frameworks, such as TensorFlow, for advancing evaluation methodology to
enhance the quality of the nuclear data provided. These products are shared on the IAEA
GitHub account for making developments more inclusive, traceable and reproducible,
and also to lower the entrance barrier to new contributors. Methodological progress is
documented in reports and journal publications.
Regarding the dissemination of databases, new APIs, formats, webservices and
software tools are developed at IAEA to facilitate the access to nuclear data and to
enhance the quality assurance procedures. Especially, programmatic access to nuclear
data is anticipated to facilitate the application of machine learning for automatic data
classification, outlier detection and estimation of nuclear quantities. Furthermore, the IAEA
is involved in international projects and expert groups, such as OECD-NEA WPEC SG50,
concerned with the development of experimental database suited for ML applications.
• Project Domain: Physics, Nuclear Data
• Technology/Platform: R, Python, TensorFlow / GitHub
• Link and Multimedia:
o IAEA's GitHub repository: https://github.com/IAEA-NDS
• Related Sustainable Development Goals (SDGs): SDG 7 – Affordable and Clean

Energy; SDG 9 – Industry, Innovation, and Infrastructure
• Contact information: nds.contact-point@iaea.org
17
Project 11: AI for Fusion

• Project Description: The IAEA has launched a 5-years Coordinated Research Project
(CRP) on AI for Accelerating Fusion R&D (2022–2027) with the objective of accelerating
fusion R&D with machine learning and AI, through the creation of a platform and cross-
community network for innovation and partnership. The project features four work
packages. These are:
o Real-time Magnetic Fusion Energy (MFE) System Behaviour Prediction, Identification

& Optimization Using ML/AI Methods: To accelerate fusion R&D by establishing a
multi-machine database of experimental and simulation MFE data (adhering to FAIR/
Open Science principles) for ML/AI-driven applications, and through increased access
to knowledge and information of ML/AI methods for MFE.
o Inertial Fusion Energy (IFE) Physics Understanding through Simulation, Theory
and Experiment Using ML/AI Methods: To accelerate fusion R&D by establishing a
database of experimental and simulation IFE data (adhering to FAIR/Open Science
principles) for ML/AI-driven applications, and through increased access to knowledge
and information of ML/AI methods for IFE.
o Feasibility of Magnetic Fusion and Inertial Fusion Image Database: To determine
the feasibility of an image database from MFE and IFE data (adhering to FAIR/Open
Science principles) for ML/AI-driven applications with potential to accelerate fusion
R&D.
o Community Engagement & Workforce Development: To accelerate
community engagement and capacity building, as well as create and provide with
access to knowledge and information in the area of ML/AI methods applied to fusion
R&D.
As part of this activity, in 2023 the IAEA and ITU conducted a Crowdsourcing Challenge
on AI for Fusion which provided a platform for collaboratively explore the potential of
AI in enabling predictive modelling for fusion energy systems. Through this Challenge,
participants used data from three distinct fusion experimental machines to develop
a cross-machine disruption prediction model using ML, with strong generalization
capabilities. Finally, the IAEA designated the Massachusetts Institute of Technology
Plasma Science and Fusion Center as its Collaborating Centre on research and education
in the area of AI applications in fusion and plasma science for the period 2023–2027.
• Project Domain: Fusion Energy
• Related Sustainable Development Goals (SDGs): SDG 9 – Industry, Innovation and
Infrastructure
o AI for Fusion: https://nucleus.iaea.org/sites/ai4atoms/ai4fusion

o AI for Fusion Energy Challenge: https://aiforgood.itu.int/about-ai-for-good/ai-for
-fusion-energy-challenge
• Contact information: Matteo Barbarino (M.Barbarino@iaea.org)
Project 12: International Network on Innovation to Support Operating Nuclear Power

Plants (ISOP)
• Project Description: The IAEA International Network on Innovation to Support Operating
Nuclear Power Plants (ISOP), established in 2019, aims to facilitate the engagement on the
deployment of near-term innovation in the nuclear power sector from diverse expertise,
both internal and external to the IAEA. This project features several work packages,
delivering unique outputs. These are:
18
o Creation and coordination of the ISOP Working Group on AI as the first pilot working
group organized under the auspices of the ISOP Network. The AI working group
was established in July 2022. The objective of the AI working group is to strengthen
awareness and engagement in activities linked to recent and near-term deployment
IAEA
of AI in the nuclear power sector The ISOP AI WG serves as an effective platform
to share these ideas and put forth the vision for possible activities in the area of
near-term deployable AI for the nuclear power sector. As of the end of 2023, the
group comprised 66 international experts, from 18 Member States representing
nuclear utilities, regulators, technical support organizations, national laboratories and
academia.
o Nuclear Energy Series Publication on Deployment of Artificial Intelligence Solutions for
the Nuclear Power Industry: Considerations and Guidance: The goal of the document
is to provide overall guidance to the nuclear industry on what aspects should be
addressed when developing and deploying AI solutions in the nuclear industry to
help ensure its performance is adequate while maintaining safe and reliable operation
based on the current knowledge, up to date best practices, experiences, benefits
and challenges in the life cycle of NPPs. The work is coordinated within the ISOP AI
working group described above.

• Project Domain: Energy
• Related Sustainable Development Goals (SDGs): SDG 7 – Affordable and Clean Energy;
SDG 9 – Industry, Innovation, and Infrastructure; SDG 13 – Climate Action
o IAEA International Network on Innovation to Support Operating Nuclear Power Plants

(ISOP): https://nucleus.iaea.org/sites/connect/ISOPpublic/SitePages/Home.aspx
• Contact information: Nelly Ngoy Kubelwa (N.Ngoy-kubelwa@iaea.org)
Project 13: AI to Accelerate the Technological Development of Evolutionary and

Innovative Reactor Designs
• Project Description: This project features the following deliverables:
o IAEA Collaborating Centre on Artificial Intelligence for Nuclear Power, Center for
Science of Information at Purdue University, established in November 2023.
o IAEA Coordinated Research Project I31033 on Advancing the State-of-Practice in
Uncertainty and Sensitivity Methodologies for Severe Accident Analysis in Water-
Cooled Reactors (2019–2024): The CRP is ending in February 2024.

o IAEA Collaborating Centre on Artificial Intelligence in Nuclear Power: https://nucleus

.iaea.org/sites/CollaboratingCentre/SitePages/Purdue%20University.aspx
o IAEA Coordinated Research Project I31033: https://www.iaea.org/projects/crp/i31033
• Contact information: Alexei Miassoedov (a.miassoedov@iaea.org)
19
Project 14: Safety Implications of the Use of AI in Nuclear Installations
Project Description: The purpose of the project is to collect current experiences in Member
States on the use and application of AI techniques to support different aspects of the design,
safety assessment and operation of nuclear installations; and to understand potential safety
benefits and challenges related to safety demonstration and justification of AI applications. The
experiences will be collected by conducting technical meetings, workshops and other events
in the subject area for nuclear power plants, research reactors and nuclear fuel cycle facilities.
These events include:
a) IAEA Technical Meeting on the Safety Implications of the Use of Artificial Intelligence in
Nuclear Power Plants
o The meeting was held at IAEA headquarters in Vienna from 16 – 20 October 2023 and
attracted more than 70 participants from 25 Member States.
b) IAEA Workshop on Safety Considerations in Use of Advanced Technologies at Nuclear

Fuel Cycle Facilities
o The workshop was hosted at National Nuclear Laboratories, UK from 2 – 6 October

2023 and attracted participants from 11 Member States.
c) IAEA Technical Meeting on Safety and Operational Considerations in the Use of Advanced
Technologies at Research Reactors
o The meeting will be organized from 30 September – 4 October 2024 at IAEA

Headquarters in Vienna.
d) Side event on “Safety considerations in use of advanced and innovative technology in

research reactors” at the International Conference on Research Reactors: Achievements,
Experience and the Way to a Sustainable Future – to be held in 2024.
The discussions and insights from the work will be captured in IAEA publications.

• Project Domain: Nuclear Safety
• Contact information: Yun Goo Kim (Y.G.Kim@iaea.org), Lakshman Valiveti (L.Valiveti@iaea
.org)
Project 15: AI Activities for Nuclear Security
Project Description: This project features the following activities:
Ongoing:
• Coordinated Research Project on Enhancing Computer Security for Radiation Detection

Systems:
o This project is focused on enhancing computer security for radiation detection systems.
The project has been exploring methods for analysing communication patterns in
radiation detection equipment. This allows for the identification and flagging of
anomalous behaviour, thereby enhancing the integrity of the devices and network
infrastructure.
20
• Enhancement of Radiation Detection Systems:
o This project is investigating enhanced operability to improve the detection of

materials outside of regulatory control. This includes the development of sophisticated
IAEA
identification algorithms capable of detecting shielded materials, as well as improving
analysis and pattern recognition using spectroscopic data, contributing to a more
robust detection capability.
• Advancement in Security Equipment:
o Progress is being made in wireless adapter development i, including the use of image
recognition ML algorithms. These advancements facilitate the wireless transmission
of detector data, streamlining operations and improving response times.
• Proactive Threat Identification:
o Efforts are underway to identify emerging threats and capabilities, particularly involving
AI/ML capabilities. This proactive approach ensures that our security measures are
constantly evolving to meet the challenges posed by the dynamic nature of threats in
nuclear security.
Forthcoming:
• Upcoming Coordinated Research Project J02021 on Computer Security for Small Modular
Reactors (SMRs):
o Research is planned into cyber-secure frameworks that can integrate with operational
modes of SMRs, employing cutting-edge technologies, including AI and ML.
o Building on the foundations of previous work, the project aims to conduct further
research into AI-based cyber threat detection and response systems. These systems
will be capable of continuous monitoring and analysis of instrumentation and control
data to bolster security measures.

• Project Domain: Nuclear Security, Computer Security, Infrastructure, Data, Radiation
Monitoring
• Related Sustainable Development Goals (SDGs): SDG 9 – Industry, Innovation, and
Infrastructure
o IAEA NUSEC for WG on AI4NS materials: https://nusec.iaea.org/portal/(also https://

www.iaea.org/resources/databases/nusec)
• Contact information: Mitchell Hewes (M.Hewes@iaea.org)
Project 16: Investigating the Use of Generative AI for Public Communication during
Emergency
• Project Description: The IAEA Coordinated Research Project, “Effective Public Emergency
Communication in a Misinformation Environment: online misinformation disruption
of nuclear/radiological emergency response”, and a 2024 IAEA Technical Meeting
on “Effective Public Communication in a Disruptive Information Environment during
Emergency Response” investigate generative AI’s capabilities in producing at a mass scale
synthetic, deceptive, persuasive, multilingual disinformation content in video, image,
audio and text formats.
To support Member States in sustaining public trust and effective emergency public
communication, this activity aims at gathering expert advice from academic, governmental
and industry researchers. The resulting advice can be used to adapt emergency
21
preparedness and response procedures to mitigate the consequences of disruptive

synthetic content.
• Project Domain: Emergency Preparedness and Response
• Technology/Platform: Open
Infrastructure and SDG 17 – Partnerships for the Goals
• Contact information: Nayana Jayarajan (N.Jayarajan@iaea.org)
Project 17: Neural Artificial Intelligence for Document Indexing and Analysis (NADIA)
• Project Description: AI is used to categorize and add index terms to each record in
the IAEA’s International Nuclear Information System (INIS), to enrich searches and add
structure to the repository.
• Project Domain: Nuclear Information Collection and Curation
• Data Source: Full-text and metadata of knowledge products provided by Member States
and IAEA Publications to INIS in all areas of nuclear science and technology. This is
supplemented with metadata of scientific articles harvested from publishers.
• Technology/Platform: Python is the main programming language. Natural Language
Processing (NLP) routines and tasks are employed along with a customized BERT-based
model, called distilBERT, which is used for categorizing records into 48 subject categories
and 11,550 of the most frequently used thematic descriptors. To incorporate indexing for
several thousands of less frequently used descriptors, as well as any descriptors that will
henceforth be added periodically, semantic matching in record abstracts is used based
on custom spaCy pipelines, which takes into account semantic relationships between
descriptor terms. The solution is designed to utilize standard CPU hardware, making it
very cost-efficient to operate. The project will furthermore employ quantization to further
reduce the model’s hardware footprint and increase prediction performance.
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education; SDG 9 –
Industry, Innovation, and Infrastructure
o INIS Repository: https://inis.iaea.org/search/
• Contact information: Brian Bales (B.Bales@iaea.org)
Project 18: AI-Assisted Processing of Safeguards-relevant Information for Nuclear

Verification
• Project Description: The following ongoing projects leverage unstructured data and latest
DL technologies to further increase effectiveness and efficiency in processing information
declared by States and information available to the IAEA from open sources, including
satellite imagery:
o Computer-Assisted Review of Open Source Information: Modern Natural Language

Processing (NLP) algorithms can assist the discovery and review of relevant open
source (OS) information, including but not limited to news articles and science &
technology publications. An additional classification model categorizes documents
based on technologies related to the nuclear fuel cycle to facilitate the consistency
analysis between information declared by a State and information validated by the
IAEA, thereby supporting the assessment of the nuclear capabilities of the State. To
encompass an even wider range of searches, AI is currently being explored to look
into transcription and translation.
22
o Computer-Assisted Review of State Declarations: The continuous review of

declarations, submitted by States under a safeguards legal instrument entitled
Additional Protocol involves a high degree of manual work due to e.g, the diversity
of formats and varying levels of detail submitted by States. NLP algorithms are being
IAEA
applied on such unstructured text data to assist IAEA staff in the identification of new
entities, new research activities or new collaborations of between States of safeguards
relevance. In so doing, the IAEA can compare historical data with other safeguards-
relevant information, with the aim of identifying new declared (or undeclared) locations
and research activities.
o Computer-Assisted Review of Satellite Imagery: To assess the correctness and
completeness of State declarations, the IAEA collects and reviews a high number
of satellite imagery e.g., to detect any safeguards relevant changes in facilities. To
reduce manual work, the IAEA is exploring the use of modern DL models to assist
in the review process of satellite imagery. The DL models can highlight important
infrastructure changes, including construction, building extensions, and demolition,
helping to prioritize the use of human resources where such infrastructure changes
have been identified and pointing attention to those that may require further (human)
analysis.

• Project Domain: Safeguards Verification
• Data Source: The text-based applications utilize internal documents, documents available
in the International Nuclear Information System (INIS) repository, as well as Internet-based
sources including feeds for news articles and databases for science and technology
publications. For satellite imagery, remote sensing data from various satellite imagery
sources are used.
• Technology/Platform: Python is used as the main programming language. NLP tasks are
based on domain-adapted BERT-like models, as well as offline available Large Language
Models, such as Llama2 and Mistral. Deep learning applications for satellite imagery use
convolutional neural networks (e.g., U-Net) and transformer-based neural networks. For
model training and inference GPU hardware resources and a ML platform are available.
Infrastructure; SDG 16 – Peace, Justice and Strong Institutions; SDG 17 – Partnerships
for the Goals
• Contact information: Paul Schneeweiss (P.Schneeweiss@iaea.org)
Project 19: Deep Learning for Safeguards Surveillance

• Project Description: As part of its safeguards activities, the IAEA carries out surveillance
using cameras installed in nuclear facilities. The review of surveillance data is time-
consuming and cognitively demanding. Implementing a DL decision support system for
inspectors performing this task can allow for large efficiency gains due to significantly
improved detection accuracy and fewer false alarms than classical algorithmic methods.
In this project, the IAEA uses DL for critical safeguards activities: verification of spent
fuel. A DL object detection algorithm is used to identify and localize safeguards relevant
activities and objects in the surveillance data with a high level of accuracy. A DL system
automatically and continuously flags events in remotely transmitted surveillance data so
that it is immediately available to IAEA inspectors when they start their surveillance review,
allowing them to quickly verify declarations made by nuclear facility operators. Using
DL for spent fuel verification is relevant to the IAEA due to steadily growing inventories
of such material. Additional algorithms are used to detect the presence of canister ID
numbers and to detect the direction of spent fuel cask movements in facilities. In addition
to helping to verify States’ declarations more efficiently, the IAEA is also exploring the use
of DL for the detection of abnormal events – i.e., events not falling into usual activities –
thereby assisting in the detection of undeclared activities.
23
The IAEA DL system for greater surveillance review efficiency is currently computing
results for over 20 cameras in four nuclear facilities in two Member States. Inspectors are
acceptance testing DL results as part of their surveillance review, and they have reported
significant time savings for reviewing surveillance data. Further, accuracy has been
statistically evaluated at over 90%, far higher than classical algorithmic methods used by
inspectors. Initial trained DL models are expected to be authorized for IAEA use in 2024
and efforts will continue to expand the DL system for other use cases and facilities.
• Project Domain: Safeguards Verification
• Data Source: Deep learning models are trained on images containing safeguards-
relevant activities using data from past inspector surveillance reviews. Current focus is
on movements of spent fuel.
• Technology/Platform: Python, Pytorch, Docker containers, Linux servers, GPUs.
• YOLOv7 algorithm for object detection and localization: https://github.com/WongKinYiu/
yolov7
Infrastructure; SDG 16 – Peace, Justice and Strong Institutions; SDG 17 – Partnerships for
the Goals
• Contact information: Maikael Thomas (M.Thomas@iaea.org)

SDGs 2, 3, 4, 6, 7, 9, 13, 15, 16, 17
3. Relevant Links
https://www.iaea.org/
https://nucleus.iaea.org/sites/ai4atoms
Contact Information: Ashraf Sayed (S.A.H.Ashraf@iaea.org), Matteo Barbarino (m.barbarino@

iaea.org)
24
International Fund for Agricultural Development
IFAD
Project 1: Athena: Leveraging Artificial Intelligence and Big Data for IFAD 2.0
• Project Description: The ATHENA project seeks to unlock the potential of artificial
intelligence and machine learning to accelerate knowledge generation and strengthen
data-driven decision-making in IFAD. The project developed an AI toolbox with three
main objectives: (1) systemize IFAD’s portfolio to facilitate results measurement and
institutional learning, (2) enhance knowledge management through deployment of AI/ML
in IFAD ICT systems to make project results and lessons learned accessible and actionable
for staff, and (3) predict performance and impact of IFAD projects to inform decision-
making, optimize targeting, and maximize impact. The (AI) “tool box” contains:
o AI-based Intervention Dashboard: a searchable dashboard that classifies IFAD’s

investment portfolio by project features, including interventions, outcomes, animal
and plant products, among others using natural language processing (NLP);
o Lessons Learned Web App: an application to search for relevant “lessons learned” as
reported in previous projects to inform new designs reports;
o Trend analyses of strategic themes: historical evidence of activities related to strategic
topics, such as SDGs, food systems and ICT4D, to understand IFAD’s support and
allocation of resources to different activities historically;
o Project performance prediction model: a framework for ex-ante prediction of project
performance based on a set of project features to facilitate better design and early
action throughout project implementation;
o Project impact prediction model: a framework to predict the probability of a positive
impact of IFAD-supported interventions using impact evaluation data;
o Project targeting optimization model: framework and tool to identify beneficiary
features to maximize project impact; and
o Covid-19 impact prediction model: model to predict impact of the pandemic in IFAD’s
beneficiary countries.
The tools developed by the project fill a gap within IFAD and the field by aiding and
simplifying IFAD reporting, especially for more complex and data hungry thematic areas
(i.e. food systems); leveraging under-utilized data resources, namely textual data buried
in project reports; and enabling ex-ante data driven design and decision-making by
closing the gap between policymakers and project evaluation and by translating data and
project insights into actionable metrics. Together, these tools not only enhance IFAD’s
knowledge management but also embed learning and data-driven decision-making into
existing project design and implementation processes
• Department/Division: Programme Management Department (PMD)
• Project Type/Output: Report, Academic paper, Dataset, Seminar/meeting, Software tool
25

• Project Domain: Agriculture, Poverty
• Data Source:
o AI-based Intervention Dashboard: Corporate IFAD data on the investment portfolio

(financing, sectors, and project type) and textual data from project reports.
o Lessons Learned App: Textual data from project reports.
o Trend analyses of strategic themes: Corporate IFAD data on the investment portfolio
(financing, sectors, and project type) and textual data from project reports.
o Project performance prediction model: Corporate data on project performance
ratings at design, during implementation, and at completion, corporate financing
data and project features, and external open-source data on country-specific risk
factors and characteristics from World Bank (WDI), IMF (WEO), and other sources.
o Project impact prediction model: Household survey data from IFAD impact
assessments.
o Covid-19 impact prediction model: Google Mobility data, Google Trends data,
John Hopkins Coronavirus Resource Center data, and INFORM Covid- 19 risk data,
containing data on movement and search prevalence as well as actual reported
Covid-19 incidence and risk factors.
• Publicly Available Data: No

• Technology/Platform:
o AI-based Intervention Dashboard: Python, AWS Elasticsearch & Kibana

o Lessons Learned App: R Shiny
o Trend analyses of strategic themes: Python, R
o Project performance prediction model: Stata, R
o Project impact prediction model: Stata, Python
o Covid-19 impact prediction model: Python, R

• Related Sustainable Development Goals (SDGs): All SDGs
o Academia: several consultants from various academic institutions contributed to the

project over the two years of its implementation. Please refer to the acknowledgments
sections in the final reports.
o Phase 1 Report: https://www.ifad.org/en/web/knowledge/-/publication/accelerating

-knowledge-generation-for-data-driven-decision-making
o Phase 2 Report: https://www.ifad.org/en/web/knowledge/-/leveraging-artificial
-intelligence-and-big-data-for-ifad-2.0-phase-2?p_l_back_url=%2Fen%2Fweb
%2Fknowledge%2Fpublications
o Open-Source AI/ML: Open- Source AI/ML and code transparency are essential
elements that ensure that dashboards and apps can be updated in a real time fashion,
when new data comes in as well as integrated with the organization’s data ecosystem.
o Human element to improve algorithmic performance: the human element is essential
to improve the accuracy of algorithmic performance and overall quality of the models.
In the case of the AI-based intervention dashboard, IFAD staff and domain experts
have provided accurate taxonomies and training datasets that have fed the models,
producing classifications that are “realistic”.
26
o The complexity of IFAD project documentation is a key challenge for models that
require standardized data. Not only are project reports written in four different
languages, but they also vary in format and length. Data processing requires the
development of multi-lingual algorithms and sensitive data filtration strategies to
IFAD
ensure relevant text is extracted for analysis.
o Integration with existing ICT systems and business model: Sustainable and sustained
AI/ML and “big data” use cases require appropriate data repositories, server space,
and secure data storage within the business model.
o Institutional buy-in and support for innovation: Support and buy-in from key actors
and a willingness to experiment is crucial to the successful adoption and integration
of new data-driven tools for decision-making.
o The next steps would include the following activities: users’ validation and scaling-up
of the algorithms and tools generated so that they can be integrated in existing IFAD
systems (for automated reporting and briefs). Additionally, future phases of the project
would also foresee additional work to explore and validate the prediction models by
leveraging additional data sources and integrating additional cost data to predict
return on investments.
• Contact Information: Alessandra Garbero, Phd., Lead Regional Economist, Near East,
North Africa, Europe and Central Asia Division (NEN) (a.garbero@ifad.org).
Project 2: Mapping food systems national pathways with AI and ML techniques

• Project Description: The analysis utilizes a supervised learning approach to classify
sentences within the national pathways and dialogues documents. This method allows
for the identification of prevalent topics across various dimensions, including food system
drivers, components, and outcomes. The taxonomy for classification is derived from the
HLPE Food System Framework, ensuring a comprehensive and relevant analytical lens.
• Project Type/Output: Report, Dataset, Seminar/meeting
• Project Domain: Agriculture, Food systems analyses
• Data Source:
o Official public documents: Food systems Country level National Pathways and National
Food Systems Dialogue reports
• Technology/Platform: The technology used is primarily the GPT-3.5 Turbo model, which
is an advanced variant of OpenAI's Generative Pre-trained Transformer (GPT) model.
This model has been trained on large volumes of text data and can understand text
composition and generate natural language responses.
The AI model is accessed and utilized through the Azure OpenAI Service, whereby
Microsoft Azure's cloud computing platform is being utilized to host and access the
GPT-3.5 Turbo model for food systems analysis. This allows for scalable and efficient
deployment of the AI model.
Prompt engineering techniques were employed to influence the model's understanding
of the context of food systems and its ability to classify statements based on predefined
categories. The output of the model is relevance scores for each statement, indicating
the model's confidence in the statement's pertinence to each category.
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty, SDG 2– Zero
Hunger, SDG 3 – Good Health and Well-Being, SDG 5 – Gender Equality, SDG 6 – Clean
Water and Sanitation, SDG 8 – Decent Work and Economic Growth, SDG 10 – Reduced
27
Inequalities, SDG 12 – Responsible Consumption and Production, SDG 13 – Climate

Action
• Lessons Learned: This analysis was carried out on specific sub-regions to primarily test
the potential of this technology for policy learning. The main aim was to uncover insights
from publicly available national pathways for food system agenda achievement through
artificial intelligence (AI) and machine learning (ML) techniques. Several key lessons
were learned. First and foremost, the project underscored the importance of leveraging
advanced technologies like AI and ML to dissect and understand complex programmatic
documents across diverse socio-economic contexts. The use of supervised learning to
classify statements within the national pathways illuminated the distinct priorities and
actions of different countries, offering nuanced insights into their food system agendas.
A critical takeaway was the recognition of the heterogeneity in food system priorities
across the specific sub-regions analysed and countries’ income classifications. This
diversity necessitates tailored approaches in policy formulation and implementation to
address the unique challenges and opportunities within each context. Furthermore, the
project highlighted the significant potential of AI and ML in enhancing policy analysis
and decision-making processes. By applying the HLPE Food System Framework for
classification, the project demonstrated how structured taxonomies could provide a
coherent framework for analysing and interpreting complex data.
Lastly, the analysis revealed the challenges and limitations of data availability and quality,
particularly in countries where these documents are not available. This gap underscores
the need for continued efforts to promote data generation and sharing across countries
to support more comprehensive and informed policy analyses.
These lessons not only reflect the project's achievements in applying AI and ML to food
system policy analysis but also pave the way for future research and applications in this
field, emphasizing the critical role of technology in advancing sustainable food systems
globally.
• Contact Information: Alessandra Garbero, Phd., Lead Regional Economist, Near East,
North Africa, Europe and Central Asia Division (NEN) (a.garbero@ifad.org).
Project 3: Omnidata
• Project Description: Omnidata is IFAD's internal platform for data and analytics,
including technologies such as Artificial Intelligence, Machine Learning, and Geographic
Information Systems for advanced analytics. Omnidata empowers IFAD’s workforce to
do more with data and AI, developing data dashboards and AI solutions for use cases
across many different and thematic areas such as biodiversity, south-south triangular
cooperation, crop production, and more. Through the platform, a strong community of
practitioners are developing new skills and augmenting the way they work at IFAD across
all departments and divisions.
Through Omnidata, IFAD management and staff are working together side by side to
pilot and release solutions in a phased manner, practically demonstrating that AI provides
value not only in efficiency, but also in developing new products and insights with the
potential to transform ways of working. Omnidata is both a platform with robust technical
infrastructure, as well as a growing community of practitioners across the organization
being upskilled with new competencies required for the continuous delivery of AI
solutions for IFAD. The benefit of Omnidata’s platform and community approach is not
to solely focus on one or two AI use cases, but on democratizing technology and building
capacity to address many complex problems where AI can provide value.
• Department/Division: Corporate Services Department, Information and Communications
Technology Division
• Project Type/Output: Internal platform and community for data and analytics
28
• Project Domain: Agriculture, Environment, Gender, Health, Poverty, Weather

• Data Source:
o Internal data from corporate systems; External data on development, agriculture,
IFAD
socioeconomics, demographics, etc.; Geospatial data on land use, biota, climate,
environment, etc.; Unstructured data from reports and other documents.
o Oracle Business Intelligence, Microsoft Azure, Azure Data Lake, Azure AI Studios, Azure
ML Studios. Microsoft Power BI, GeoNode, GeoServer, QGIS, Python, JavaScript,
OpenAI, LLaMA, Mistral

o https://customers.microsoft.com/en-us/story/1729190323007055712-ifad-nonprofit
-azure-openai-service
• Lessons Learned: Choosing a single use case on which to pilot AI may be simple but
might not unlock institution-wide value and transformative impact over time. Creating a
platform underpinned by a community – as was done for Omnidata − was instrumental in
sensitizing and mobilizing the workforce to address actual needs across IFAD. This also
enabled wider impact in areas ranging from core operations to addressing biodiversity
to legal matters. This approach can be further leveraged in the years to come.
Advanced analytics and AI work typically ignites new ideas, insights, and applications for
new ways of working. This process is iterative and continuous. IFAD’s hands-on approach
to use cases is the most effective means of explaining in practice how AI can support
IFAD’s work. Embedding this approach more systematically to inform and augment work
where there is value will be the focus going forward.
The transformative potential of technologies such as AI has been confirmed. IFAD has
developed the requisite capacity and is now well positioned, not only internally but within
the United Nations system, to make the most of advanced analytics and AI – key elements
in support of the United Nations 2.0 agenda and ongoing discussions within the United
Nations System Chief Executives Board for Coordination.
• Contact Information: Ricardo Rendon Cepeda, ICT Analyst, Emerging Technologies
(r.rendoncepeda@ifad.org).
Project 4: Sustainability Action Plan Repository and Chatbot

• Project Description: IFAD’s Sustainability Action Plan aims to provide operational and
technical actions to strengthen IFAD’s performance related to sustainability of benefits
at the project level through an approach that commonly supports IFAD’s portfolio.
Due to the diversity of IFAD projects and the numerous factors that may impact project
sustainability, the action plan does not aim to address each one individually, but focuses
on providing the necessary building blocks to support widespread knowledge, behaviors,
and capacity to improve project sustainability.
As interventions to enhance sustainability of project benefits are context specific, a
key output of the action plan is an artificial intelligence (AI)-enhanced repository on
Sustainability. The repository offers a rich, interactive, and visual interface for country
teams to explore IFAD’s resources on sustainability of benefits. Users are able to extract
information from curated publications on sustainability, according to the criteria and
dimensions detailed in the action plan. The repository allows country teams to develop
a tailored approach to increasing sustainability and developing sound exit strategies for
better ownership in a wide variety of contexts. The repository consists of how-to-do notes,
analysis, and useful resources developed by IFAD, as well as tools and guidance from
29
external sources. These resources are processed with text-centric AI and automation tools
that extract key concepts and generate focused summaries and data visualizations.
The repository is also connected to a Generative AI (GenAI) chatbot interface, which
allows users to ask any questions about sustainability. The chatbot searches through
the repository contents and provides a referenced answer to the question using Large
Language Models (LLMs). This solution is implemented as a Retrieval Augmented
Generation (RAG) search engine with a natural language interface.
Technology Division
• Project Domain: Agriculture, Environment, Gender, Health, Poverty, Weather
• Data Source:
o Publicly available reports, documents, how-to-do notes, analysis, frameworks, and

other resources on sustainability across social, environmental, financial, technical, and
institutional dimensions.
o Microsoft Azure, Azure Data Lake, Microsoft Power BI, Python, JavaScript, OpenAI,
LLaMA, Mistral

(r.rendoncepeda@ifad.org)
Project 5: Corporate Portfolio Stocktake on Food Systems

• Project Description: IFAD’s corporate portfolio stocktake provides a comprehensive
overview of the performance and quality of the organization’s operations and results,
paying special attention to food systems, sustainability, and scaling up in 2023. To better
understand IFAD’s operating context and enhance the analysis of the global portfolio,
interactive data dashboards enhanced by AI technologies were developed to evaluate the
portfolio in terms of relevance and alignment on four sub-components of food systems:
production, livelihoods, climate change, and resilience.
Using Large Language Models (LLMs), AI translation tools, and other natural language
processing (NLP) techniques, statements from national pathways, summit dialogues,
and country strategies were processed, transformed, ranked, and classified according to
IFAD-specific and multi-dimensional food system taxonomies.
Technology Division
• Project Domain: Agriculture, Environment, Poverty
• Data Source:
o Summit Dialogues and National Pathways from the UN Food Systems Coordination
Hub; Country Strategic Opportunities Programmes (COSOPs) and Country Strategy
Notes (CSNs) from IFAD
30
o Link to data: https://www.unfoodsystemshub.org; https://www.ifad.org/en/cosop
IFAD
o Microsoft Azure, Microsoft Power BI, Python, OpenAI

Hunger, SDG 3 – Good Health and Well-Being, SDG 6 – Clean Water and Sanitation, SDG
8 – Decent Work and Economic Growth, SDG 9 – Industry, Innovation, and Infrastructure,
SDG 11 – Sustainable Cities and Communities, SDG 12 – Responsible Consumption and
Production, SDG 13 – Climate Action, SDG 14 – Life Below Water, SDG 15 – Life on Land,
SDG 16 – Peace, Justice, and Strong Institutions
Project 6: Global Portfolio Analysis of Fragile Situations

• Project Description: The global portfolio analysis of fragile situations presents a new way
of using Artificial Intelligence (AI) technologies to analyze IFAD's operating context as
stated in country strategies, project design reports, and integrated project risk matrices.
The analysis uses IFAD's definition of fragility and six additional sub-definitions of fragility
to evaluate descriptions of IFAD's operating context, resulting in AI-generated fragile
situation scores and fragile situation justifications to provide deeper insights into IFAD's
portfolio.
The analysis revealed that, as fragility becomes increasingly complex and the context of
work around fragility changes, the specific application of IFAD’s definition of fragility can
identify a broader range of fragile situations across the IFAD portfolio. The analysis also
showed that fragility occurs in different degrees across different dimensions, highlighting
the importance of assessing fragility in the specific context in which IFAD operates.
Advanced technologies like AI, Generative AI (GenAI), and Large Language Models
(LLMs) can be combined with detailed source data such as IFAD's corporate documents
to reveal new insights on fragility in order to better understand fragile situations across
IFAD’s portfolio, especially through interactive data dashboards that allow for multiple
lines of investigation across countries and projects.
Technology Division
• Project Type/Output: Report, Software tool
• Project Domain: Agriculture, Environment, Gender, Health, Human Rights, Justice, Poverty
• Data Source:
o Programme and Project documents from IFAD; Regional and Country Documents
from IFAD
o Link to data: https://www.ifad.org/en/programme-and-project-documents; https://
www.ifad.org/en/regional-and-country-documents
o Microsoft Azure, Microsoft Power BI, Python, OpenAI

Hunger, SDG 3 – Good Health and Well-Being, SDG 5 – Gender Equality, SDG 6 – Clean
31
Water and Sanitation, SDG 8 – Decent Work and Economic Growth, SDG 9 – Industry,
Innovation, and Infrastructure, SDG 10 – Reduced Inequalities, SDG 11 – Sustainable Cities
and Communities, SDG 12 – Responsible Consumption and Production, SDG 13 – Climate
Action, SDG 14 – Life Below Water, SDG 15 – Life on Land, SDG 16 – Peace, Justice, and
Strong Institutions
Project 7: Farmer Chat

• Project Description: With the objective to support improvements to farm productivity
and climate change adaptation, IFAD is partnering with Digital Green, a not for profit
organisation and are engaging to pilot ‘Farmer.Chat’ an AI-enabled digital assistant for
field extension agents and develop a ‘proof of concept (POC)’ as well as lessons from
implementation with IFAD supported projects in Nigeria and Kenya.
Extension Agents will be supported to utilize this digital assistant through a smartphone
application, called ‘Farmer.Chat’ to provide customised advisory to smallholders in local
language’. Farmer.Chat Extension agents can triage farm-level needs with local language
text and voice inputs and text and audio responses to facilitate the efficient delivery of
customized advisory recommendations to small-scale farmers.
The proposed engagement will take a phased approach, with Digital Green collaborating
with IFAD's country offices and local partners to scope and deliver "Proof of Concept"
(POC/pilot) pilot initiatives. The main objective of the pilot is to test out the technology
capabilities to support farmers and local agents in the value chains that are supported
by IFAD projects and factor in the Natural language processing as a value addition to
support inclusivity. The implementation and operational model for the digital assistant will
a focus on enhancing advisory services as a base for replication in future IFAD supported
projects.
• Department/Division: Sustainable Production Markets and Institutions Division
• Data Source:
o Information typically comes from research reports such as of international agencies,

the CGIAR, FAO, national agricultural research and extension systems, university
research reports, soils and meteorological data sources, some video content, etc. But
the important thing to note is that any/all content included in the Content corpus is first
scrutinized and attributed to a legitimate, credible source and then is included ONLY
after the designated Content Manager, e.g. Ministry of Agriculture, Kenya Agriculture
Livestock Research Organization, National Agricultural Extension and Research Liaison
Service, etc has approved its inclusion.
o Open AI, MSR, Meta AI, Google, Telegram

• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty, SDG 13 – Climate
Action
o UN Partners: IFAD
32
o Government: Government of Kenya and Nigeria

o Civil Society: Digital Green
o Academia: National Agricultural Extension and Research Liaison Service, Nigeria
IFAD
• Contact Information: Brenda Gunde (b.gunde@ifad.org)
Project 8: SMARTFARM: A data and digital technology driven farm and farm
management solution for climate resilience
• Project Description: The SMARTFARM project objective is to enable smallholder farmers
in Ethiopia and Rwanda to adapt to climate change and improve their crop productivity
and food and nutrition security through the delivery of real-time weather and climate data,
and through farm advisory services. The project is funded by the GEF Challenge fund for
Adaptation Innovation with IFAD as the main executing agency and CropIn Technologies
as the lead implementing agency.
SMARTFARM will promote the use of real-time weather and climate data, along with data-
driven farm advisory services available to smallholder farmers. SMARTFARM leverages
digital technologies such as Artificial Intelligence, Machine Learning, Remote Sensing
& mobile telecommunication to offer data driven agriculture digital services weather and
climate advisory (WACS) and Data Driven Agriculture Advisory (DDAS) on innovative and
collaborative digital platforms, i.e., cloud web-based & mobile application.
Together, these will increase the adoption of climate-resilient agriculture practices and
enhance rural communities' resilience to climate change. Digitalization of farm location
data will be used to provide context-specific responses based on prediction data using
advance analytical techniques to generate advisories.
The target group is made up of 130,000 SHFs, including women SHFs, (100,000 in
Ethiopia and 30,000 in Rwanda), cumulative over a two-year period, with the help of
2,000 village/agroentrepreneurs. The project will be implemented over two years and
will have the following components:
Component 1. Deployment, adoption and scale up of SMARTFARM to increase the
climate adaptation and resilience of 130,000 SHFs over a 2-year period, including women
SHFs, with the help of 2,000 village/agroentrepreneurs.
Component 2. Capacity-building of selected farmers’, producers’ and rural organizations
and institutions, 2,000 village/agroentrepreneurs (of whom at least 25 per cent will be
women) and implementation partners for knowledge and asset transfer.
Component 3. Creation of partnerships, knowledge and tools to promote engagement
among off-takers, buyers and institutions for credit access and market linkages
• Department/Division: Sustainable Production Markets and Institutions Division
• Project Domain: Agriculture, Environment, Poverty
• Data Source:
o IBM weather data, Landsat, World Bank Climate Change Knowledge Portal and other
institutional repositories like the World Meteorological Organization. Some of the
data is publicly available. Insitu data and farm level data will also be collected through
village-based agents.
o Own and use Microsoft Azure.
33
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty, SDG 13 – Climate

Action
o UN Partners: IFAD
o Government: Government of Ethiopia and Rwanda
o Private Sector: CropIn Technologies
o https://www.cropin.com/invest-in-gef-ifad-climate-adaptation-project-in-africa
• Contact Information: Brenda Gunde (b.gunde@ifad.org)
Project 9: Voice Companion for Farmers by Viamo

• Project Description: AI-powered Voice Companion for agriculture extension, delivering
instant, localized climate-smart agriculture information in local languages.
• Department/Division: Innovation Unit
• Project Type/Output: Startup funded by IFAD in the context of the M4D Joint Open
Innovation Challenge (2023)
• Project Domain: Agriculture, Environment
Hunger
o https://viamo.io/
o https://challenges.moonshots4dev.org/en/challenges/challenge2023/pages/about
-us?lang=en
o https://www.ifad.org/en/web/latest/-/m4d-joint-open-innovation-challenge-pitch
-event
• Contact Information: Gladys H. Morales (g.moralesguevara@ifad.org)
Project 10: Farmer Lifeline Technologies

• Project Description: Farmer Lifeline's AI devices empower smallholder farmers by
detecting and predicting crop pests and pathogens, promoting climate-smart practices
for enhanced food security.
o TensorFlow, Python Django

Hunger
34
o Private Sector: Farmer Lifeline Technologies
IFAD
o https://farmerlifeline.co.ke/
-us?lang=en
-event
• Contact Information: Gladys H. Morales (g.moralesguevara@ifad.org), Carmela Lopez (ca.

lopez@ifad.org)
Project 11: Agpreneur

• Project Description: Agpreneur is an all-in-one Agri-fintech Platform providing tailored
financing, dynamic education, predictive input demands, and optimized market
interactions for African farmers.
Hunger
o Private Sector: Agpreneur
o https://agpreneur.com.ng/
-us?lang=en
-event
• Contact Information: Gladys H. Morales (g.moralesguevara@ifad.org), Carmela Lopez (ca.

lopez@ifad.org)

SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17
3. Relevant Links
https://www.ifad.org/en/
Contact Information
Alessandra Garbero, Phd., Lead Regional Economist, Near East, North Africa, Europe and
Central Asia Division (NEN) (a.garbero@ifad.org)
Ricardo Rendon Cepeda, ICT Analyst, Emerging Technologies (r.rendoncepeda@ifad.org)
35
International Labour Organisation
Project 1: Algorithmic management in the logistics and healthcare sectors

• Project Description: This research explores how the algorithmic management practices
that are often associated with platform work, such as rating systems, surveillance and
control through tracking devices, online logging of work hours, the use of diverse forms of
employment, etc. are being increasingly utilised by traditional companies in the logistics
and healthcare sectors, thus leading to ‘platformisation’ of work. The project will look
at the impact of such practices on work organisation, efficiency and working conditions
in European (France, Italy) and non-European (India, South Africa) countries. It will also
explore how the data that is collected through these practices is used by the firms, and who
has control and rights over such data. The objective of this research is to understand the
extent of the penetration of these practices in the logistics and healthcare sectors in both
European and non-European countries, whether the experiences are similar or different,
and the role of governments and social partners (workers’ and employers’ organisations)
in addressing some of the challenges due to the rapid technological transformations. It
will analyse whether the algorithmic management practices in the logistics and healthcare
sectors leads to improved autonomy, flexibility, and working conditions for the workers. It
will examine how the existing social and human rights standards are enforced in a context
in which key employer functions are carried out by algorithms and are thus opaque. It
will explore what public policies are required to address some of the challenges relating
to worker surveillance, working conditions, thus ensuring greater transparency in the
algorithms.
• Department/Division: Research Department
• Project Type/Output: Report/Academic paper/Seminar/meeting
• Project Start Year: January 2021
• Project End Year: December 2022
• Project Domain: Health, Human Rights, Logistics, Improving working conditions, Reducing
inequalities
• Data Source: The project will adopt a case study approach and collect data from both
workers and managers in enterprises in the logistics and healthcare sector in India and
South Africa (non-European countries) and France and Italy (European countries).
• Publicly available data: No
• Related Sustainable Development Goals (SDGs): SDG 8 – Decent Work and Economic
Growth; SDG 10 – Reduced Inequalities
36
o Academia: In the process of deciding the academic partners in India and South
Africa. This project is in collaboration with the Joint Research Centre, Seville of the
European Commission.
ILO
• Relevant Links and Multimedia: https://www.ilo.org/employment/Whatwedo/Projects/
building-partnerships-on-the-future-of-work/lang--en/index.htm
• Contact information: Uma Rani (amara@ilo.org)
Project 2: Research on worker privacy and personal data protect

• Project Description: The ILO conducted a comparative legal study on the protection of the
personal data of workers and their right to privacy in the light of the ILO Code of Practice
on the protection of workers’ personal data. It also conducted a preliminary analysis of
the issues raised in relation to digital monitoring of workers and algorithmic management
as they have grown in importance with the development of the digital economy and
teleworking, including during the COVID-19 crisis. The outcome of this research will
contribute to the preparation of the Tripartite Meeting of Experts on decent work in the
platform economy to be held in 2022.
• Entity Name: INWORK
• Department/Division: WORQUALITY
• Project Type/Output: Other: Ongoing research
• Project Domain (field): Human Rights, Justice, Labour
• Publicly available data: Yes
Growth
• Contact information: Martine Humblet (humblet@ilo.org)
Project 3: Competency Profiling App (funded through the PROSPECTS partnership)

• Project Description: The number of international migrants and refugees is growing
rapidly. Between 2000 and 2017, the number increased from 173 million to 258 million
(an increase of almost 50%). To reap the benefits of migration, states need to enable
migrants and refugees to integrate into the labour market and society through access to
employment opportunities. One important factor that prevents this relates to the under-
utilization of migrants’ and refugees’ skills in countries of destination and upon return.
It is therefore vital not only to support governments in adopting policies and legislation
that facilitates the access of migrant workers and refugees to the labour market but also
to support them in developing and contextualizing technical solutions that may assist
them to reduce the strain on public services whilst improving service delivery to the host-
population.
Therefore, the ILO Skills and Employability Branch is piloting a web-application in Kenya,
Egypt and Lebanon
o for 1500 refugees, migrants and host populations in developing countries

o that allows individuals to capture and present their past experiences, skills and
competences acquired both formally and informally.
The multi-lingual and minimal text-typing methodology allows individuals to produce a

profile of their skills and competencies summarized in a standardized Curriculum Vitae,
and in more detailed occupational competency profiles. Counsellors of employment
37
services, UNHCR, NGOs or other service providers can also assist in filling in and
completing the profile.
• Entity Name: HQ
• Department/Division: ILO Skills and Employability Branch
• Project Type/Output: Software tool/Application
• Project End Year: Ongoing pilot
• Project Domain: Education and training, Job and Employment
• Data Source: Individual level data is collected through implementing partners in Egypt,
Kenya and Lebanon.
• The European System of Occupational Classifications ESCO, with more than 13.000 skills
and 3000 occupations, is utilized as a reference framework for coding / classifying skills.
An AI engine learns which skills tend to 'appear in combination' and prompts the right
follow-up questions to the user / employment service provider.
• Technology/Platform: Testing and applying an AI-based competency profiling tool in
Egypt, Kenya and Lebanon
Decent Work and Economic Growth
o UN Partners: under PROSPECTS

o Government: ABA (Egyptian public employment services)
o Private Sector: Project Partners: Skilllab (start-up that is developing the app)

• Contact information: Christine Hofmann (hofmann@ilo.org)
Project 4: Digitalization of national TVET and skills systems: Harnessing technology

to support LLL
• Project Description: The project explores the potential for digitalisation of different
functions in TVET, skills development and Life--long learning systems. Digital transition
of TVET and skill systems goes far beyond taking training products and services online.
A holistic and coordinated approach to digitalisation should be taken that looks at each
high-level function of a national skills system, and its potential for digitalisation.
This report describes and updates the picture of digital vocational education and training,
providing an overview of the issues surrounding digitalisation across the key functional
areas of skills systems.
It gives an introduction to key frameworks and tools; concrete examples of national
initiatives, adaptable digitalisation models and practical guides; as well as providing
initial guidance on implementation, to deploy a strategic approach to the digitalisation
of national TVET and skills development systems at the country level.
• Department/Division: SKILLS Branch
• Project Type/Output: Report
• Project Status: Complete
• Project Domain: Education and training
38

• Contact information: Karine Sonigo (sonigo@ilo.org)
ILO
Project 5: The Skills Gap In Jordan And Impact On Unemployment
• Project Description: Mehnati provides a solution for labour market information. The initial
conceptual idea was born in a project ‘Applying the G20 Training Strategy Project’ that
was funded by the Russian Federation through ILO Jordan office. The goals was to bridge
the employability gap and to promote development. The design of the concept was
undertaken in close collaboration with all relevant stakeholders with the aim to offer a
smooth and easy to use platform to Employers, Job Seekers and Training providers.
Various national stakeholders including the chamber of industry, chemicals and garments
sector skills councils as well as a number of training providers participated in the design
of the concept. This process revealed the high need for the Mehnati platform in different
sector and geographical places. Also, it revealed the need to cover various skill types,
ranging from TVET, Modular bridge learning, and work readiness for digital gig based
workers.
To ensure that Mehnati learns from previous experiences and capitalizes on know-
how achieved across various ILO projects, it will be implemented in phases or mini-
projects, with the ultimate objective of creating value across the spectrum of ILO reach
in different sectors and geographical places. Mehnati will be integrated into the national
e-counselling platform Jordan has created with support of the ILO.
• Entity Name: RO Arab States
• Project Status: Development/Ongoing
• Project Start Year: 2020 (piloting)
• Project Domain: Education and training; Gender; Poverty
• Data Source: LMI, Training content, Employment opportunities
• Technology/Platform: This currently being specified. But the portal uses AI for the “job fit
test”, which helps automate some aspects of career counselling in the experience
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 4 – Quality
Education; SDG 5 – Gender Equality; SDG 8 – Decent Work and Economic Growth; SDG
9 – Industry; Innovation, and Infrastructure; SDG 10 – Reduced Inequalities; SDG 17 –
Partnerships for the Goals
• Contact information: Kishore Kumar Singh (singhkk@ilo.org)
Project 6: Online digital labour platforms in China: Working conditions, policy issues
and prospects
• Project Description: Digital labour platforms have been proliferating in China since 2005,
making China one of the world’s largest platforms economies. This paper summarizes
the results of an ILO survey, conducted in 2019, of workers’ characteristics and working
conditions on three major digital labour platforms. Using the survey data generated, it
provides first-hand information on worker demographics, motivations, and experiences.
This paper also compares the findings between the Chinese platforms and dominant
Western platforms, the object of previous ILO studies. The paper concludes with a
discussion about the need for institutional reforms and suggests some possible avenues
for implementing policies to improve working conditions.
• Project Type/Output: Working Paper
39

• End Year: 2021
• Project Domain: Human Rights, Labour
• Data Source: Economic data
Growth
• Relevant Links and Multimedia
o https://www.ilo.org/wcmsp5/groups/public/---ed_protect/---protrav/---travail/
documents/publication/wcms_768699.pdf
• Contact Information: Martine Humblet (humblet@ilo.org)
Project 7: Platform work and the employment relationship

• Project Description: This working paper analyses national and supranational case law
and legislation about the employment status of platform workers. It does so by referring
to the ILO Employment Relationship Recommendation, 2006 (No. 198). It finds that this
Recommendation provides for a valuable compass to navigate the issues that emerge
from the analysis of the existing case law and legislation about platform work.
• Project Domain: Human Rights, Justice, Labour
• Data Source: Legal Information
Growth
• Links and Multimedia: https://www.ilo.org/global/publications/working-papers/WCMS
_777866/lang--en/index.htm
Project 8: Digital Work in Eastern Europe: Overview of Trends, Outcomes, and Policy
Responses
• Project Description: This paper presents the emergence and growth of digital labour
markets in Eastern Europe over the period 1999-2019. It presents the profiles of digital
workers, their working conditions and discusses how these are shaped by the business
models of digital labour platforms.
40
• Data Source: Legal and economic data.

Growth
ILO
• Relevant Links and Multimedia: https://www.ilo.org/wcmsp5/groups/public/-- - dgreports/
---inst/documents/publication/wcms_794543.pdf
Project 9: How Do You Lip Read a Robot? – Recruitment AI has a Disability Problem
• Project Description: Information sharing webinar arising from discussion within ILO Global
Business and Disability Network on the risks associated with using AI powered recruitment
software, based on emerging evidence that it leads to the exclusion of candidates with
different types of disabilities.
• Department/Division WORKQuality, Gender, Equality, Diversity and Inclusion Branch
• Project Type/Output: Seminar/ Meeting
• Project Domain: Human Rights, Equality and non-discrimination, Employment
• Relevant Links and Multimedia: https://youtu.be/ndA-Z_wJ31s
• Contact information: Stefan Tromel (tromel@ilo.org)
Project 10: Tripartite Meeting of Experts on decent work in the platform economy
• Project Description: The ILO Centenary Declaration for the Future of Work, adopted on
21 June 2019, calls on all ILO Members to put in practice “policies and measures that
ensure appropriate privacy and personal data protection, and respond to challenges
and opportunities in the world of work to the digital transformation of work, including
platform work”. On 27 March 2021, at its 341st Session, the Governing Body decided “to
request the Office to convene a tripartite meeting of experts on the issue of “decent work
in the platform economy” in the course of 2022”. This meeting is expected to take place
in September 2022.
• Project Type/Output: Seminar/Meeting
• Data Source: Economic and legal data.
Growth
41
Project 11: Technical meeting on the future of work in the arts and entertainment
sector
• Project Description: At its 341st Session, in March 2021, the Governing Body of the
International Labour Office endorsed the convening of a technical meeting on the future
of work in the arts and entertainment sector, which took place 13-17 February 2023.
The purpose of the meeting was to discuss opportunities and challenges for decent
work in the sector in the context of digital technologies, globalization, environmental
sustainability, demographic changes and a human-centred COVID-19 recovery, with the
aim of adopting conclusions, including recommendations for future action.
In order to inform the meeting, a report for discussion was published including trends
and developments regarding digitalization in the retail sector. The discussion took place
based on points for discussion, and the meeting resulted in agreed conclusions including
recommendations for the Office and for the ILO constituents.
• Department/Division: SECTOR
• Project Type/Output: Seminar/meeting
• Project Domain: Future of work, AI, Digitalization, Decent work
Growth
• Contact information: Margherita Licata (licata@ilo.org)
Project 12: Technical meeting on Tdigitalization in the retail sector as an engine for
economic recovery and decent work
• Project Description: At its 346th Session, in October–November 2022, the Governing
Body of the International Labour Office endorsed the convening of a technical meeting
on digitalization in the retail sector as an engine for economic recovery and decent work,
which took place 25-29 September 2023..
The purpose of the meeting was to discuss opportunities and challenges for the future
of work in retail and commerce in the context of digitalization as a vehicle to ensure a
human-centred economic recovery, including from the COVID-19 pandemic, with the aim
of adopting conclusions, including recommendations for future action..
In order to inform the meeting, a report for discussion was published including trends
and developments regarding digitalization in the retail sector.The discussion took place
based on points for discussion, and the meeting resulted in agreed conclusions including
recommendations for the Office and for the ILO constituents.
• Project Domain: Future of work, Digitalization, Decent work
Growth
• Contact information: Margherita Licata (licata@ilo.org)
42
Project 13: Issues paper for the Technical meeting on the future of decent and
sustainable work in urban transport services
• Project Description: Urban passenger transport systems are crucial to the achievement of
ILO
sustainable cities and communities and contribute towards a zero-carbon future. Yet, the
industry is faced with disruption from the pandemic, a technological revolution, as well as
with plummeting ridership and occupational safety and health challenges. The meeting
discussed challenges and solutions relating to the future of decent and sustainable
work in urban passenger transport operations and services, with the aim of adopting
conclusions, including recommendations for future action.
The document includes a section on technological innovation and AI, which emphasizes
ITU’s “AI for good” mandate. The paper aims to inform the meeting’s discussion,
highlighting the trends steering major sectoral changes and analysing how these impact
employment, labour and social protection, and the sector’s regulatory environment.
• Project Domain: Transport
• Reported as part of 2022Compendium on UN AI Activities? Yes
(3.6 Road Safety); SDG 11 – Sustainable Cities and Communities (11.2 Sustainable public
transport)
• Links and Multimedia: https://www.ilo.org/sector/activities/sectoral-meetings/WCMS
_726153/lang--en/index.htm, https://www.ilo.org/sector/activities/sectoral-meetings/
WCMS_818255/lang--en/index.htm
• Contact information: Alejandra Cruz Ross (cruzross@ilo.org)

SDGs 1, 3, 4, 5, 8, 9, 10, 11, 17
3. Relevant Links
www.ilo.org
Contact information: Ms Irmgard Nübler, Senior Economist (nubler@ilo.org)
43
International Monetary Fund
Project: Powering the Digital Economy: Opportunities and Risks of Artificial

Intelligence in Finance
• Project Description: This paper discusses the impact of the rapid adoption of artificial
intelligence (AI) and machine learning (ML) in the financial sector. It highlights the
benefits these technologies bring in terms of financial deepening and efficiency, while
raising concerns about its potential in widening the digital divide between advanced
and developing economies. The paper advances the discussion on the impact of this
technology by distilling and categorizing the unique risks that it could pose to the
integrity and stability of the financial system, policy challenges, and potential regulatory
approaches. The evolving nature of this technology and its application in finance means
that the full extent of its strengths and weaknesses is yet to be fully understood. Given the
risk of unexpected pitfalls, countries will need to strengthen prudential oversight.
• Department/Division: Money and Capital Markets Department and Information and
Technology Department
• Project Type/Output: White Paper
• Project Domain: Finance
Growth; SDG 10 – Reduced Inequalities
• Data Source: Many Sources
• Links and Multimedia: https://www.imf.org/en/Publications/Departmental-Papers-Policy
-Papers/Issues/2021/10/21/Powering-the-Digital-Economy-Opportunities-and-Risks-of
-Artificial-Intelligence-in-Finance-494717
• Technology/Platform: Research Paper

SDGs 8 and 10
3. Relevant Links
www.imf.org
Contact information: EL Bachir Boukherouaa (eboukherouaa@imf.org)
44
International Maritime Organization
IMO
Project 1: Maritime Autonomous Surface Ships (MASS)
AI is part of the MASS functionality as the ship system (e.g. for collision avoidance) will have
to learn from a large number of scenarios (Ship situations) to decide for the best anti-collision
action to be taken. There will be many other AI applications for MASS. In this context, IMO
developed a set of interim guidelines for the conduct of MASS trials, stipulating that trials
should be conducted in a manner that provides at least the same degree of safety, security and
protection of the environment as provided by the relevant instruments. IIMO has also agreed
on a road map for the development of a goal-based MASS Code which, as a first step, will
be non-mandatory and is expected to be effective on 1 January 2025. . With the experience
gained, paired with new emerging concepts and technology, IMO will be working towards a
mandatory Code, which is envisaged to enter into force on 1 January 2028.
• Project Type (Status): Framework/Strategy/Policy and Regulation (Development)

• Project Domain: Shipping
• AI Approach: AI as one of the drivers for enabling MASS operation; the IMO as the body
responsible for regulating MASS, together with standard-setting organizations (e.g. ISO,
IEC) , will need to consider safety and security aspects of AI systems used for MASS
• Related Sustainable Development Goals (SDGs): SDG 8 Decent work and Economic
growth, SDG 9 Industry, Innovation and Infrastructure, SDG 11 Sustainable Cities and
Communities, SDG 13 Climate Action, SDG 14 Life below Water
• Project Website (links): Autonomous shipping (imo.org)
Project 2: Marine Environmental Protection
IMO under its Global Industry Alliance to support Low Carbon Shipping (Low Carbon GIA) is
working towards promoting Just-In-Time (JIT) arrivals of ships through the use of ship and port
specific data with an aim to reduce fuel consumption and GHG emissions at sea and in ports.
IMO adopted MEPC in 2022 resolution 366 (79) which invites IMO Member States to encourage
voluntary cooperation between the port and shipping sectors to contribute to reducing GHG
emissions from ships. It is expected that AI will help with many elements of this cooperation,
especially voyage optimization, weather routeing, and just-in-time (JIT) arrivals in ports.
The in July 2023 adopted 2023 IMO Greenhouse Gas (GHG) Strategy (resolution MEPC.377(80)
also acknowledges the importance of addressing the human element, including the impact
on seafarers and other maritime professionals, in the safe implementation of the Strategy. It
45
is expected that AI could assist in training seafarers with the safe handling of alternative low-
carbon fuels and other decarbonization technology.
• Project Type (Status): Ideation

• Project Domain: Marine environmental protection
• Datasets: Port call data
• Related Sustainable Development Goals (SDGs): SDG 9 Industry, Innovation and
Infrastructure, SDG 13 Climate Action
• Project Partners: –Members of the Low Carbon GIA, and other maritime stakeholders
Project 3: Digital Review
IMO is undertaking a digital review, to ensure a future-viable IMO, as part of a broader Functional
Review. The Secretariat aims to ascertain what is working well and what is not working well,
what is needed and what is redundant, and to develop a digital strategy and roadmap for the
next 5 years to ensure the Secretariat embraces digital opportunities in a way which will make
it future viable with regards to digital access.
• Project Type (Status): Framework/Strategy/Policy (Development)

• Project Domain: Internal management
• AI Approach: Digital review
• Related Sustainable Development Goals (SDGs): SDG 9 Industry, Innovation and
Infrastructure; SDG 13 Climate Action; SDG 14 Life below Water; SDG 16 Peace, Justice
and Strong Institutions; SDG 17 Partnership for the Goals
Project 4: Knowledge building
A seminar/workshop to strengthen knowledge of the maritime community/IMO staff, as well

as delegates.
• Project Type (Status): Training (Ideation)

• Project Domain: Capacity building
• AI Approach: Events

SDGs 3, 8, 9, 11, 13, 14, 16 and 17
3. Relevant links
www.imo.org
Contact information: Ms Galuh Rarasanti, Senior Maritime Advisor (grarasan@imo.org)
46
International Organization for Migration
IOM
Project: Expanding IOM’s Displacement Tracking Matrix’s (DTM) Data Foundations
Project Description: the process of integrating OpenAI tools into the Displacement Tracking
Matrix (DTM) core systems. This process is ongoing in three significant, interlinked and mutually
re-enforcing areas:
a) Semantic search in the Central Data Dictionary
DTM’s Central Data Dictionary (CDD) is built to provide access to all DTM missions worldwide
so that users in the field can select survey indicators from a global question bank. Using a
proprietary Form Builder application and Form validator, data collection tools and indicators
deployed now share a standardised structure and codification, resulting in standardised coded
DTMs datasets. Missions in diverse contexts can then gather the comparable information in
the same format – data that can then be fed into a central database, extracted, and analysed
in the same way. Yet, with hundreds of indicators in the Dictionary, field users need to find
the questions that fit their contexts and gain them the answers they can realistically use to
provide the best support possible to affected communities. With that need in mind, an AI-
powered semantic search has been deployed to ease the process of finding indicators and
building survey tools. For example, within the semantic search, a field mission can search for
“barrier questions” and receive a list of the standard indicators that fit, whether barriers to
access for people with disabilities or language barriers, so that they can select the ones they
need more quickly. In addition to speed and ease for missions, the semantic search facilitates
improved indicator selection by offering a reduced range of relevant indicators so missions
can see options they might not have thought of without being overwhelmed by unrelated
questions. It also allows the user to search natively in one of the five supported languages
and will allow DTM to support additional languages natively in the application more easily.
This tool ultimately allows users to develop survey tools from a pre-developed indicator bank
for contextual relevance, whilst adding common structuring and codification to the output
datasets.
b) Retro-fitting historical data to the new DTM DataLake
DTM’s ongoing work to centralize and harmonize displacement data across region and country
contexts includes the re-development of the medallion structure Data Lake. New data will
only be ingested into the DTM Data Lake if the naming structure is compliant with the Central
Data Dictionary. By default, data gathered going forward will be compliant since they will be
gathered using forms generated using the Data Dictionary as mentioned above. However,
historical data (pre-2024) will need to be retroactively recoded to the new data names. OpenAI
47
search functions and clustering tools will be used to find and match similar questions that
can be changed to the current standardized versions to sort and format existing data. Using
AI, historical data will be recoded on scale, allowing DTM to benefit from its long history
and contextual information within a new, harmonized data structure, which will enable other
innovations, such as displacement projects, and support anticipatory action.
c) DTM Chatbot for non-technical querying of its data systems
Utilizing the combined capabilities of OpenAI's large language model and Langchain's SQL
Agent, DTM have developed a chatbot integrated with our aggregate baseline assessment
database. This innovative chatbot empowers users to effortlessly query the database using
natural language and receive responses in a similarly natural language format. Leveraging the
SQL Agent functionality, the chatbot is adept at answering inquiries based on both the structure
(schema) and contents of the database.
With this chatbot, we gain flexibility in extracting insights from our database. For instance, users
can inquire about various statistics such as the total number of Internally Displaced Persons
(IDPs), Migrants, or Returnees across different countries, regions, round and timeframes.
Additionally, complex queries are supported, enabling users to delve into specific details,
such as identifying the country, date and round with the highest number of IDPs over the past
decade.
This integration not only streamlines the process of accessing critical information but also
enhances the accessibility of our data for users across diverse skill levels. The chatbot serves
as a user-friendly interface, bridging the gap between complex database structures and end-
users, thereby facilitating more informed decision-making and analysis.
Keywords: Central Data Dictionary (CDD), humanitarian data collection, interoperability,

harmonization of data systems
• Department/Division: Global Data Institute (GDI) – Displacement Tracking Matrix

• Project Type/Output: Global consolidation of DTM data into the DTM Data Lake
• Project Domain: data structure, displacement, human mobility
• Related Sustainable Development Goals (SDGs): 10.7 facilitate orderly, safe, and
responsible migration and mobility of people, including through implementation of
planned and well-managed migration policies
o Angular, .NET Core, HTML, CSS, and JavaScript for the frontend of the system
o Laravel 8 (PHP 7.4 and MySQL 8) for the back end of the system
o Google Maps Services and GeoIP for the Geolocation of the interview
o Python (pandas, scipy, nltk, VADER, langchain) and Jupyter notebook
o PowerBI for daily monitoring dashboard
o Large Language Models (OpenAI)
o Azure Services
– Azure App Service

– Azure Cosmos DB (NoSQL)
– Azure Cosmos DB for Mongo DB
48
– Azure SQL Server

– Azure Front Door
– Active Directory services
IOM
– Azure Storage account
– Data Factory
– Azure Open AI
o Civil Society: Microsoft, Kainos

SDG 10
3. Relevant Links
www.dtm.iom.int
Contact information: Robert Trigwell (rtrigwell@iom.int)
49
International Trade Centre
Project 1: Strengthening policy support for SMEs in developing countries using

artificial intelligence
• Project Description: TradeAI is an AI-powered online trade policy advisory application that
offers policymakers quick and complimentary access to relevant policy solutions for their
most common trade-related challenges. This app empowers policymakers to understand
trade strategy, connect with other professionals, and create their own streamlined trade
and SME (Small and Medium Enterprises) competitiveness strategies.
The project leverages AI to identify potential policy solutions that can address a country's
trade competitiveness constraints. It utilizes image recognition, natural language
processing (NLP) techniques, and neural networks to discover patterns in data and
generate new policy insights. TradeAI was initiated in 2020 as a pilot project to evaluate
how AI could help the International Trade Centre (ITC) identify pertinent information within
extensive repositories of trade policy documents. Throughout 2021 and 2022, the project
delved into ITC’s comprehensive database of national trade strategies, employing a mix
of NLP methods, information extraction techniques, and neural networks to thoroughly
analyze the trade competitiveness issues countries typically encounter, along with the
solutions policymakers and experts commonly recommend. In 2023, TradeAI initiated a
pilot to integrate generative AI into the app, enhancing its capability to offer innovative
policy solutions.
• Department/Division: Research Strategies for Export / Division of Market Development
• Project Type/Output: Online Application
• Project Domain: International Trade
• Data Source: Trade Strategy Map data (ITC’s repository of trade strategies)
• Data publicly available: No, partial datasets only
• Technology/Platform: Open Source - Python
• Related Sustainable Development Goals (SDGs): SDG 8 – decent work and economic
growth, and SDG 17 – partnership for the goals – by providing policy-relevant insights to
policymakers in the area of trade strategy.
• Relevant Links and Multimedia:
o Project site: http://ai.intracen.org/

o Blog post: https://www.linkedin.com/groups/9238012/
• Lesson Learned: Supervised machine learning can be used to accurately identify relevant
trade and SME-related content automatically. This capability significantly streamlines the
tasks traditionally performed by junior policy analysts, who previously spent days on these
activities. Now, similar (and consistently reliable) results can be achieved instantaneously
50
with the click of a button. The substantial time savings free up resources for tasks that add
more value.
• Another key insight is the ability of AI to uncover patterns of interest to policymakers,
which are often hidden within vast amounts of text data, including tables. The outcomes
ITC
of this project underscore the vital role AI plays in the design of trade and SME policies.
By automatically mining extensive quantities of trade-related text data, AI generates
invaluable policy insights and fresh perspectives on familiar issues. This shows that AI
paves the way for the formulation of more effective policies, which in turn, benefit SMEs.
Innovative services powered by AI have the potential to enhance SME competitiveness
significantly. The development of such services is critical.
• Contact Information: Alberto AMURGO PACHECO (amurgo@intracen.org)
Project 2: Corporate Strategy for AI in ITC

• Project Description: ITC has developed an organizational strategy for AI in ITC’s own
operations – with a focus on how AI tools can make the organization more productive and
innovate with new approaches to trade development.
• Project Type (Status): Ongoing
• Project Domain: International Trade and Development
• Datasets: Internal surveys
• Related Sustainable Development Goals (SDGs): SDG 8 – decent work and economic
growth
• Project Partners: ITC
• Contact Information: Alberto AMURGO PACHECO (amurgo@intracen.org)

SDG 8, 17
3. Relevant links
http://ai.intracen.org/
Contact Information: Alberto Amurgo Pacheco (amurgo@intracen.org)
51
International Telecommunication Union
Project 1: AI for Good

• Project Description: AI for Good is the leading action-oriented, global & inclusive United
Nations platform on AI. Its goal is to identify practical applications of AI to advance the
United Nations Sustainable Development Goals and scale those solutions for global
impact.
The “digital bouquet of flowers” has been arranged into three streams (Learn, Build,
Connect). For the complete list of service offerings, please visit this page. These service
offerings are available for all UN partners to play an active role in moving the needle
towards achieving the Sustainable Development Goals.
AI for Good consists of an all year online program which in 2022 broadcast over 160
webinars, and the annual in-person AI for Good Global Summit (which did not take place
from 2020 to 2022 due to Covid-19 but which will resume in 2023).
The content of the online AI for Good platform is organized into the following “Discovery
channels” (plus some other topics), which comprise of technical talks which dig deeper
into thematic areas transformed by Artificial Intelligence/Machine Learning as well as into
challenges of current AI/ML technology. Each “Discovery” episode dedicates in general
one hour to one researcher to present their latest findings in one of the topic areas below:
o AI for Earth and Sustainability Science

o AI and Climate Science
o AI for Manufacturing
o Geospatial AI (GeoAI)
o Machine Learning in 5G networks (ML5G)
o Trustworthy AI
o AI and Health
o AI for Biodiversity
o AI and Robotics
o Data-centric Machine Learning for Good
o AI & Work
The AI for Good Neural Network is an AI-powered smart matchmaking community

platform that is designed to help users build connections with innovators and experts,
link innovative ideas with social impact opportunities, and bring the community together
to discuss AI applications for social good.
Expanding on ITU’s AI for Good programme, the Neural Network offers content and
collaboration opportunities aligned to each of the 17 SDGs. Through the Neural Network,
community members can connect to each other, receive personalized content, and
pursue engagement aligned to their profiles, goals and needs.
52
The smart matching mechanism – designed according to the principles of the Global
Initiative in AI and Data Commons – connects AI innovators to anyone with an AI-related
problem, as a step towards globally scaled AI solutions. For example, it can generate
matches for open data and AI algorithms, cloud storage and computing power, problem
ITU
statements and expertise, funding and mentorships, domain transfer, SDG alignment, and
more.
The solution is meant to stimulate unprecedented cooperation across borders and
boundaries, foster impactful SDG-focused partnerships in the field of AI, and directly
serve Goal 17: Revitalize the global partnership for sustainable development.
Join the Neural Network and build your profile to enable smart matching and personalized
suggestions and discover 1000s of hours of on-demand content, networking features and
virtual exhibitions, in addition to almost daily live sessions and interactive content.
LEARN – Design your personalized programme with smart content suggestions
BUILD – Take the smart-matching quiz to meet your future AI for Good partners
CONNECT – Join our networking sessions to meet world-class AI experts
The AI for Good Neural Network is open to all with an interest in how AI can positively
impact the future of humankind. Join the AI for Good Neural Network to help build
the future of AI. 15,000 people have joined the Neural Network in just over 1 year and
continues to grow rapidly.
An onboarding tutorial video of how the Neural Network works can be viewed here:
https://www.youtube.com/watch?v=KULDHdb8xvM
• Project Type: Platform/Event/Networking/Report/Meeting
• Reported as Part of 2022 Compendium on AI Activities? Yes
• Partnership: Partnership: AI for Good is organized by ITU in partnership with 40 UN
Sister Agencies, and co-convened with Switzerland. In addition, various universities and
organizations support AI for Good.
• Project Website (links): https://aiforgood.itu.int/ ; https://aiforgood.itu.int/about-ai-for
-good/discovery ; https://aiforgood.itu.int/neural-network/
• Contacts: Mr Reinhard Scholl (reinhard.scholl@itu.int); Mr Frederic Werner (frederic.
werner@itu.int)
Project 2: UN Activities on Artificial Intelligence

• Project Description: Since 2018, ITU has issued the annual “Compendium of UN Activities
on Artificial Intelligence”, aiming to introduce activities being carried out by the UN
system. A joint-effort between ITU and 47 UN agencies and bodies, all partners of AI
for Good or members of the Interagency Working Group on AI, the 2022 version of the
report includes the collection of activity report from 40 UN agencies, providing details
on UN agencies experiments with AI to improve their response to global challenges. It
also includes additional analysis and summary to provide a comprehensive overview of
the trends and tracks within the UN system.
• Project Type: Report
• Partnership: 40 UN entities
• Project Website (links):
53
o 2018 Compendium https://www.itu.int/dms_pub/itu-s/opb/gen/S- GEN-UNACT-2018

-1-PDF-E.pdf
-1-PDF-E.pdf
-1-PDF-E.pdf
-PDF-E.pdf
-PDF-E.pdf
• Contact Information: Jin Cui (jin.cui@itu.int), Sadhvi Saran (sadhvi.saran@itu.int)
Project 3: Focus Group on Artificial Intelligence for Health (FG-AI4H)

• Project Description: The ITU-WHO Focus Group on Artificial Intelligence for Health (FG-
AI4H), driven in close collaboration by ITU and WHO, is working towards the establishment
of a framework and associated processes for the performance benchmarking of ‘AI for
Health’ algorithms. The group is working on 20+ topic areas ("use cases") addressing
health issues including breast cancer, neurodegenerative diseases, autism, vision loss,
skin lesions, cardiovascular diseases, venomous snakebites and many more. An overview
article was published in The Lancet - “WHO and ITU establish benchmarking process for
artificial intelligence in health” – a weekly peer-reviewed general medical journal which is
among the world’s oldest, most prestigious and best known general medical journals. See
also the Whitepaper for the ITU/WHO focus Group on Artificial Intelligence for Health.
• AI Approach: Framework/Strategy/Methodology Formation
• Partnership: World Health Organization
• Project Website (links): https://www.itu.int/en/ITU-T/focusgroups/ai4h
• Contact Information: Simao Campos (simao.campos@itu.int), Bastiaan Quast (bastiaan.
quast@itu.int)
Project 4: Focus Group on AI for autonomous and assisted driving (FG-AI4AD)

• Project Description: The ITU Focus Group on AI for autonomous and assisted driving
(FG-AI4AD) supports standardization activities for services and applications enabled
by AI systems in autonomous and assisted driving. FG-AI4AD studied the behavioural
evaluation of AI (when it is responsible for the dynamic driving task of a vehicle), in
accordance with the 1949 and 1968 Convention on Road Traffic of the UNECE Global
Forum for Road Safety.
To build public trust it is fundamental that the performance of AI on our road meets, or
exceeds, the performance of a competent and careful human driver. The FG aimed to
create international harmonisation on the definition of a minimal performance threshold
for these AI systems (such as AI as a Driver). This work has the potential to facilitate
adoption of AI on our roads and aims to reducing road injuries, which are already the
leading cause of death for children and young adults aged 5–29 years (more so than
HIV and tuberculosis). In fact, AI can play a significant role to reduce 1.3 million road
deaths and 25 million injuries (SDG 3.6) occurring each year, whilst also encouraging
safe, affordable, accessible and sustainable transport systems (SDG 11.2). However, the
widespread and socially acceptable deployment of AI on our roads is dependent upon
technology achieving public trust. The Focus Group has raised attention from public and
private entities and is becoming a popular forum for discussion. The Group concluded
54
its activities in September 2022. During its mandate FG-AI4AD has developed three
deliverables, which were transferred to ITU-T SG16 for further deliberation:
o “Automated driving safety data protocol – Specification”
ITU
o "Automated driving safety data protocol – Ethical and legal considerations of continual
monitoring g”
o "Automated driving safety data protocol – Practical demonstrators”
See: https://www.itu.int/en/ITU-T/ITS/standardization/Pages/sg16.aspx
The Focus Group also pioneered the discussion on what is referred to as the “The Molly
Problem”.
AI for Road Safety initiative: ITU, the UN Secretary-General’s Special Envoy for Road
Safety, and the Office of the UN Envoy on Technology launched the new AI for Road
Safety initiative in October 2021 to promote an AI-enhanced “safe system” approach to
reduce fatalities based on six pillars: road safety management, safer roads and mobility,
safer vehicles, safer road users, post-crash response, and speed control.
The AI for Road Safety initiative is in line with the UN General Assembly Resolution (UN
A/RES/74/299) on Improving global Road Safety, which highlights the role of innovative
automotive and digital technologies, as well as in line with the UN Secretary General’s
roadmap on digital cooperation. The initiative also supports achieving the UN SDG target
3.6 to halve by 2030 the number of global deaths and injuries from road traffic accidents,
and the SDG Goal 11.2 to provide access to safe, affordable, accessible and sustainable
transport systems for all by 2030.
• Project End Year: ongoing
• Project Domain: Vehicles
• Project Website (links): https://www.itu.int/en/ITU-T/ITS/AIRoadSafety/Pages/default
.aspx
• Contact Information: Stefan Polidori (stefano.polidori@itu.int)
Project 5: Focus Group on Environmental Efficiency for AI and other Emerging

Technologies (FG AI4EE)
• Project Description: The ITU Focus Group on Environmental Efficiency for AI and
other Emerging Technologies (FG AI4EE) is working to provide guidance on the
environmentally efficient operation of emerging technologies. Additionally, this project
seeks to study the impact of these technologies on the ecological/environmental
feasibility of the broader ICT ecosystem by exploring AI, increasing automation and smart
manufacturing. The group’s work also supports ITU’s ongoing research regarding the
environmental requirements of IMT-2020 (5G) systems. FG-AI4EE has worked on over
20 deliverables which cover topics related to requirements, assessment, measurement
and implementation guidelines concerning the environmental efficiency of AI and other
emerging technologies. Participation is open; there are no membership requirements.
The Group concluded its work in December 2022.
• Project Domain: Environment, Energy Efficiency
• Project Website (links): https://www.itu.int/en/ITU-T/focusgroups/ai4ee
• Contact Information: Charlyne Restivo, (charlyne.restivo@itu.int)
55
Project 6: Focus Group AI for Natural Disaster Management (FG-AI4NDM)

• Project Description: The Focus Group on AI for Natural Disaster Management (FG-
AI4NDM) capitalizes on the growing interest and novelty of AI in the field of natural
disaster management to help lay the groundwork for best practices in the use of AI for:
assisting with data collection and handling, improving modelling across spatiotemporal
scales, and providing effective communication.
• To achieve these objectives, FG-AI4NDM will develop a community of engaged
stakeholders and experts and build on past progress made by ITU in this domain.
• Participation in the Focus Group is open to all interested stakeholders
• Project Domain: Disaster Management
• Partnership: WMO and UN Environment
• Project Website (links): https://www.itu.int/en/ITU-T/focusgroups/ai4ndm
• Contact Information: Mythili Menon (mythili.menon@itu.int)
Project 7: Focus Group on AI and IoT for Digital Agriculture (FG-AI4A)

• Project Description: To address the core challenges and opportunities within the
agricultural sector, the Focus Group on Artificial Intelligence (AI) and Internet of Things
(IoT) for Digital Agriculture (FG-AI4A), explores the potential of emerging technologies
including AI and IoT in supporting data acquisition and handling, improving modelling
from a growing volume of agricultural and geospatial data, and providing effective
communication for interventions related to the optimization of agricultural production
processes. The Focus Group will also examine key concepts, and relevant gaps in current
standardization landscape related to agriculture, and will underscore the best practices
and barriers related to the use of AI and IoT-based technologies within the agricultural
domain.
To achieve these objectives, FG-AI4A cooperates closely with FAO converging multiple
stakeholders and experts from across the globe, serving as an open platform to explore
the potential of AI and IoT to support innovative practices for agricultural production
processes.
• Project Domain: Agriculture, Smart Cities, Smart Communities, Sustainable Development
• Partnership: FAO
• Project Website (links): https://www.itu.int/en/ITU-T/focusgroups/ai4a
• Contact Information: Mythili Menon (mythili.menon@itu.int)
Project 8: Focus Group on metaverse (FG-MV)

• Project Description: Recently, metaverse has become one disruptive area of innovation
with great potential to change our economy, way of living and communicating and
society. In this nascent phase of the metaverse, the industry has not converged towards
common terms and definitions. The metaverse concept has attracted considerable public
attention.
The ITU-T Focus Group on metaverse was established in December 2022. The group
analyses the technical requirements of the metaverse to identify fundamental enabling
technologies in areas from multimedia and network optimization to digital currencies,
Internet of Things, digital twins, and environmental sustainability.
56
It also provides a collaboration platform for dialogue, for identifying stakeholders

with whom ITU-T could collaborate, and for enabling the inclusion of non-members
to contribute to the technical pre-standardization work. The Focus Group work will be
enriched with the identification of relevant use cases.
ITU
To stimulate global dialogue on metaverse, a series of ITU Forum on Embracing the
metaverse will be held along with the Focus Group meetings.
• Department/Division: Telecommunication Standardization Bureau (ITU)
• Project Domain: Metaverse
• Related Sustainable Development Goals: Potentially all 17 SDGs
o FG-MV webpage: https://www.itu.int/en/ITU-T/focusgroups/mv

o Press release: https://www.itu.int/en/mediacentre/Pages/PR-2023-01-19-TSB-Focus
-Group-metaverse.aspx
o ITU metaverse activities: https://www.itu.int/metaverse/
ITU metaverse related events: https://www.itu.int/metaverse/meetings-events/
Publications on metaverse: https://www.itu.int/metaverse/publications/
• Contact information: Cristina Bueti (cristina.bueti@itu.int)
Project 9: Global Initiative on AI and Data Commons

• Project Description: The Global Initiative on AI and Data Commons is a program and
collaborative platform to support the implementation of beneficial AI based solutions to
accelerate progress towards the 2030 Sustainable Development Goals. A Roundtable
on the Global Initiative on AI and Data Commons was convened at ITU headquarters on
30-31 January 2020, attended by around 100 participants (including AI specialists, data
owners, and infrastructure providers from the private sector, academia, governments,
UN agencies and standards bodies). The roundtable highlighted the need for the Global
Initiative to maximize collaboration in order to:
o Match problem owners with providers of solutions using AI and data;

o Scale and sustain AI-based projects;
o Make available and accessible capabilities, resources, datasets, know-how, guidelines,
frameworks, standards as a common good.
• At the roundtable, two working groups (on repositories and on marketplaces) were
established and one project was identified (Global AI services platform, initially introduced
at an AI for Good Global Summit) to progress toward achieving the mission of the Global
Initiative, summarized here. On 16 July 2020, as part of the AI for Good Webinar series,
the Global Initiative launched the Global Data Pledge project to help identify, support
and make available data as a common global resource.
• The Global Initiative on AI and Data Commons is now initiating a public collaborative
effort named “Project Resilience”. The vision, in the continuity of efforts towards AI for the
common good, is to create a public AI service where a global community of innovators
and thought leaders can enhance and utilize a collection of data and AI approaches both
in the context of the current pandemic and for similar future challenges. The goal is to
collaboratively design and build an open AI system that could inform and help tackle
global decision-augmentation problems.
57

• Project Website (links): https://www.itu.int/en/ITU-T/extcoop/ai-data-commons/Pages/
default.aspx
• Contact Information: Martin Adolph (martin.adolph@itu.int)
Project 10: AI/ML Competitions (“Challenges”)

• Project Description: Since 2020, thousands of students and professionals are competing
in the ITU AI/Machine Learning Challenges. In 2023, The ITU AI/ML Challenges were
hosted in three (3) main themes: AI/ML in 5G Challenge, GeoAI Challenge, and
tinyML Challenge. Through the Challenge, ITU encourages and supports the growing
community driving the integration of AI/ML in networks and at the same time enhances
the community driving standardization work for AI/ML, creating new opportunities for
industry and academia to influence the evolution of ITU standards. Participants attempt
to address the UN Sustainable Development Goals (SDGs) related problems using real-
world data. In addition, participants will acquire hands-on experience in AI/ML in areas
relevant to solving SDGs and compete for prizes, recognition, and certificates. These
Challenges offer participants an opportunity to showcase their talent, test their concepts
on real data and real-world problems. Participants are offered free computing resources
(GPUs!) from ITU to train and optimize their machine learning models. The solutions can
be accessed in several repositories on the Challenge GitHub;
AI/ML in 5G Challenge: https://github.com/ITU-AI-ML-in-5G-Challenge.
GeoAI Challenge: https://github.com/ITU-GeoAI-Challenge
Most of the solutions submitted to the Challenge are innovative as well as improvements
with respect to the baselines. To share the solutions with the larger community, every-
year, ITU issues a call for papers for a special issue on AI and machine learning solutions
in 5G and future networks of the ITU Journal on Future and Evolving Technologies (ITU
J-FET). In this special issue, hosts (i.e., the originators of the problem statements) and
participants of the ITU Challenge submit their solutions and learnings for publication. This
special issue is dedicated to exploration of Artificial Intelligence and Machine Learning in
5G and future networks as well as enabling technologies and tools in networks.
• Project Type/Output: Datasets, Code, Papers, Standards
“AI Challenges” are competitions where a “host” defines a problem statement and
provides a dataset. Anyone in the world is invited to solve this problem statement using
machine learning.
• Project End Year: each competition has a 12-months cycle.
• Project updates: The ITU AI/Machine Learning in 5G Challenge is running for the 5th year
in 2024. GeoAI and tinyML challenges are planned to run for the 3rd time in 2024.
• Project Domain: Agriculture, Environment, Energy, Gender, Health, Poverty,
Telecommunications, and Weather
• Data Source: Data sets provided are real data, synthetic data or openly available data.
Data for the ITU AI/Machine Learning in 5G Challenge use network data, i.e., data which
occur in a communications network. Data for the GeoAI Challenge can be any data with
a location component, i.a. satellite data, user generated content (like data from social
media), sensor data or underwater data.
• Links to the challenges:
o https://aiforgood.itu.int/about/aiml-in-5g-challenge/
o https://aiforgood.itu.int/about-ai-for-good/geoai-challenge/
o https://aiforgood.itu.int/about-ai-for-good/tinyml-challenge/
58
• Data is publicly available: Yes

• Technology/Platform: To solve the problem statements, any of the above environments
or toolsets can be used.
• Related Sustainable Development Goals (SDGs): All the SDGs
ITU
• Partnerships:
o UN Partners: UN partners are UNGGIM (United Nations Global Geospatial Information

Management) Academic Network, UNGGIM Geospatial Network, UN Open GIS
(Geographic Information System) Initiatives, FAO, UNICEF, UNODC, UNGSC and IAEA
o Private Sector: see https://aiforgood.itu.int/about/aiml-in-5g-challenge/ for private
and academic partners of the ITU AI/Machine Learning in 5G Challenge
o Academia: see https://aiforgood.itu.int/about/aiml-in-5g-challenge/ for private and
academic partners of the ITU AI/Machine Learning in 5G Challenge
• Contact information: Thomas Basikolo (thomas.basikolo@itu.int); Andrea Manara for

GeoAI (andrea.manara@itu.int)
Project 11: United for Smart Sustainable Cities (U4SSC)

• Project Description: The United for Smart Sustainable Cities (U4SSC) is a global UN
initiative coordinated by ITU, UNECE, UNEP and UN-HABITAT. It currently involves 19
UN bodies. U4SSC is a global platform for smart cities stakeholders, which advocates
for public policies to encourage the use of ICT to facilitate the transition to smart
sustainable cities. The initiative aims to: Generate guidelines, policies and frameworks
for the integration of ICTs and emerging technologies into urban operations, based on
the SDGs, international standards and urban key performance indicators (KPIs); and help
streamline smart sustainable cities action plans and establish best practices with feasible
targets that urban development stakeholders are encouraged to meet. The topics of
this phase of U4SSC include: city platforms, urban economic resilience at the city-level,
procurement guidelines for smart cities, artificial intelligence in cities, metaverse in cities,
digital wellbeing and digital transformation for people-oriented cities. In the context of the
Thematic Group on AI in Cities, a deliverable is being developed to explore how AI-based
innovations be effectively deployed in the urban domain, while laying forth a framework
which encompasses the governance principles for successfully implementing AI in cities
in line with the Sustainable Development Goals (SDGs). The Thematic Group on AI in
Cities is also working on autonomous cities. The initiative delivers policy guidelines and
training materials through the work on specific outputs elaborated via regular e-meetings
and one main meeting once per year. U4SSC stakeholders also elaborated a set of Key
Performance Indicators (KPIs) for smart sustainable cities which includes 91 indicators
(core and advanced) divided in the three dimensions of sustainable development:
economy, environment, and society and culture. The indicators are fully aligned with the
Sustainable Development Goals (SDGs) and serve as a tool for evidence-based decision
making, self-assessments, progress monitoring and achieving the SDGs at the local-level.
They are being implemented by 200 cities of different sizes and development worldwide.
• Department/Division: Telecommunication Standardization Bureau (ITU), Housing and
Land Management, Forests, Land and Housing Division (UNECE), United Nations
Environment Programme (UNEP), UN-HABITAT.
• Project Type/Output: Multi-agency partnership
• Project Domain: Sustainable urban development and digital transformation
• Related Sustainable Development Goals: SDG 11 – Sustainable Cities and Communities;
SDG 17 – Partnership for the Goal
59
• Partnerships:
o UN Partners: ITU, UNECE, UN Habitat, CBD, ECLAC, FAO, UNDESA, UNDP, UNECA,
UNESCO, UNEP, UNEP-FI, UNFCCC, UNIDO, UNOP, UNU-EGOV, UN-Women,
UNWTO and WMO
o U4SSC webpage: https://u4ssc.itu.int/

o U4SSC publications: https://u4ssc.itu.int/publications/
o U4SSC KPIs project: https://u4ssc.itu.int/u4ssc-kpi/
o U4SSC KPIs publications: https://u4ssc.itu.int/u4ssc-kpis-report/
• Contact information: Cristina Bueti (cristina.bueti@itu.int), Vahid Khatami (vahid.khatami@

un.org), Gulnara Roll (gulnara.roll@un.org), Edlam Yemeru (yemeru@un.org)
Project 12: Digital Transformation Dialogues (DTD)

• Project Description: The International Telecommunication Union (ITU) is organizing the
Digital Transformation Dialogues (DTD). It offers a dynamic platform to facilitate a deeper
understanding of emerging technologies to reshape traditional processes, improve
operational efficiency and unlock new possibilities for innovation and standardization.
The Digital Transformation Dialogues seeks to address evolving themes associated with
digital transformation, foster cooperation among city stakeholders, and examine the role
of standardization within this domain. The Digital Transformation Dialogues serve as a
unique platform for highlighting the latest work and outcomes of the ITU-T Focus Groups,
Initiatives and ITU-T Study Groups. The Digital Transformation Dialogues will encompass:
o Digital Transformation Webinars

o Fireside Chats
o Ask the expert: ITU-T Standard in Focus Sessions

• Project Domain: Digital Transformation
• Related Sustainable Development Goals: All 17 SDGs
o DTD webpage: https://www.itu.int/cities/digitaltransformationdialogues/
Project 13: AI in radiocommunications

• AI could be used during the process of making and distributing television and radio
content. It is now being used to optimise workflows for broadcasting programme making,
to improve audio and visual quality evaluation, to efficiently utilize the frequency spectrum
in television and radio distribution and recently even to create new programmes by mining
archives as well as automatically targeting content to specific audiences or individuals.
For example, AI is being used for extracting content from vast archives; automatically
localising content for international distribution; and generating access services such as
captioning, audio description, text to speech and signing far faster and far more accurately
than could be achieved in the past.
AI could be used for spectrum management and radio monitoring activities.
60
For example, deploying machine learning technology for methodologies for assessing
or predicting spectrum availability; introducing big data processing and other AI
technologies in the automation of spectrum management and radio monitoring activities.
• Project Domain: Communication
ITU
• Project Website (links): https://www.itu.int/en/action/ai/emerging-radio-technologies/
Pages/default.aspx
• Contact Information: Ruoting Chang (ruoting.chang@itu.int)
Project 14: Toolkit on Digital Transformation for People-Oriented Cities and

Communities
• Project Description: The International Telecommunication Union (ITU) together with other
13 UN entities, have developed the Toolkit on Digital Transformation for People-Oriented
Cities and Communities. The Toolkit on Digital Transformation for People-Oriented Cities
and Communities is a comprehensive guide designed to help cities and communities
leverage digital technologies for sustainable development. It provides practical strategies
and tools to address challenges and harness opportunities in the digital age. 12 modules
have been developed covering the topics from Smart Sustainable City Governance to
Smart Manufacturing.
o Toolkit webpage: https://toolkit-dt4c.itu.int/
Project 15: Digital Transformation Resource Hub

• Project Description: The Digital Transformation Resource Hub collates a range of quality
publications on various digital transformation topics such as smart sustainable cities, AI,
IoT, blockchain, digital twin, metaverse, and digital transformation trends.
o Digital Transformation Resource Hub webpage: https://www.itu.int/cities/dt-resource

-hub/
61
Project 16: Artificial Intelligence Landscape Survey

• Project Description: Artificial Intelligence (AI) is introducing new ways of analyzing,
disseminating and even generating information, and reshaping entire economic sectors
of activity. In addition to new opportunities, AI is bringing significant potential risks for
policy makers and regulators to consider. Pursuant to Plenipotentiary Resolution 214
(Bucharest, 2022), ITU is seeking to gather more information about AI-related policy
and regulatory initiatives of ITU Member States and how these efforts align with their
ongoing Digital Transformation endeavors. The results of this survey will complement
other ongoing data collection initiatives (within and beyond ITU), support the assessment
of Member States readiness in the field of AI and serve as a valuable public resource for
ITU and other UN agencies in fostering meaningful partnerships to support countries in
their ongoing AI-related efforts.
• Project Type: Survey
• Reported as Part of 2022 Compendium on AI Activities? No
• Project Website (links): https://aiforgood.itu.int/ai-landscape-survey/
• Contact Information: Jin Cui (jin.cui@itu.int), Sadhvi Saran (sadhvi.saran@itu.int)
Project 17: AI Sub-Group of Early Warnings for All Initiative

• Project Description: Under the EW4All initiative, ITU and partners have created an AI
sub-group that aims to leverage AI, particularly in disaster management, warning
dissemination, and response capabilities, while acknowledging and mitigating associated
risks through careful design and implementation protocols. The AI sub-group focuses on
adding structure to the application of AI to support the attainment of the EW4All Initiative
objectives. AI is mentioned as one of the outputs of Pillar 3 in the EW4All Executive
Action Plan, launched in November 2022: “Leverage artificial intelligence to support the
development of 'client' profiles and scale the dissemination of action alerts." AI is also
featured under Pillar 1's strategic outcome seven on Innovation for Risk Knowledge;
however, developments in AI provide opportunities to transform the practice of disaster
management and support the implementation of all Pillars in the Early Warning System
value chain. The sub-group have organised a workshop at the 2024 AI for Good Global
Summit in Geneva on 31 May to demonstrate AI use cases for enhancing multi-hazard
early warning systems, fostering dialogue, collaboration, and engagement with countries
eager to pilot innovative AI applications. ITU will showcase how AI is being piloted in
the disaster connectivity map, pinpointing areas lacking electricity access, cellular,
broadband, or siren-based technologies for early warning notifications. The results from
three pilot countries (Fiji, Tonga and Vanuatu) will identify intervention needs and tailor
early warning strategies accordingly.
• Project Domain: Emergency Telecommunications, Climate Change, Disaster Risk, Early
Warnings
• AI Approach: Framework/Applications/Use Cases
• Partnership: WMO, UNDRR, IFRC, UNDP, UNFCCC, Google, Microsoft, GSMA, Group on
Earth Observations
• Project Website (links): https://www.itu.int/en/ITU-D/Emergency-Telecommunications/
Pages/AI-Sub-Group-EW4All-.aspx
• Contact Information: Vanessa Gray (vanessa.gray@itu.int)
62
Project 18: AI/ML solutions for Climate Action

• Project Description: Climate change is one of the biggest environmental challenges
affecting the planet and humanity. The consequences of climate change are resulting in
ITU
broad and deep social and economic consequences, impacting the lives and livelihoods
of billions of people. The United Nations (UN) are playing a leading role in the global
effort to fight and address climate change through a range of activities, initiatives and
programs. The combination of ICTs and AI technologies offer a promising approach to
tackle the climate crisis by reducing emissions, increasing resource use efficiency, and
building resilience to the impacts of climate change.
The project to crowdsource innovative AI solutions was supported by, the International
Atomic Energy Agency (IAEA), the UN Food and Agriculture Organization (FAO), the
UN Educational, and Scientific and Cultural Organization (UNESCO). Top solutions and
projects pitched their solutions at the 2023 United Nations Climate Change Conference
(COP28), in Dubai, United Arab Emirates, 30 November – 12 December 2023 in the
context of the Green Digital Action.• Project Start Year: 2023
• Project Domain: Climate Change
• Partnership: IAEA, FAO, UNESCO
• Project Website (links): https://aiforgood.itu.int/about-ai-for-good/aiml-solutions-for
-climate-change/
• Contact Information: Thomas Basikolo (thomas.basikolo@itu.int)
Project 19: Artificial Intelligence & Machine Learning Capacity Building

• Project Description: The ITU Telecommunication Standardization Bureau (TSB)
launched an AI capacity building project aimed at strengthening AI and ML capacities in
developing countries. This project is a part of TSB's broader efforts to close the digital
divide and enhance technological competencies globally, particularly in regions that
are technologically underserved. At the core of this project is the mentoring activities,
which pair AI experts with local practitioners in developing countries to provide training,
guidance, and knowledge exchange. These mentorship programs are tailored to address
the specific needs and challenges of each region, ensuring that the benefits of AI and ML
technologies are accessible and relevant to local contexts.
In addition, the project also contributes to the series of in-person workshops hosted by
the Science, Technology and Innovation Unit of the International Centre for Theoretical
Physics (ICTP). These workshops focus on tinyML technology—highly efficient ML models
that can run on low-power devices at the edge of the network. TinyML is particularly suited
for developing countries due to its low resource requirements and its ability to operate
independently of extensive cloud infrastructure, making advanced AI applications feasible
in remote and low-bandwidth environments. These workshops includes training on tinyML
technologies and use case relevant for SDGs. Participants from developing countries are
equipped with skills necessary in their respective communities.
The ITU's initiative not only aims to educate and train individuals in these emerging
technologies but also seeks to create a sustainable ecosystem where local developers can
continue innovating with AI and ML. Through this project, ITU supports the implementation
of practical, impactful AI solutions that can lead to significant advancements in healthcare,
agriculture, education, and other critical sectors in developing regions. The goal is to
empower these communities with the tools and knowledge necessary to harness the
potential of AI and ML for their development and prosperity.
63
• Project related links:
o https://indico.ictp.it/event/10185/overview
o https://tinyml.seas.harvard.edu/SciTinyML-23/
• Contact Information: Thomas Basikolo (thomas.basikolo@itu.int), Marco Zennaro

(mzennaro@ictp.it)
Project 20: Guidelines for staff on the Responsible use of GenAI in content creation
• Project Description: This project is about developing guidelines for ITU staff on the
responsible use of genAI in content creation. UN system peers have expressed interest in
collaborating to potentially elevate the guidelines to UN system use. Artificial intelligence
(AI) has the potential to advance global prosperity by tackling critical issues that align
with the United Nations Sustainable Development Goals (SDGs). Novel AI systems are
continuously introduced to enhance creativity, innovation and efficiency of content
creation and to elevate the overall performance of communications. Generative AI (genAI)
is a field in AI that focuses on creating content such as text, images, sound and video.
genAI systems also bring a spectrum of risks and challenges specific to content creation.
These include the potential for malevolent use of genAI in manipulating or distorting
information and images, challenges in safeguarding privacy and protecting personal
data and ensuring the integrity of the information disseminated. If not properly managed,
these risks could harm the communicator/communicating organization, users and society.
What are the guardrails to put in place and to observe for maximum benefit and minimum
harm?
• Project Type/Output: Guidelines for ITU/UN staff
• Project Domain: AI, ethics and governance
o (Adapting the principles laid out in the following document to content creation)
Principles for the Ethical Use of Artificial Intelligence in the United Nations System:
https://unsceb.org/sites/default/files/2022-09/Principles%20for%20the%20Ethical
%20Use%20of%20AI%20in%20the%20UN%20System_1.pdf
• Contact information: Monika Gehner (monika.gehner@itu.int)
Project 21: Demystifying Digital series

• Project Description: The ITU Demystifying Digital series aims to demystify technology
for a predominantly non-technical audience. It brings together ITU staff and experts with
members of the diplomatic community, facilitating a better understanding of emerging
trends and fostering discussions on their potential impact.
These informative sessions are hosted by Permanent Missions upon their request. The
topics of interest are identified by the host. So far, the ITU has successfully conducted
Demystifying Digital sessions on the Metaverse, Sustainable and Green Standards, and
Quantum, with several more planned for the upcoming months.
Past sessions include:
o “WSIS+20 & AI for Good Summit”, Permanent Mission of the Dominican Republic,
March 2024
o “How the Internet Works”, ITU and ICANN (Geneva and New York), January 2024
o “Global Digital Compact”, ITU and GDC Co-Facilitators, December 2023
64
o “Digital Skills for Sustainable Development”, Permanent Mission of Jamaica, November

2023
o “Demystifying the Citiverse and Smart Cities”, EU Mission, September 2023
o “ITU Specific Support Programmes (SIDs), Activities and Initiatives”, Commonwealth
ITU
Small State Office, July 2023
o “WRC-23”, Permanent Mission of Canada, June 2023
o “Quantum”, Permanent Mission of Denmark in Geneva, April 2023
• Project Type: Seminar/meeting

• Partnership: Permanent Mission of member states
• Project Website (links): https://www.itu.int/hub/membership/our-members/permanent
-missions-in-geneva/demystifying-digital-series/
• Contact Information: Preetam Maloor (Preetam.maloor@itu.int), Sadhvi Saran (sadhvi.
saran@itu.int)
Project 22: Learning Labs Series

• Project Description: The ITU LearningLabs initiative provides experiential hands-on
learning opportunities for ITU staff on new and emerging technologies, in partnership with
external experts, with a view to facilitating (a) use of such technologies by staff in the day-
to-day work; and (b) research and analysis relevant to ITU’s activities. LearningLabs is part
of the emerging trends observatory, which also involves: (a) analysing and showcasing
technology megatrends and potential societal impact; (b) demystifying technology for a
largely non-technical audience; and (c) developing a network of academics to serve as a
key resource for ITU thought leadership.
Past sessions include:
o “Microsoft on ChatGPT”, Microsoft, May 2023

o “Space, the next generation”, Aerospace, May 2023
o "Cybersecurity", Saudi National Cybersecurity Authority (NCA), September 2023.
o "Generative AI Enablement", Amazon Web Services (AWS), April 2024.
• Project Type: Seminar/meeting

• Partnership: Microsoft, Aerospace, Saudi National Cybersecurity Authority, Amazon Web
Services
• Contact Information: Preetam Maloor (Preetam.maloor@itu.int), Sadhvi Saran (sadhvi.
saran@itu.int), Jin Cui (jin.cui@itu.int)
Project 23: ITU Academy course on “The governance of Artificial Intelligence”

• Project Description: This course, conducted through ITU Academy training centres
by UNU-Merit, examines the complex impact of artificial intelligence on society and
emphasizes the urgent need for comprehensive governance. It covers ethical, normative,
stakeholder, and technical aspects of AI governance across four modules, providing
65
participants with practical tools and knowledge to address challenges and dilemmas in
AI implementation and regulation. This introductory course targets mainly civil servants
of national and regional administrations.
• Project Type: Training course
• Project Status: Closed
• Related Sustainable Development Goals (SDGs): SDGs 4, 9, 16 and 17.
• Partnership: UNU-Merit
• Contact Information: ituacademy@itu.int
Project 24: ITU Academy course on “Key technologies and governance of Internet
of Things, Big Data and Artificial Intelligence”
• Project Description: This course, conducted through ITU Academy training centres by
National Institute of Telecommunications (NIT), focuses on the integration of Internet of
Things (IoT), Big Data, and Artificial Intelligence (AI) in technical, business, and regulatory
contexts. It comprehensively addresses IoT technologies, standards, and policies; explores
Big Data processing techniques and architectures; and delves into AI applications in
Internet and telecom networks, concluding with a focus on Internet governance issues
related to IoT, Big Data, and AI. This course is targeted at managers, engineers and
employees from regulators, government organizations, telecommunication companies
and academia.
• Partnership: National Institute of Telecommunications (NIT)
Project 25: ITU Academy course on “Digital transformation regulation”

• Project Description: This ITU Academy course aims to enhance participants'
understanding of digital transformation policies and regulation, focusing on strategic
elements such as regulatory governance, evidence-based decision making, regulatory
sandboxes, and consumer protection. Targeted at ICT regional regulatory associations,
ICT/Telecommunications regulators, other relevant regulatory agencies and ICT policy-
makers, it included sessions on “transformative technologies and evolving regulation”.
• Partnership: EMERG and EaPeReg
66
Project 26: ITU Academy course on “Emerging technologies for resilient digital
transformation”
• Project Description: This ITU Academy course is designed to provide a deep dive
ITU
into emerging technologies and their applications, explores topics like blockchain,
AI, Distributed Ledger Technology (DLT), and digital transformation strategies. It
encompasses lectures and interactive sessions on digital identity applications in
finance and healthcare, practical AI applications in various sectors, and insights into 5G
technology. Additionally, the course includes site visits, workshops, and case studies
focusing on government digital transformation and the application of GovStack
frameworks. It is targeted at Ministries, Regulators, Digital Financial Service players from
businesses, government agencies, non-profit groups, academia that are interested in
building their capacity in emerging technologies and Enterprise Architecture.
• Partnership: NBTC
Project 27: GovStack Initiative

• Project Description: GovStack Architecture Framework guides government investment
in establishing their customized digital government public infrastructure and technology
stack, which is an integrated set of digital capabilities, e.g., digital identity, digital
payments, information exchange, registries, consent, e-signature, GIS, digital wallets,
AI models in collaboration with OSEE Project, etc. that constitute a digital infrastructure
or foundation for digital transformation and digital economy. The Stack powers the
whole of government with a set of shared digital services that any government agency
or department can use to build quickly and at a fraction of the cost of new government
digital services that are trusted, interoperable, and scalable across different sectors
without having to design, test and operate the underlying systems and infrastructure
themselves. GovStack structures its offerings around three areas:
GovSpecs: building blocks form the heart of GovStack. These standards for interoperable
software components are published open source and can be used by governments to
procure secure and reusable software. GovStack offers twelve building blocks to build
Digital Public Infrastructure.
GovTest: prototyping is key for successful digital implementations. GovStack ́s sandbox
allows to experiment and prototype new services based on the Building Block approach.
GovLearn: digital online programs train civil servants and technical personnel on the
Building Block approach. Different leaders ́ forums connect GovStack implementing
countries and foster peer-to-peer learning in the field of digital public infrastructure.
Women in GovTech Challenge and GovStack Architechs trainings provide university
credited specializations on GovStack approach.
GovStack engage with countries by identifying priority use cases on government services
to be digitialized using GovStack approach. This year GovStack contributed with use
cases to be part of the AI Hack ‘AIntuition’: Retrieval Augmented Generation (RAG) for
Public Services and Administration Tasks organized by OSEE. The models resulted from
the Hackathon will be integrated as AI building blocks in GovStack.
• Project Type: Capacity Building, Service Design and Delivery
• Project Status: On-Going
67

• Related Sustainable Development Goals (SDGs): SDGs 16.
• Partnership: GIZ, Ministry of Foreign Affairs Estonia (EE), DIAL
• Contact Information: hani.eskandar@itu.int
Project 28: ITU Open Source Programme Office (OSPO)

• Project Description: Open-source software (OSS), explainable AI algorithms, open data,
and open content (collectively referred to as “open-source technologies”) can accelerate
a country’s digital transformation and therefore, its ability to achieve the SDGs.
A part of the Open Source Programme Office work (and in collaboration with the AI for
Good), ITU BDT has launched a community challenge to promote the development of
open-source Retrieval Augmented Generation (RAG) solutions optimised for the public
sector: https://zindi.africa/competitions/retrieval-augmented-generation-rag-for-public
-services-and-administration-tasks
Public sector institutions, such as government agencies and municipalities, manage
critical services impacting citizens' lives, including administrative tasks like regulatory
compliance monitoring, procurement management, certifications issuance, and public
communication.
The efficiency of these services relies on streamlined processes. AI virtual assistants,
powered by Large Language Model (LLM) applications, can aid in this by facilitating
seamless interaction with data, providing procedural guidance, and automating
workflows.
To be effective, Generative AI applications in the public sector must deliver contextualized
responses. RAG enhances LLM-powered virtual assistants' capabilities, allowing
efficient access to information from vast document repositories. This ensures accurate
and contextually relevant responses, supporting both employees and citizens in their
interactions with public sector information.
Around 170 participants (individuals and teams) have enrolled in the challenge.
The community challenge and knowledge exchange among stakeholders in the open-
source ecosystem are anticipated to foster the development of open-source RAG
solutions tailored for the public sector. These solutions would empower public institutions
to harness AI effectively, thereby enhancing service delivery efficiency and quality for
citizens.
• Project Domain: Digital Public Infrastructure
• Partnership: UNDP
• Project Website (links): Open Source Ecosystem Enabler (itu.int)
• Contact Information: David Manset (david.manset@itu.int )
Project 29: ITU Open-Source Ecosystem Enabler (OSEE)

• Project Description: As part of the ITU Open Source Ecosystem Enabler (OSEE) project,
ITU is promoting the integration of open-source solutions and strategies into national
digital infrastructure, projects, and strategies, including open-source AI for public sector
use cases. To raise awareness of available resources and to promote dialogue and
experience exchange between a diverse range of stakeholders, a series of webinars have
been organized:
o Open-Source (generative) AI for Public Services Innovation (https://aiforgood.itu.int/

event/open-source-generative-ai-for-public-services-innovation/);
68
o Innovating Educations: Navigating Challenges in Open-Source (Generative)

AI Integration (https://aiforgood.itu.int/event/innovating-education-navigating
-challenges-in-open-source-generative-ai-integration/);
o AIntuition: Unlocking Efficiency for Public Sector with Retrieval Augmented Generation
ITU
Applications (https://aiforgood.itu.int/event/aintuition-unlocking-efficiency-for-public
-sector-with-retrieval-augmented-generation-applications/).
The webinars have attracted stakeholders from public institutions, international

organizations, NGOs, the private sector, and academia. The recordings of the webinars
are available on the AI for Good platform.
• Project Domain: Digital Public Infrastructure
• Partnership: UNDP
• Project Website (links): Open Source Ecosystem Enabler (itu.int)
• Contact Information: David Manset (david.manset@itu.int )
Project 30: Giga and AI

• Project Description: While much of the initial school data comes from government
datasets, many lack comprehensive, accurate and up-to-date locations for all the schools
within their borders. To overcome this challenge, Giga uses AI and ML algorithms that
can identify accurate school locations in satellite imagery tiles.
Here are some of the structures that help AI determine school buildings. And through
this technology, we were able to identify and map more than 29,000 schools in a handful
of countries.
o Giga aims to develop a global database of school locations, in support of our goal to
connect every school to the internet by 2030.
o We’ve developed ML models for school mapping for several countries, including
Mongolia, Sudan, Colombia, Senegal and Ghana.
o Our ML models tend to achieve model performances (F1-score) > 90%
o High-resolution Satellite Images. Global high-resolution satellite images (60 cm/px)
from Maxar made available with support from the US State Department.
o High Performance Computing Resources. High Performance Computing (HPC)
clusters with NVIDIA GPU support from Dell allows us to run computationally intensive
AI models.
o Experiment with different types of models and model architectures. Conduct
hyperparameter tuning to identify the optimal configuration settings per model.
o Evaluate model performances using an independent test set and identifying
opportunities to further improve the performance. Develop more precise locations,
e.g. using class activation maps.
o Download nationwide Maxar satellite images and run the model on each image to
generate school/non-school predictions.
o Validate the model outputs. This include re-confirming existing school locations,
identifying and fixing erroneous school locations, and discovering previously
unmapped schools.
• Project start year: 2018

• Project Domain: School connectivity (digital infrastructure)
• Partnership: ITU-UNICEF
• Related SDGs: SDG 4 & SDG 9
69
• Project Website (links): https://projectconnect.unicef.org/map

• Contact Information: Braden Phillips (braden.phillips@itu.int), Naroa Zurutuza (nzurutuza@
unicef.org)

All SDGs
3. Relevant Links
https://www.itu.int/en/action/ai/Pages/default.aspx
Contact Information
Mr Preetam Maloor, Head of Emerging Technologies Division (Preetam.maloor@itu.int)
70
Office of the United Nations

High Commissioner for Human Rights
OHCHR
Project 1: Expert seminar on artificial intelligence and the right to privacy
Human Rights Council resolution 42/15 requested UN Human Rights to organize a one-day
expert seminar to discuss how artificial intelligence, including profiling, automated decision-
making and machine-learning technologies may, without proper safeguards, affect the
enjoyment of the right to privacy. The seminar took place as a public online event over two
half-days on 27/28 May 2020. One important area of discussion were the specific challenges
for the right to privacy that the rapidly increasing use of AI brings about. The seminar also
highlighted the key role that privacy plays in safeguarding other human rights affected by
AI. It also articulated safeguards and processes that States, businesses and international
organisations are required to put in place to promote and protect the right to privacy in the
digital age.
• Project Type (Status): Full-fledged development (Framework/Strategy/Policy)

• Project Domain: Right to privacy
• AI Approach: Events
• Project Website (links): https://www.ohchr.org/EN/Issues/DigitalAge/Pages/SeminarAr
tificialIntelligence.aspx
• Contacts: Mr Scott Campbell, Senior Human Rights Officer (scott.campbell@un.org)
Project 2: Report on artificial intelligence and the right to privacy
Human Rights Council resolution 42/15 also requested UN Human Rights to analyse the
widespread use of artificial intelligence by States and businesses and their impact on the
enjoyment of the right to privacy, as well as economic, social and cultural rights. The report
calls for a ban on AI applications that are incompatible with international human rights law, and
for moratoriums to be imposed on the sale and use of high-risk AI systems, including a ban on
remote biometric systems, unless and until adequate safeguards are put in place to protect
human rights. The report recommends that States and businesses conduct human rights due
diligence throughout the entire life cycle of AI systems.
• Project Type (Status): Report (Report)

• Project Domain: Freedom of peaceful assembly, freedom of expression, right to privacy,
economic, social and cultural rights
71
• Project Website (links): www.undocs.org/A/HRC/48/31

Project 3: Report on the right to privacy in the digital age
As a follow-up to the report on artificial intelligence and the right to privacy, published in
September 2021, UN Human Rights will present a report examining recent trends and challenges
regarding the right to privacy and clarifying related human rights principles, safeguards and
best practices.

• Project Domain: Right to privacy
• Project Website (links): The report will become available here: https://www.ohchr.org/
en/privacy-in-the-digital-age
Project 4: Expert consultation and report on 1) the practical application of the Guiding
Principles on Business and Human Rights on the activities of technology companies;
and 2) technical standard-setting and human rights
Human Rights Council resolution 47/23 requested UN Human Rights to convene two expert
consultations, to discuss the relationship between human rights and technical standard-setting
processes for new and emerging digital technologies and the practical application of the
Guiding Principles on Business and Human Rights (UNGP) on the activities of technology
companies, and to submit two reports. A virtual expert consultation was held on 7 and 8 March
2022 to discuss the practical application of the Guiding Principles on Business and Human
Rights to the activities of technology companies, focusing on 1) the State duty to protect human
rights; 2) the role of the UNGPs in tech policy and regulation; 3) the corporate responsibility
to protect human rights; and 4) access to remedy. The report demonstrates the value and
practical application of the UNGPs in preventing and addressing adverse impacts on human
rights related to technology companies and provides a set of recommendations for States,
technology companies, regional and international organizations, civil society, and the United
Nations. The expert consultation and report on technical standard-setting and human rights
will be completed in 2023.
• Project Type (Status): Event and report on the practical application of the UNGPs
(concluded in 2022) / Event and report on human rights and technical standard-setting
(to be completed in 2023)
• Project Domain: Human rights
• Project Website (links): https://www.ohchr.org/en/business-and-human-rights
Project 5: Report on peaceful protests and new technologies
In its resolution 38/11, the Human Rights Council requested the United Nations High
Commissioner for Human Rights to prepare a thematic report on new technologies, including
information and communications technology (ICT), and their impact on the promotion and
72
protection of human rights in the context of assemblies, including peaceful protests. The report,
presented at the 44th session of the Human Rights Council highlights not only the character
of new digital technologies as enablers of the enjoyment of human rights but also delves
OHCHR
into issues linked to various surveillance technologies, including AI-based surveillance (such
as facial recognition) of organizers of and participants in peaceful assemblies. Among other
recommendations, it calls for a moratorium on the use of facial recognition in the context of
peaceful assemblies.

• Project Domain: Freedom of peaceful assembly, freedom of expression, right to privacy
Project 6: B-Tech Projectdeploying and using new technologies
UN Human Rights has launched the B-Tech Project which develops authoritative guidance and
resources to enhance the quality of implementation of the United National Guiding Principles
on Business and Human rights with respect to a selected number of strategic focus areas in the
technology space. It focuses on the following thematic areas, all of which touch upon important
aspects of the development, deployment and use of AI: (1) Addressing Human Rights Risks
in Business Models; (2) Human Rights Due Diligence and End-Use; (3) Accountability and
Remedy; and (4) A Smart Mix of Measures: Exploring regulatory and policy responses to human
rights challenges linked to digital technologies.

• Project t Domain: Human rights
• Project Website (links): https://www.ohchr.org/EN/Issues/Business/Pages/B- TechProject
.aspx
Project 7: Development of UN system-wide guidance on human rights diligence in

the context of developing, deploying and using new technologies
In his Roadmap for Digital Cooperation, the Secretary-General asked UN Human Rights to
develop UN System-Wide Guidance on Human Rights Due Diligence for Digital Technology
Use (A/74/821) to support all UN entities to implement and strengthen human rights due
diligence (HRDD) policies, processes and practices for the use (including development,
acquisition and sharing) of digital technologies. The guidance provides a practical introduction
to HRDD to assist each entity in developing, implementing and strengthening its HRDD for
digital technology use, as well as actions to get started and strengthen HRDD over time.
• Project Type (Status): In development, scheduled completion by end of 2022 (Guidance)

73
Project 8: United Nations Hub for Human Rights and Digital Technology
As part of the implementation of the UN Secretary-General’s Call to Action for Human Rights, UN
Human Rights launched the UN Hub for Human Rights and Digital Technology, which provides
a central repository of authoritative guidance from various UN human rights mechanisms on
the application of human rights norms to the use and governance of digital technologies,
including artificial intelligence.
• Project Type (Status): Website (website)

• Project Website (links): www.digitalhub.ohchr.org (beta version)
Project 9: Co-lead of implementation of the data protection pillar of the UN Data

Strategy
In June 2016, the Secretary-General presented the UN Data Strategy for Action by Everyone,
Everywhere. Data Protection and Privacy is one of the priority areas in the strategy. OLA, EOSG
and UN Human Rights are the co-leads of the implementation of this priority area).

• Project Domain: Human rights, Data protection
• Project Website (links): https://www.un.org/en/content/datastrategy/
Project 10: Universal Human Rights Index
The Universal Human Rights Index (UHRI) is designed to facilitate access to human rights
recommendations issued by three key pillars of the United Nations human rights protection
system: the Treaty Bodies established under the international human rights treaties as well as the
Special Procedures and the Universal Periodic Review (UPR) of the Human Rights Council. Many
of these outputs have been manually tagged for eight years. We have used this training dataset
to build a natural language classifier, using a neural network, to create recommendations for
how outputs should be classified.
• Project Type (Status): Proof of concept (Software product)

• AI Approach: Software application
• Datasets: Universal Human Rights Index
• Related Sustainable Development Goals: All SDGs
• Project Partners: HuriDocs, Danish Institute for Human Rights
• Membership or Secretariat-driven: Secretariat-driven
• Project Website (links): https://uhri.ohchr.org/en
• Resources/Skills: Natural language processing, software development
• Technology: PyTorch
• Challenges: Data quality and consistency in the tagging of training data is key. We are
redeveloping the model on the basis of improved training data, with more consistent
tagging.
74
Project 11: Digital Image Verification and Classification Project
OHCHR
In the past, human rights investigations faced challenges in gathering sufficient data, but with
the advent of portable consumer technologies the challenge has evolved. The amount of data
is not such a pressing issue, but filtering information to create useful evidence is a challenge.
The recent experience of the Commission of Enquiry on the protests in Gaza or the current
experience of the Commission of Enquiry on the Syrian Arab Republic are instructive. Both
initiatives have received huge quantities of video and image data from networks of informants,
a big challenge to authenticate, classify and analyse into useful evidence. This project works
to address this need by further developing existing open source tools, available to the human
rights ecosystem, and creating an internal instance for conducting the same analysis on
confidential information.
• Project Type (Status): Full-fledged development (Software product)

• AI Approach: Further development of an existing software application
• Datasets: DEFACTO: Image and Face Manipulation Dataset
• Related Sustainable Development Goals: All SDGs, especially SDG 16 – Peace, Justice,
And Strong Institutions
• Project Partners: Information Technologies Institute of the Centre for Research and
Technology Hellas (ITI-CERTH)
• Membership or Secretariat-driven: Secretariat-driven
• Project Website (links): https://www.invid-project.eu/
• Resources/Skills: Forensic image analysis and classification. Software development
• Technology: PyTorch
• Challenges: Video tampering detection is computationally expensive, and we are looking
for more efficient ways to perform this task.
Human Rights Council, Advisory Committee
Project 12: Report of the Advisory Committee of the Human Rights Council on New
and Emerging Digital Technologies and Human Rights (A/HRC/47/52)
Pursuant to the adoption by the Human Rights Council resolution “New and emerging digital
technologies and human rights” (A/HRC/RES/41/11) at the forty-first session, the Advisory
Committee presented a report on the impacts, opportunities, and challenges of new
technologies with regard to the promotion and protection of human rights, including mapping
of relevant existing initiatives by the United Nations (UN) and recommendations on how human
rights opportunities, challenges, and gaps arising from new technologies could be addressed
by the Human Rights Council and its special procedures and subsidiary bodies in a holistic,
balanced, and pragmatic manner. The report addresses a range of issues linked to the use
of AI and notes that AI decision-making, even when unintended, may result in discriminatory
outcomes if the decision-making is based on biased algorithms. It highlights that a rigorous
human rights due diligence of automated decision-making tools is necessary.
• Project Type (Status): Full-fledged development (Report)
75

• Project Website (links): https://www.ohchr.org/en/hr-bodies/hrc/advisory-committee/
thematic-reports
• Contacts: Mr Eric Tistounet, Chief, Human Rights Council Branch (eric.tistounet@un.org);
Mr Scott Campbell, Senior Human Rights Officer (scott.campbell@un.org)
Special Procedures of the Human Rights Council
Project 13: Human Rights Council report of the Special Rapporteur on contemporary
forms of racism, racial discrimination, xenophobia and related intolerance on Racial
discrimination and emerging digital technologies: a human rights analysis (A/
HRC/44/57)
The Special Rapporteur analyses different forms of racial discrimination in the design and use
of emerging digital technologies, such as AI, and focuses in particular on the structural and
institutional dimensions of this discrimination. She also outlines the human rights obligations
of States and the responsibility of corporations to combat this discrimination.

• Project Website (links): https://www.undocs.org/A/HRC/44/57
• Contacts: Ms Beatriz Balbin Chamorro, Chief, Special Procedures Branch (beatriz.balbin@
un.org); Mr Scott Campbell, Senior Human Rights Officer (scott.campbell@un.org)
Project 14: Human Rights Council report of the Special Rapporteur on contemporary
forms of racism, racial discrimination, xenophobia and related intolerance on racial
and xenophobic discrimination and the use of digital technologies in border and
immigration enforcement (A/HRC/48/76)
The Special Rapporteur highlights how digital technologies, including AI systems, are being
deployed to advance xenophobic and racially discriminatory ideologies which have become
prevalent. The report also links the trends in immigration surveillance whereby AI-driven
predictive models are prone to creating and reproducing racially discriminatory feedback
loops.

Project 15: Human Rights Council report of the Special Rapporteur on the rights of
persons with disabilities (A/HRC/49/52)
Pursuant to the Human Rights Council resolution 44/10, the Special Rapporteur provides a
study on artificial intelligence and the rights of persons with disabilities. The Special Rapporteur
examines the opportunities and the risks posed by artificial intelligence technology to the
enjoyment of the human rights of persons with disabilities, and emphasizes that the practical
76
benefits of artificial intelligence may be realized once their human rights risks are sufficiently
addressed.
OHCHR
Project 16: Human Rights Council report of the Special Rapporteur on extreme
poverty and human rights on the non-take-up of rights in the context of social
protection (A/HRC/50/38)
The Special Rapporteur notes that automation of benefits, by reducing administrative

complexity for potential recipients, can reduce “non-take-up”, a phenomenon where social
protection benefits often go unused even though they are designed to protect individuals
throughout their lives. At the same time, the Special Rapporteur highlights that automation
carries risks of exclusion for the most vulnerable groups, including people unregistered at birth,
undocumented migrants, individuals without a fixed address, or informal workers.

Project 17: UN General Assembly report of the Special Rapporteur on extreme

poverty and human rights on digital welfare states and human rights (A/74/493)
In the context of greater digitization of social protection and assistance systems and the
increasing role played by automated decision-making through the use of algorithms and
artificial intelligence in such systems, the Special Rapporteur warns of a digital welfare dystopia
and recommends that rather than obsessing about fraud, cost savings, sanctions, and market-
driven definitions of efficiency, the starting point should be on how welfare budgets could be
transformed through technology to ensure a higher standard of living for the vulnerable and
disadvantaged.

• Project Website (links): www.undocs.org/A/74/493
77
Human Rights Committee
Project 18: General comment No. 37, Article 21: Right of peaceful assembly
The General Comment No. 37 on the right of peaceful assembly was adopted on 23 July
2020 during the 129th online session of the Human Rights Committee. The General Comment
addresses extensively question linked to the use of digital technologies, including AI-based
tools, both by organizers of and participants in assemblies and state authorities.
• Project Type (Status): Full-fledged development (Other)

• Project Website (links): https://www.ohchr.org/EN/HRBodies/CCPR/Pages/GCArticle21
.aspx
• Contacts: Ibrahim Salama, Chief, Human Rights Treaties Branch (ibrahim.salama@un.org);
Scott Campbell, Senior Human Rights Officer (scott.campbell@un.org)
Committee on the Rights of the Child
Project 19: Committee on the Rights of the Child - General Comment No. 25:
Children’s rights in relation to the digital environment
The General Comment No. 25 on children’s rights in relation to the digital environment was
adopted on 2 March 2021 during the 131st session of the Human Rights Committee. The
General Comment lays out how States parties should implement the Convention in relation
to the digital environment and provides guidance on relevant legislative, policy and other
measures to ensure full compliance with their obligations under the Convention and the
Optional Protocols in light of the opportunities, risks, and challenges in promoting, respecting,
protecting and fulfilling all children’s rights in the digital environment.

• Project Website (links): https://www.ohchr.org/EN/HRBodies/CRC/Pages/GCChildrensRi
ghtsRelationDigitalEnvironment.aspx
Committee on the Elimination of Racial Discrimination
Project 20: Committee on the Elimination of Racial Discrimination: General comment

No. 36: preventing and combating racial profiling by law enforcement officials
The General Comment No. 36 on preventing and combating racial profiling by law enforcement
officials was adopted on 17 December 2020 during the 102nd session of the Human Rights
Committee. The General Comment focuses on algorithmic decision-making and AI in relation
to racial profiling by law enforcement officials, and has observed that the increasing use of
new technologies, including AI, has the potential to deepen racism, racial discrimination,
78
xenophobia and other forms of exclusion. focuses on algorithmic decision-making and AI in

relation to racial profiling by law enforcement officials.
OHCHR
• Project Website (links): https://www.ohchr.org/EN/HRBodies/CERD/Pages/GC36.aspx

SDGs 1, 2, 3, 7, 9, 10, 11, 13, 16 and 17
3. Relevant Links
https://www.ohchr.org/EN/pages/home.aspx
Contact Information: Mr. Scott Campbell (scott.campbell@un.org)
79
Office of the Secretary-General's Envoy on Technology
• Project: Secretary-General’s Advisory Body on Global AI Cooperation Project Description:
As part of the follow-up to the Secretary General’s Roadmap for Digital Cooperation, the
Office of the Secretary-Generals’ Envoy on Technology is coordinating the implementation
of the Secretary-General’s proposal to establish a multi-stakeholder advisory body on
global artificial intelligence cooperation. The body will provide guidance on artificial
intelligence that is trustworthy, human-rights based, safe and sustainable, and promotes
peace. The advisory body will bring a diverse group of relevant entities in the AI
landscape, including experts, scientists and academia, to address issues around inclusion,
coordination, and capacity-building by sharing and promoting best practices, as well as
exchanging views on artificial intelligence standardization and compliance efforts.
• Project Type/Output: Policy Framework
• Project Status: Development
• Project Start Year: June 2020
• Project Domain: All Project Domains, as part of overall effort to strengthen
• Data Source: Various.
• Related Sustainable Development Goals (SDGs) : SDG 1 – No Poverty; SDG 2 – Zero
Hunger; SDG 3 – Good Health and Well-Being; SDG 4 – Quality Education; SDG 5 –
Gender Equality; SDG 6 – Clean Water and Sanitation; SDG 7 – Affordable and Clean
Energy; SDG 8 – Decent Work and Economic Growth; SDG 9 – Industry, Innovation,
and Infrastructure; SDG 10 – Reduced Inequalities; SDG 11 – Sustainable Cities and
Communities; SDG 12 – Responsible Consumption and Production; SDG 13 – Climate
Action; SDG 14 – Life Below Water; SDG 15 – Life on Land; SDG 16 – Peace, Justice, and
Strong Institutions; SDG 17 – Partnerships for the Goals
o UN Partners: Various
o Private Sector: Various
o Civil Society: Various
o Academia: Various
• Links: https://www.un.org/techenvoy/content/artificial-intelligence

SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17
3. Relevant Links
https://www.un.org/techenvoy/
80
United Nations Global Pulse
UNGP
Project 1: Operational response simulation tool for epidemics in refugee and IDP
settlements
• Project Description: The spread of infectious diseases presents many challenges to
healthcare systems across the world. Given their density and available infrastructure,
refugee and internally displaced person (IDP) settlements can be particularly susceptible
to the dangers of disease spread.
Since the beginning of 2020, we have been working with public health decision-makers
to seek to understand how COVID-19 spreads in these settlements. We initially focussed
our efforts on the Cox’s Bazar settlement in Bangladesh, working with teams from UNHCR
and WHO, and have since performed modeling of other settlements around the world.
Our model simulates the movements and interactions of each individual in the settlement,
incorporating information about family structures and demographic attributes, to
understand how COVID-19 might spread under various intervention strategies and
scenarios.
During the height of the pandemic we worked on simulating the effects on public
health interventions in settlements under different scenarios and assumptions. We are
now focussing on robustly understanding plausible excess mortality rates in the Cox’s
Bazar settlement given newly available data collected by Community Health Workers. In
addition, we are developing new methods for the collection of data on contact patterns
between individuals in settlements – a crucial input to many epidemiological models, and
an important data point for understanding disease transmission routes. We are currently
working towards publishing, to the best of our knowledge, the first set of contact matrices
for a refugee settlement.
With almost 80 million forcibly displaced people in the world, we hope that this work will
inspire more modeling groups to focus on these vulnerable populations, which have been
traditionally under-served by such efforts, to ensure no one is left behind. To this end, we
have also been working with 18 different institutions, both from inside and outside the
UN system, to form a community of practice around disease modeling in refugee and
IDP settlements. This collaboration has led to a report documenting shared challenges
as well as a global call to action.
• Department/Division: Executive Office of the Secretary General
• Project Type/Output: Academic paper/Software tool/Report
• Project End Year: Ongoing
• Project Domain: Health, crisis response and humanitarian assistance
• Data Source: Census, epidemiological and survey data
81
• Publicly Available Data: Partially

• Technology/Platform: Python, Javascript, AWS
• Related SDGs: SDG 3 – Good Health and Well-being; SDG 10 – Reduced Inequality; SDG
17 – Partnership for the Goals
o UN Partners: UNHCR Innovation, WHO, UNHCR, OCHA

o Private Sector: IBM
o Academia: Durham University, University College London, London School of Hygiene
and Tropical Medicine, Massachusetts Institute of Technology, Delft University of
Technology, University of Manchester
• Lessons Learned: People living in refugee and IDP settlements are highly vulnerable
to disease spread, however, few modelling works exist which address these groups
specifically. As part of forming a community of practice around this challenge, we have
systematically documented our lessons learned as a collective in our group report. The
link to this can be found below.
• Links:
o https://www.unglobalpulse.org/microsite/epidemic-modelling-in-settlements/
o https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1009360
o https://gh.bmj.com/content/7/3/e007822
o https://epimodel.unglobalpulse.net
• Contact information: Joseph Aylett-Bullock (joseph@unglobalpulse.org)
Project 2: Using Social Media Tools to Monitor and Fight the COVID-19 Infodemic
• Project Description: This project consists of two core components. First, in partnership
with the WHO we have been conducting ongoing social media listening exercises in the
Africa region. The data is explored and analyzed with the help of a third-party platform
but we have developed a custom classifier to categorize mentions as positive or negative
from the perspective of the WHO. UN Global Pulse has produced over 40 reports to assist
WHO AFRO in monitoring its brand and understanding the conversations associated with
COVID-19 and poliovirus.
Second, in collaboration with WHO we are supporting a team from Stanford University
which aims to test interventions to reduce vaccine hesitancy among social media
users. Machine learning will be used to segment users into different vaccine hesitancy
types based on their responses to an online chatbot survey, and a contextual bandits
experiment will be used to dynamically assign personalized treatments to reduce vaccine
hesitancy according to the user type.
• Project Domain: Health, misinformation
• Data Source: Our social media listening relies on data collected from platforms such as
Twitter, blogs, and news media. The research on reducing vaccine hesitancy uses survey/
chatbot log data collected through engagement with participants on Facebook.
82
• Technology/Platform: Our social media listening exercise adopts an AI-driven approach

supported by a commercially available consumer research tool with a human in the loop
to manually validate the performance of the tool and provide contextual insights.
UNGP
The vaccine hesitancy interventions will be deployed using the Facebook platform
(Facebook ads will be used to recruit respondents, and Facebook messenger will be
used to implement the surveys and treatments).
o UN Partners: UN Global Pulse, World Health Organization

o Academia: Stanford University Golub Capital Social Impact Lab (research collaborator)
• Lessons Learned: Lessons learned from the social media listening exercise include the
importance of clearly defining project objectives, the intended use of the data, timeframe,
and a taxonomy at the start of the project in order to effectively filter the vast quantity of
social media data into a useful product. At the same time, adopting an iterative approach
was important for improving the sensitivity and specificity of the taxonomy as well as the
ability of the analysis to address the questions most important to the WHO.
While the social media analysis process has been simplified by the use of a third-party
tool, the setup, monitoring and analysis steps are still very labor intensive. Automatic tools
are not without flaws, which reinforces the importance of having a human in the loop to
validate the findings. Challenges included the “black-box” nature of the third party tool’s
algorithms and the need to define a custom classifier because the tool’s built-in sentiment
analysis failed to capture some of the nuances of interest.
• Contact information: Katherine Hoffmann Pham (katherine@unglobalpulse.org)
Project 3: A computational framework for predictive modeling of refugee and IDP

movements
• Project Description: Predicting forced displacement is an important undertaking of
many humanitarian aid agencies, which must anticipate flows in advance in order to
provide vulnerable refugees and Internally Displaced Persons (IDPs) with shelter, food,
and medical care. While there is a growing interest in using machine learning to better
anticipate future arrivals, there is little standardized knowledge on how to predict
refugee and IDP flows in practice. Researchers and humanitarian officers are confronted
with the need to make decisions about how to structure their datasets and how to fit
their problem to predictive analytics approaches, and they must choose from a variety
of modeling options. In an academic paper and an accompanying set of practitioner-
focused “modeling cards”, we attempt to facilitate a more comprehensive understanding
of this emerging field of research by providing a systematic model-agnostic framework,
adapted to the use of big data sources, for structuring the prediction problem.
• Project Type/Output: Policy Framework/Academic paper
• Project Domain: Human Rights, Forced displacement
• Related Sustainable Development Goals (SDGs): SDG 10 – Reduced Inequality; SDG 16
– Peace, Justice, and Strong Institutions
• Partnership(s)/Collaborator(s)
o UN Partners: UNHCR
83
• Links:
o http://unglobalpulse.net/predictingdisplacement/
o https://www.tandfonline.com/doi/abs/10.1080/1369183X.2022.2100546
• Lessons Learned: The field of predictive analytics for humanitarian response is still at a
nascent stage, but due to growing operational and policy interest we expect that it will
expand substantially in the coming years. In the course of preparing this framework, we
have found that relatively little is known about the structure of this prediction problem
at the high level, and that there are a number of big-picture questions for which we
lack empirical evidence, such as: how far in advance displacement can be predicted;
how much data is needed to train effective models; and whether models transfer across
borders and humanitarian settings. We hope that this framework will help to encourage
future research on these questions; facilitate comparisons between existing models; and
spark a broader discussion on best practices for predicting forced displacement.
Project 4: PulseSatellite: A collaboration tool using human-AI interaction to analyse

satellite imagery
• Project Description: Humanitarian response to natural disasters and conflicts can be
assisted by satellite image analysis. In a humanitarian context, very specific satellite image
analysis tasks must be done accurately and in a timely manner to provide operational
support. PulseSatellite is a collaborative satellite image analysis tool which leverages neural
network models that can be retrained on-the-fly and adapted to specific humanitarian
contexts and geographies. The tool grew out of a long standing collaboration with
UNOSAT which began by building an AI model for counting structures in refugee and
IDP settlements. This was then expanded to a web-based toolkit - PulseSatellite - that
can be easily adapted to other remote sensing applications and which allows for the
incorporation of models created by other users. Currently, we have three models loaded
into the system - one that allows users to map structures in refugee settlements, a roof
density detection model (e.g. for slum mapping), and a flood mapping application.
PulseSatellite is now open for use by other UN agencies.
• Department/Division : Executive Office of the Secretary General
• Project Type/Output: Academic paper/Software tool
• Project End Year: Ongoing with various UN partners on a needs-based system
• Project updates: Over the past year we have continued to work with UNOSAT to develop
and improve on the machine learning models included in the PulseSatellite tool. A
particular focus has been on rigorous testing of the flood mapping capabilities of the
tool by testing the model in a variety of settings. The model has been used to produce
the UN’s first AI generated flood maps which have been used in operational contexts,
and has now been deployed in many countries across the world.
We are also working closely with UNOSAT to develop benchmark datasets for shelter
(refugee camp) mapping, building footprint detection and damage assessment. We plan
to use these to test many of the available well trained and top-performing models, but in
the context of UN-focused datasets (e.g. with more of a Global South and development
context than many of the standard machine learning benchmarks) and make this available
as a service to the UN system.
In addition, we are working with WHO and UNOSAT, alongside the US CDC, to develop
methods for rapid estimations of population counts in refugee camps for use in public
health programming, such as inputs to computational disease models as well as for
resource planning.
84
• Project Domain: Crisis response and humanitarian assistance

• Data Source: Satellite data
UNGP
• Technology/Platform: Python, Javascript, TensorFlow, PyTorch, Keras, AWS
• Related Sustainable Development Goals (SDGs): SDG 10 Reduced Inequality; SDG 13
Climate Action; SDG 17 Partnership for the Goals
o UN Partners: UNOSAT, WHO

o Public sector: US CDC
o Academia: Durham University
• Links:
o https://www.unglobalpulse.org/microsite/pulsesatellite/
o https://ojs.aaai.org/index.php/AAAI/article/view/7101
• Lessons Learned: Operational contexts are rapidly changing, meaning that AI models
may not always perform well. Through using a human-in-the-loop approach we have
found that models can be adapted to such changing settings, however, this still requires
(sometimes significant) manual intervention from analysts. In addition, there are various
workflows which have been more tailed to AI models but which do not always follow
the workflow of satellite image analysis using our tool. In future work we will continue to
work with end users to test and help refine PulseSatellite. We will also expand into more
applications, and continue to open PulseSatellite as a tool to agencies across the UN
system.
• Contact information: Tomaz Logar (tomaz@unglobalpulse.org)
Project 5: Online radio monitoring for public health social listening and beyond
• Project Description: Radio remains the most reliable and affordable medium of accessing
and sharing information in most of the developing world. Indeed, studies have shown that
radio remains more prevalent as a means of communication in many parts of the world
than social media. Since 2019, UN Global Pulse has worked with the WHO to explore the
use of data from radio talk shows to signal early warnings of health risks and health-related
matters. We have developed a radio monitoring pipeline which can ‘listen’ to online
radio stations, transcribe the audio using machine learning speech-to-text models, and
analyse the content using a series of NLP methods for display in a frontend dashboard.
The dashboard was designed to be used by infodemic managers and decision makers
to inform public health interventions and communication strategies. However, we aim to
convert it into a generally applicable radio mining tool that can be used in a variety of
settings.
• Department/Division : Executive Office of the Secretary General
• Project Type/Output: Academic paper/Dataset/Software tool
• Project updates: This year, we have continued to provide insights from radio data to the
WHO as part of the Africa Infodemic Response Alliance. Through partnerships, we are
currently working to adapt the dashboard to conduct radio monitoring in a wide range
of additional application contexts, ranging from gender stereotypes and xenophobia
to preparedness for environmental disasters. We are also independently conducting
research on mentions of hate speech in a global sample of radio data.
85
• Project Domain: Telecommunications

• Data Source: Radio data from online stations
• Technology/Platform: Python, Spacy, Gensim, Plotly Dash, AWS
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Well-being;
SDG 10 – Reduced Inequality; SDG 17 – Partnership for the Goals
o UN Partners: WHO, UNDP Accelerator Labs, UN Women
• Links: https://www.unglobalpulse.org/2021/05/who-and-un-global-pulse-are-building-a
-social-listening-radio-tool-to-aid-the-covid-19-infodemic-response/
• Lessons Learned: One of the technical challenges encountered has been the wide
variability in the quality of speech-to-text machine learning models when applied to
speakers of the same language but with different accents. We have worked to find diverse
training datasets to fine tune existing language models and performed benchmarking
tests to find the most generalisable approach. We have also conducted user testing of
the dashboard with WHO infodemic managers, which helped shape the development
of the tool. In the future we also hope to be able to plug in our radio data to existing
WHO analysis tools such as EARS and EIOS, as well as to transfer the technology for radio
analysis to other partners for use in different contexts.
Project 6: Imagining post-Covid-19 UN: foresight for organizational realignment

and adaptation
• Project Description: Through its SG Lab Futures Initiative, UN Global Pulse fostered a
strategic foresight exercise and dialogue to frame the role of the UN post-COVID-19
through scenarios and future visioning. The activity leveraged partnerships with two
private sector entities to access AI tools as a backbone for foresight research, including
scenario building to support the UN leadership in accessing the futures and foresight
capacity.
The purpose of the exercise is to systematically analyze the driving forces and future
trends underpinned by the COVID-19 pandemic, synthesizing them into alternative
futures scenarios. The scenarios provided a framework for discussing implications for
the UN long-term, including country-level operations.
• Project Type/Output: Seminar/meeting/Strategic dialogue
• Project Domain: Strategic transformation and planning
• Data Source: 1. Horizon scanning included a combination of structured and unstructured
qualitative data collected from web scraping and uploaded documentation to customize
an internal search engine functionality. 2. Social media analysis
• Technology/Platform: IBM Watson
SDG 8 – Decent Work and Economic Growth; SDG 10 – Reduced Inequality; SDG 13 –
Climate Action; SDG 16 – Peace, Justice, and Strong Institutions;SDG 17 Partnership for
the Goals
86
o UN Partners: HLCP foresight network, UNESCO, WHO
UNGP
1. Introducing the benefits of new methodologies and mixed methods approaches to

foresight takes time and engagement with the stakeholders.
2. Blending qualitative and quantitative data in foresight analysis can help contextualize
and provide a broader understanding of long-term challenges within a system.
3. Applying AI tools developed for commercial purposes in the UN context requires
collaboration from a multi-disciplinary team from both sides.
• Contact information: Tiina Neuvonen (tiina.neuvonen@un.org)
Project 7: Understanding population movement related to COVID-19 border closures

• Project Description: UN Global Pulse and UNHCR are working to calculate and anticipate
the number of displaced persons a) that have already crossed the Brazil-Venezuela border
and b) that can potentially cross in order to understand their need for humanitarian
support and overall strengthen protection efforts. This project consists of: (i) a queue
modeling tool for simulating border crossings under different conditions, (ii) a nowcasting
effort to calculate the amount of urban population and potentially identify interest in
population movements to Brazil using big data sources, and (iii) predictive models for
forecasting future arrivals/population movements.
• Project Type/Output: Report/Interactive dashboard tool
• Project Domain: Migration
• Data Source: We draw on a range of traditional and nontraditional data sources, including:
Twitter and Facebook data; Google trends and Google mobility data; radio data; conflict
data from ACLED; data on COVID cases, deaths, and symptoms; economic indicators;
and data on arrivals from UNHCR.
• Technology/Platform: The data cleaning and modeling pipeline is primarily coded in
Python and stored in Jupyter notebooks. Some scripts have been written in R. The code
is stored on Github. The front-end dashboard has been coded with Plotly Dash.
• Related Sustainable Development Goals (SDGs): SDG 3 Good Health and Well-being and
SDG 10 Reduced Inequality
o UN Partners: UN Global Pulse, UNHCR
• Links:
o https://www.unglobalpulse.org/project/understanding-population-movement-from
-venezuela-to-brazil-related-to-covid-19-border-closures/
o https://brazil-venezuela-flows.unglobalpulse.net/
o https://m edium. com/u nhcr- innovation- service/p redicting- the- unpredictable
-preparing-for-potential-future-scenarios-1b22cd7f8da2
• Lessons Learned: One of the core challenges to prediction/forecasting/nowcasting is

that there is a large period of unusual data from the COVID pandemic. At the time of this
project, we could only make assumptions about how COVID-19 might have changed
87
population movement intentions, since we did not have reliable information on the extent
of suppressed demand and the true number of people crossing illegally. We attempted
to address these limitations by adapting our modeling approach. First, we developed
a simulation tool for arrivals which does not rely on precise information about border
crossings, but rather allows decision-makers to experiment with a variety of assumptions
about crossing volumes, the demand for shelters, and relocation capacity. Second, we
are conducting ongoing data collection to “nowcast” potential border flows in real time.
Finally, we tested a range of different arrivals forecasting models, comparing predictions
across methods and developing uncertainty estimates for each model.
Project 8: Ukraine response – Data Science Cell

• Project Description: In response to the Ukraine crisis there were many actors in the
data science space who were providing support and performing analysis to inform
the response efforts of the UN. Early in the crisis, we initiated a Data Science Cell to
bring together entities from across the UN system to coordinate and collaborate on
these topics. The cell has grown to over 60 members from over 15 institutions. The
cell also has three subgroups: population estimation (facilitated by UNHCR and IOM);
socioeconomic impacts (facilitated by the World Bank-UNHCR Joint Data Center on
Forced Displacement); and situational awareness (facilitated by UNDP). These subgroups
are currently being used as spaces for focused interagency collaboration on these topics.
• Project Type/Output: working group
• Project End Year: Ongoing based on need
• Data Source: Varied
• Publicly Available Data: Mixed
• Technology/Platform: Varied depending on use case
• Related Sustainable Development Goals (SDGs): 10 (Reduced Inequality), 16 (Peace,
Justice and Strong Institutions), 17 (Partnerships for the Goals)
o UN Partners: UNHCR, OCHA, World Bank, UNHCR-World Bank JDC, UNICEF, IOM,
WFP, UNDP, WHO, UNOCC, UNOSAT, WTO, UNFPA, IFAD
o Non-UN Partners: NRC/DFS, Oxford University, QCRI, ACAPS
• Lessons Learned: Data science is being used in a vast number of ways across the UN
system to respond to crises and humanitarian emergencies. There was and is a strong
desire to collaborate among institutions to build better systems and overcome challenges.
As a community we hope to release a report outlining our shared challenges and lessons
learned throughout our response to the Ukraine crisis, which we aim to use as a foundation
for continuing to improve the data science community’s crisis responses in the future.
• Contact information: Joseph Aylett-Bullock, Katherine Hoffmann Pham (joseph@
unglobalpulse.org, katherine@unglobalpulse.org)
Project 9: Data Science Cell subgroup – situational awareness

• Project Description: Situational awareness can be broadly defined as the “knowledge,
understanding and anticipation of a situation through monitoring and reporting of current
events, analysis and predictive assessments” (2019 DPO JOCs). As part of the data science
cell, one of the three subgroups is looking at understanding and developing data science
88
methods for improved situational awareness during a crisis, based on our experience
in responding to the Ukraine crisis. We are currently focussed on damage assessment,
conflict monitoring and understanding the affected geography in the first month and then
UNGP
the first six months of a crisis. We hope to produce a literature review of current methods,
conduct an analysis of where gaps in this literature exist, and work towards developing
methods to address these gaps. This subgroup is facilitated by UNDP, and UN Global
Pulse is a member of the group.
• Project Type/Output: working group
• Data Source: Varied
• Publicly Available Data: Mixed
• Technology/Platform: Varied depending on use case
• Related Sustainable Development Goals (SDGs): 10 (Reduced Inequality), 16 (Peace,
Justice and Strong Institutions), 17 (Partnerships for the Goals)
o UN Partners: UNDP, UNOSAT, UNICEF, UNOCC
• Contact information: UN Global Pulse focal point - Joseph Aylett-Bullock (joseph@

unglobalpulse.org); subgroup facilitator - Gaia Rigodanza (gaia.rigodanza@undp.org)
Project 10: Spotlight Initiative – Mining Gender Perceptions from Public Radio
Discussions
• Project Description: UNGP in Kampala is using its AI-powered public radio social listening
tool to mine data on perceptions around Gender Based Violence (GBV), Violence Against
Women and Girls (VAWG), Harmful Practices (HP), Sexual Reproductive Health Rights
(SRHR), and Violence Against Children (VAC) in Uganda. This data source complements
perception surveys but provides the present state of affairs and the evolving context
of violence against women and girls in the community to allow for timely and targeted
interventions. Following engagements with the Recipient UN Organizations (RUNOs), the
team of data analysts at UNGP in Kampala registered the following achievements:
o Maintained the radio flow from 22 radio stations broadcast in Luganda covering
Kampala and its surrounding districts.
o Defined a detailed taxonomy of keywords to mine relevant data and defined a tagging
taxonomy (capturing perpetrators, victims, and violence types).

• Project Type/Output: Dashboard/Report
• Project updates: Biweekly dashboard updates are produced by UNGP and shared with
stakeholders to inform their programming.
• Project Domain: Telecommunications
• Data Source: Radio data from FM stations
89
• Technology/Platform: Python, Bash, AWS, My SQL, PHP, HTML, Ruby, C++

• Related Sustainable Development Goals (SDGs): SDG 5 – Gender Equality and SDG 10
– Reduced Inequality
o WHO, MGLSD, UNICEF, UN RCO, UN WOMEN, UNFPA and Ministry of Gender

Labour and Social Development in Uganda
• Links: https://www.unglobalpulse.org/project/ending-violence-against-women-and-girls
-in-uganda/
• Lessons Learned: Radio remains the most popular source of information in Uganda,
used by different people irrespective of their demographics. For example, according
to the Spotlight Initiative’s baseline report in 2020 the majority of respondents (70.8%
of women and 73.3% of men) received information on VAWG/HPs and SRHR through
radio. Radio shows allow two-way communication between radio studios and people in
the community through call-ins, making them a rich representation of the people that are
digitally marginalized.
Project 11: Developing Ethical AI Frameworks and Data Exchanges for Uganda and
Ghana
• Project Description: AI and other emerging technologies present opportunities for
the achievement of national priorities and to increase national and regional welfare.
To realize these opportunities, and to safeguard against risks and potential harms, key
foundations need to be put in place, such as access to data (a.o., to train AI), and the
development and use of emerging technologies need to be guided by ethical rules and
principles. UNGP has been supporting the Government of Ghana and the Government
of Uganda on developing solutions to address these two issues. This support has led to
the development, in each country, of a Roadmap for an Ethical AI Framework informed by
the local social and societal context, and a Roadmap for the development of Community-
Centric Data Exchanges at the national level.
• Project Type/Output: Ethical AI Framework, and community-centric Data Exchange
Roadmap to unlock data as a resource for the attainment of national and regional
priorities (each for both Uganda and Ghana).
• Project updates: The recommendations made in the Ethical AI Framework and the data
exchange roadmap are being used to inform the follow-up work in developing a data
strategy for Uganda and implementing a use case of the data exchanges focusing on the
tourism sector. Both projects are underway.
• Domain: Digital technologies
Growth; SDG 12 – Responsible consumption and Production; SDG9 – Industry, Innovation
and Infrastructure(each for both Uganda and Ghana), SDG11 – Sustainable Cities and
Communities
• Partnership
o Government: Government of Uganda (Ministry of ICT and National Governance)
90

SDG 3, 8, 10, 13, 16, and 17
UNGP
3. Relevant links
https://www.unglobalpulse.org/
Contact information
Rene Clausen Nielsen (rene.nielsen@un.org) and Miguel Luengo Oroz (miguel@unglobalpulse

.org)
91
United Nations Human Settlements Programme
Project 1: AI in Cities: Risks, Applications and Governance

• Project Description: UN-Habitat has collaborated with Mila to produce a policy paper
on the application of artificial intelligence in urban contexts. The paper AI in Cities:
Risks, Applications and Governance, takes a risk-based approach highlighting the risks
of AI throughout its lifecycle and application in cities. The report first describes AI, its
different types, the opportunities it offers for cities, and its current limitations. It identifies
key sectors for intervention for cities, along with examples of AI applications within each
of those sectors. The report then presents a detailed risk assessment framework with
reflective guiding questions. Finally, the report offers a set of recommendations and areas
of action to consider when developing an AI strategy.
• Department/Division: Knowledge & Innovation Branch
• Project Type/Output: White paper
• Project Status: Published
• Project Domain: Human Rights, Environment
• Related Sustainable Development Goals (SDGs): SDG 11 – Sustainable Cities and
Communities
o Civil Society: MILA
• Related Sustainable Development Goals: SDG 11

• Contact Information
o Edlam Abera Yemeru (yemeru@un.org) and Livia Schaeffer Nonose (livia.

schaeffernonose@un.org)
Project 2: International Guidelines on People-Centred Smart Cities

• Project description: Member states adopted the resolution mandating UN-Habitat to
develop International Guidelines on People-Centered Smart Cities in June 2023 at the
UN-Habitat Assembly in its second session. The process will culminate in the development
of International guidelines on people-centered smart cities which will address key areas of
technology application in cities requiring international guidance, due to normative gaps.
Among the areas to be tackled, AI and new technologies are specifically mentioned as
in need to be regulated by international non-binding frameworks that can be used by
92
national and local policy-makers to ensure cities and technologies used are really built in
a people centered fashion.
UN-Habitat
Strong emphasis is also put on the importance to preserve citizens right to privacy both
online and offline, while building the capacity of policy makers to ensure there is enough
understanding about the risks and opportunities related to the new technologies. In
leading on the development of the guidelines, UN-Habitat will build on existing work
and data collected through the AI need assessment, the global smart cities governance
study and others, and is currently organizing global, regional, thematic and stakeholder
consultations as the basis of the guidelines development process.
• Project Type/Output: International guidelines
Communities

Project 3: AI for people-centered smart cities workstream in the UN inter-agency

working group on AI
• Project Description: In early 2022, a new workstream on AI for people-centered smart
cities, co-chaired by UN-Habitat ITU, UNESCO, and UNDP, was established in the UN’s
Inter-agency working group on AI. The objectives of the workstream are to understand the
needs of cities and the capabilities of UN System organizations in relation to AI, and how
to leverage AI for people-centered smart cities while promoting ethical guidelines for AI
that are human-rights based, strengthening inter-agency cooperation, and implementing
key performance indicators and other metrics to ensure that these values are respected.
Currently, the project is kicking off a global assessment of responsible AI use in cities
to map out the opportunities and constraints for local governments to responsibly use,
implement and govern AI technologies in cities.
• Project Type/Output: Reports
Communities
o UN Agencies: UNDP, ITU, UNESCO

o Civil Society: IDRC, UNU-E-GOV, U4SSC, CC4DR
93

Project 4: Global Assessment of Responsible AI in Cities

• Project Type/Output: Reports
• Project Domain: Human Rights, Environment, Strategies, Capacity Building
Communities
o UNU-E-GOV
o Civil Society: IDRC - International Development Research Centre


Project 5: BEAM: Building & Establishment Automated Mapper: Mapping informal

settlement in eThekwini, South Africa
• Project Description: The city of eThekwini has the largest number of informal settlements of
any municipality in South Africa. To keep track of the fast changes in the built environment
and develop a pipeline of public service delivery and upgrading projects, it is essential
that the city has access to evidence-based information. That requires up-to-date records,
particularly on the scale, location, and number of informal areas.
To assist the eThekwini's Human Settlement Unit (HSU) in automating their building
mapping process, United Nations Technology Accelerator for Cities (UNITAC) developed
BEAM, a model that uses machine learning to radically accelerate the spatial recognition
of settlements and building structures on aerial imagery. BEAM can mark pixels specific
to a building's spectral profile with an accuracy of 94%. It looks to provide HSU with up-
to-date geo-referenced base maps.
In addition to the tool development, there will be a technology and knowledge transfer
component. This capacity-building process aims to enable the eThekwini Municipal
Government to leverage BEAM to more effectively target their upgrading and basic urban
service delivery interventions. A project advisory group has been established to address
the BEAM's potential risks, challenges, and opportunities.
• Department/Division: UNITAC Hamburg (linked to UN-Habitat’s Flagship II Program:
People Centered Smart City)
• Project Type/Output: Application Tool & Manual; Academic Paper; Project Advisory
Group; External Communications, including blog posts and video content
• Project Domain: https://unitac.un.org
94
• Data Source:
o Population demographics: 580 urban informal settlements encompassing 314,000
UN-Habitat
household
o Geography: South Africa
o Locations: eThekwini
• Link to data: Aerial photography is provided by the city

• Data publicly available: Not yet
Communities, SDG 9 – Industry Innovation and Infrastructure, SDG 10 Reduce Inequalities,
and SDG 17 – Partnership for the Goals
o UNITAC partnership between the United Nations Human Settlements Programme (UN-
Habitat), the United Nations Office for Information and Communication Technology
(OICT), and HafenCity University
o Human Settlement Department, City of eThekwini, South Africa
o Using Artificial Intelligence for better city planning

o Leveraging local expertise to enhance urban sustainability with BEAM
• Lesson Learned:
Perception of AI in the Field of Informality
Mapping a community inherently makes it “legible” to public authorities. Doing so can
have tremendous advantages for the planning and allocation of critical public health
and infrastructure services, especially in slum settlements. However, such data collection
must respect both the privacy of residents and their unique cultural dynamics. Similar
initiatives need to ensure that all developed tools adhere to United Nations Human Rights
guidelines. The legal agreements with project partners, including MoUs and NDAs, need
to have clauses on human rights and social safeguards that denote language, such as
that the data and results produced by the tool will not negatively impact people on the
ground. It is vital to integrate within the project objective, the understanding that such
tools are only to be used towards the upholding of rights and freedoms of all.
Without a doubt, such initiatives should be communicated openly and transparently to
the public. A Community of Practice or Project Advisory Committee should be established
to bring diverse perspectives on project development and implementation. At the same
time, communication for #tech4good initiatives needs to be considerate of public
perception and biases toward AI solutions, so that there are no misinterpretations of
project objectives
Knowledge Transfer and Training Sessions can be Difficult
The tool was co-designed with project partners. However, knowledge transfer and training
sessions can become difficult in the long run. Long-term hosting and maintenance of the
tool require ongoing resources for cloud computing and HR. The cost may exceed the
project’s lifetime. There is a risk that this tool cannot be ensured in the long run.
The Model needs to be Re-trained for New Cities
A key limitation for project upscaling is that the tool needs to be re-trained for each
city. In addition to limited personnel, another challenge is access to geospatial images.
BEAM currently relies on aerial photography. However, not every city has access to or
can afford frequent aerial images. An alternative will be the use of satellite imagery. Not
only are satellite images available on numerous platforms, but using them will also be
95
advantageous during the project upscaling phase, where some cities may not have the
capacity to produce aerial imagery.
BEAM tool is currently being piloted in 8 cities in Central America.
Unfortunately, high-resolution (ideally 1m Ground Sampling Distance or below), freely-
accessible satellite images are not always easily attainable. As such, it is worthwhile to
consider the possibility of merging multiple sources and types of geospatial images.
Livia Schaeffer Nonose (livia.schaeffernonose@un.org) and Sophie Naue (sophie.naue@
un.org)
Project 6: Hola ISUD plan (Integrated Strategic Urban Development Plan for Hola
Town, Tana River County, Kenya).
• Project description: UN-Habitat in collaboration with UNEP (UN Environment Programme)
is committed to supporting the preparation and elaboration of the Hola ISUD, with a
strong link between urbanization, biodiversity, and ecosystem services. The preparation
of Hola ISUDP will form part of a larger territorial strategy: The Go Blue project, an
Innovative Land-Sea Planning and Management for Sustainable and Resilient Kenyan
Coast, which is a partnership between the EU (European Union) and the Government of
Kenya to advance the blue economy agenda through coastal development. For the Hola
ISUDP, AI was primarily used in the remote sensing process of extraction of land cover
information from satellite imagery. Land cover is a key layer necessary to perform most
of the GIS-based multicriteria analyses, helping to analyze a territory to detect historical
changes and trends, particularly in a poor data environment, in which we mostly rely on
information extraction from satellite imagery. The land cover layer was obtained through
a supervised process of machine learning (considered as a subcategory of AI) applied
to Copernicus Sentinel-2 imagery with 10 m resolution, by which the software is trained
to recognize the different land cover classes (built up, tree cover, shrubland, agriculture,
water, etc.). The creation of a high-quality land cover layer was critical to producing
different analyses related to natural hazards like flooding vulnerability, drought potential,
or land degradation, to take into consideration, among different other factors, the impact
or influence of each type of cover on the specific hazard. Moreover, it was important for
environmental assessments as biodiversity condition land cover change and ecosystem
service value, as well as for multicriteria analyses aimed to assess the agriculture or urban
development suitability. Therefore, we used two key open-source datasets originated
through AI processes, to fill the data gap and to support different analyses and maps.
The first one is the high-resolution density map provided by Facebook data for good in
collaboration with CIESIN (Center for International Earth Science Information Network)
of Columbia University. The map estimates the population at 30 x 30 m resolution for
different categories (youth, elder, women, men). The data are generated by distributing
the aggregate population data from the available census on building footprint and
density extracted using advanced AI methodology. The data were used in the project
to produce a population density map, to estimate the population affected by natural
hazards, the amount of population served or unserved by public facilities, and to support
strategic scenarios of growth and development. The second one is the open buildings
dataset provided by Google, which produced a large dataset of 516 M buildings covering
64% of Africa extracted from high-resolution imagery in 2021, by applying AI. We used
the buildings layer for different applications from the production of topographic maps
to city scale and neighborhood strategic plans.
• Department/Division: UN-Habitat, Global Solution Division, Planning Finance and
Economic Section (PFES)
• Project type/output: The main objective is to prepare the Integrated Strategic
Urban Development Plan For Hola Town. The ISUDP is a document that will define
a comprehensive long-term vision of the Municipality and guide sustainable urban
development for the upcoming ten years.
96
• Project status: Completed

UN-Habitat
• Project end year: 2023
• Project domain: Urban planning and design, spatial analysis
• Data source: Google earth engine https://developers.google.com/earth-engine/datasets/
Open buildings https://developers.google.com/earth-engine/datasets/catalog/GOOGLE
_Research_open-buildings_v1_polygons
High-resolution density map: https://dataforgood.facebook.com/dfg/docs/methodology
-high-resolution-population-density-maps
• Technology/Platform: ArcGis, Google Earth Engine
• Related SDGs: SDG 3 – Good Health and Well-Being; SDG 11 – Sustainable Cities and
Communities; SDG 13 – Climate Action
• Contact information: Edoardo Forzano (edoardo.forzano@un.org)

SDG 3, 9, 10, 11, 13, 17
3. Relevant Links
https://unhabitat.org/
Contact Information
Edlam Abera Yemeru (yemeru@un.org) and Livia Schaeffer Nonose (livia.schaeffernonose@

un.org)
97
United Nations World Tourism Organization
UN Tourism, the United Nations World Tourism Organization, works with Artificial Intelligence
principally through providing opportunities for startups that are either primarily involved in this
technology or utilise it in their solutions. Spanning several startup competitions powered by UN
Tourism, up to 60 top solutions from 35 countries are engaged in emerging technologies directly
relevant or adjacent to AI. They are being supported by the UN Tourism Innovation Network
through connection to Member States and investors as well as training and development
support. By doing so, the Organization enables the digitalization of the tourism sector and
facilitates broader adoption of technologies worldwide for the benefit of local communities
and the sector stakeholders.
Project 1: UN Tourism – Women in Tech Startup Competition Middle East 1st Edition
Countries across the Middle East have been integrating innovation and women empowerment
within their long-term economic development plans. With women across the region becoming
more interested in STEM degrees and launching startups, they need support and access to
growth and funding opportunities. In this regard, the Women in Tech Startup Competition
offers a platform for innovation opportunities in tourism for new entrepreneurs.
Categories:
• Tourism & Travel Experience (food and beverage, transportation and accommodation,
travel, and retail industries)
• Social Impact (technology in wellness, health, urban development, rural development,
sustainability, and education)
• Events & Community (social and messaging platforms, e-sports and gaming platforms,
community-based applications, events-related platforms, and education technology)
• Future Tech (AI, AR/VR, crypto currency, internet of things, blockchain technology, digital
twining and the metaverse)
This competition closed with a total of 143 applicants from 11 countries in the Middle East. All
the winning startups of the competition were based off AI-solutions:
• Sparkle Haze (UAE): Award-winning voice AI assistant speaker for luxury hospitality.
98
• Plastus Biotech (KSA): Patented tech converts organic waste into 100% biodegradable
bioplastic.
• Fosha (Egypt): AI-powered app for travel service bookings.
UNWTO
• Xenios Academy (UAE): AI powered EdTech startup with gamified training programs for
cafes, restaurants, and hotels to bridge skill gaps.
Official website: https://www.unwto.org/startup-competition/unwto-women-in-tech-middle

-east
Project 2: UN Tourism - Awake Tourism Challenge
Following a careful evaluation process, 15 winners out of 2,000 participants from 120 countries
were released. The majority of applications came from Africa, Europe and the Americas, and
mainly from the Tourism Tech for Good and Local Community Involvement categories. Top
reade-to-implement solutions have joined the UN Tourism Innovation Network and look
forward to engaging projects all over the world.
Official website: https://www.unwto.org/unwto-awake-tourism-challenge
Project 3: UN Tourism Online Academy
UN Tourism Online Academy continues to grow now offering 40+ courses from renowned
universities and now introducing more content on technologies and their application to daily
tourism and hospitality operations. Women account for 57% of the total of students.
Official website: https://www.unwto-tourismacademy.ie.edu/
Project 4: ITU Focus Group on Metaverse and the UN Tourism Task Group on
Metaverse Tourism
UN Tourism is vice-chair of the ITU Focus Group on Metaverse and leads the Task Group on
Metaverse Tourism which attempts to compile a list of possible use cases and applications
of technology in the tourism sector and to analyse whether there is a need for standards or
guidelines. Recommendations are expected as outputs of the expert discussions.
Official website: https://www.itu.int/en/ITU-T/focusgroups/mv/Pages/default.aspx
Project 5: Other UN Tourism Startup Competitions and Innovation Challenges
UN Tourism has developed additional programmes which identify startups with use cases of
AI in different segments of the value-chain, tourism products, and destination types.
• Event Tourism Startup Competition: https://www.unwto.org/startup-competition/mega

-events-and-mice-tourism
• Startup Competition for Island Destinations: https://www.unwto.org/unwto-tourism-tech
-adventures-canarias-2023
• Hospitality Challenge: https://www.unwto.org/startup-competition-tourism-technologies
-solutions-in-hotels-new-business-models
• 4th Global Gastronomy Tourism Startup Competition: https://www.unwto.org/startup
-competition/4th-global-gastronomy-tourism-startup-competition
99

SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17
3. Relevant Link
https://www.unwto.org/
Contact Information
Natalia Bayona, Executive Director (nbayona@unwto.org)
Glenn Cauwenberghs, Senior Specialist – Innovation, Education and Investments Department

(gcauwenberghs@unwto.org)
José Diaz Ardila, Senior Specialist – Innovation, Education and Investments Department (jdiaz@
unwto.org)
100
United Nations Women
UNWOMEN
Project 1: Action Coalition for Technology and Innovation for Gender Equality
• Project Description: With the rapid digitalization of work, school and social life stimulated
by the COVID-19 pandemic, the importance of technology and innovation to achieving
gender equality and inclusive development have never been clearer nor more urgent.
From a gender perspective, specific applications of AI and machine learning have shown
the greatest risks of bias and misuse, like facial recognition and deep fakes. The AI world
today is almost entirely dominated by men and we find societal biases relating to gender
roles and identities embedded in social programs and services via automated decision-
making. Data modelling such as predictive policing or social intervention increasingly
transcends the individual to focus on groups or communities, making women more at
risks of being discriminated.
UN Women’s role and mandate is to reaffirm the need to focus on diversity and
inclusiveness when developing AI technologies. Ensuring that societal values are reflected
in algorithms and AI technologies will require no less creativity, hard work and innovation
than developing the AI technologies themselves.
In order to drive action and unite efforts from across governments, private sector and
civil society, UN Women is convening partners working on gender and technology
as part of the Generation Equality Forum. An Action Coalition focusing on innovation
and technology will be launched in 2021 to generate innovative ideas for policies and
initiatives needed to accelerate progress for more gender-responsive AI.
The Action Coalition will explore how to harness opportunities arising from the use of AI
and overcome the challenges associated with algorithms. By breaking down silos and
fostering collaboration through this new multi-stakeholder platform, the Coalition aims
to inspire public and private partners to make strong and actionable commitments that
will advance gender equality and women’s rights.
• Project Update: The Generation Equality Forum, held in Mexico and Paris in 2021,
resulted in the launch of the Action Coalition for Technology and Innovation for Gender
Equality, to ensure women and girls have equal opportunities to safely and meaningfully
access, use, lead, and design technology but also to ensure that building inclusive digital
economies is at the core of the COVID-19 recovery efforts.
This initiative is open to world leaders and grassroots leaders equally, which makes
it a unique space for co-creating innovative solutions and for generating bold and
transformative commitments in the next five years.
The Coalition’s Global Acceleration Plan recommend stakeholders to join efforts on four
priorities:
o First, to prioritize initiatives that support digital access and the development of digital
skills, to ensure equal use, creation and control of digital technology by women and
girls.
101
o Second, to invest in feminist technology and innovation, to embed gender in tech

development and create solutions that leverage technology for social impact.
o Third, to build inclusive, transformative and accountable innovation ecosystems that
ensure women and girls’ full participation in digital economy and society.
o And finally, to prevent and eliminate online and tech-facilitated GBV and discrimination
to allow women and girls in all their diversity to use digital spaces safely. More
than a thousand commitments and 40 billion dollars were pledged for Generation
Equality, many originating from governments, civil society, international organizations,
youth networks and companies engaged in building more inclusive and gender-
transformative technology and innovation ecosystems. This resulted in commitments
to improve regulatory frameworks, to systematize gender impact assessments and AI
audits, to improve the availability of disaggregated data and to bridge the massive
gender gap that currently exists in tech professions – and Artificial Intelligence in
particular.
In 2024, the collective mobilization of the Action Coalition led to the launch of the
Position Paper “ Placing Gender Equality at the Heart of the Global Digital Compact”
which includes recommendations on AI governance.
• Project Type/Output: Multi-stakeholder partnership

• End Year: 2026
• Project Domain: Gender
• Related SDGs: SDG 4 – Quality Education, SDG 5 – Gender Equality; SDG 9 – Industry,
Innovation, and Infrastructure; SDG 17 – Partnerships for the Goals
o UN Partners: ITU, UNICEF

o Government: Chile, Rwanda, Armenia, Tunisia, Finland
o Private Sector: Salesforce, Microsoft, Koc Holding
o Civil Society: A+ Alliance, Social Builder, Global Fund for Women, Digital Grassroots
o https://techforgenerationequality.org/
o https://forum.generationequality.org/action-coalitions
o https://www.unwomen.org/en/get-involved/beijing-plus-25/generation-equality
-forum
• Contact information: Helene Molinier (helene.molinier@unwomen.org)
Project 2: Women, Peace and Cybersecurity: Promoting Women´s Peace and Security
in the Digital World
• Project Description: UN Women ROAP has been conducting research in 2023 on the
implications of AI, including leveraging AI, for the implementation of the Women, Peace
and Security agenda in Southeast Asia , in partnership with United Nations University
Institute (UNU) in Macau. The research has been conducted under a broader project on
digital security in the context of the WPS agenda, “Women, Peace and Cybersecurity:
Promoting Women´s Peace and Security in the Digital World”, in partnership with the
Governments of the Republic of Korea and Australia. The project aims to ensure that
women meaningfully participate in cybersecurity decision-making processes and online
peacebuilding efforts, in the prevention of cybersecurity threats (including disinformation
and hate speech) and cybercrimes, and contribute to reforming cybersecurity laws,
102
policies and norms, where emerging technologies such as AI plays an important

part. In addition to developing research relating to AI, UN Women is also developing
UNWOMEN
e-Learning and other training material to further disseminate the research findings using
an interactive methodology. These materials are being developed in close collaboration
with key regional civil society partners and digital rights experts.
The learnings from the research will continue to inform ROAP’s WPS programming and
peacebuilding work in the coming years, and will be shared with UN Women COs & HQ,
other ROAP teams and the UN system.
• Project Status: The research report is currently going through the final stages of internal
review and will likely be launched in May 2024. The broader project under which the
aforementioned activities are conducted will conclude on 30 June 2024.
• Project Sector: Peace and Security
• Related Sustainable Development Goals (SDGs): SDG 16; SDG 5
Project 3: Andi-the Chatbot

• Project Description: UN Women MCO Caribbean is supporting a registered charity in
Trinidad and Tobago - the Coalition Against Domestic Violence (CADV), to pilot an AI
online platform – “Andi the Chatbot”, that will provide immediate information to connect
victims and survivors of domestic violence and the public to protective and essential
services in the country. The chatbot is intended to augment human support with its 24/7
availability, in addition to a locally operated helpline “800-SAVE”. The project will be
implemented by three collaborating institutions - the Coalition against Domestic Violence,
The Shelter and The Andrea Project, all of whom are working with vulnerable communities
affected by violence.
• Project Status: Andi-the Chatbot, will be launched in March 2024 and UN Women MCO
Caribbean will be providing technical support for the development of the Chatbot as well
as supporting a portion of the services required for its operationalization.
• Project Sector: Social services
• Related Sustainable Development Goals (SDGs): SDG5

SDG 4, 5, 9 and 17
3. Relevant Link
https://www.unwomen.org/en
Contact Information:
Helene Molinier (helene.molinier@unwomen.org); Adam Simpson (adam.simpson@unwomen

.org)
103
The Joint United Nations Programme on HIV/AIDS
Project: Global AIDS Monitoring and use of AI

• Project Description: Currently, the Global AIDS Monitoring (GAM) data collection is done
through an online reporting platform, and the data are published in the annual Global
AIDS Report and on AIDSinfo data visualization platform. Much of this work includes
manual processing: data entry (country rapporteurs), data management and organization,
validation and presentation/visualization (strategic information department, SID).
While the new UNAIDS strategy and targets outline far more detailed and broader data
framework, it is anticipated that the burden on collecting, validating and analyzing the
data is likely to increase both for country rapporteurs and UNAIDS. It is necessary to
explore new ways of automating the data collation from publicly accessible sources and
platforms, reducing the reporting burden on countries. Furthermore, with the increase
in data, iterative yet fast cross-analysis and learning are required to identify issues such
as inequities in services and affected communities, as well as the role of societal enablers
affecting peoples access to the services they need.
Collecting data via GAM online reporting tool is currently the primary data acquisition
mechanism for GAM. The data submitted through the platform, by the authorities from
different countries, have to be inspected by experts of UNAIDS and partner organizations
to ensure the correctness of information reported in GAM. Often the direct data collection
is complemented by analysis of existing reports and journal articles. This requires
searching and reading the requested pieces of information from lengthy documents
within many different repositories. An intelligent search functionality implemented with
state-of-the-art NLP AI models becomes an invaluable assistant.
The following AI-functionalities are required for GAM processes: prefilling GAM online
reporting tool fields with specific automatically retrieved information; helping UNAIDS
experts to quickly find the relevant information from auxiliary sources to validate the
reported data; extracting the tables from various PDF documents to UNAIDS internal
repository to enable efficient search with predefined keywords.
• UN Entity Name: UNAIDS
• Entity Name: UN Joint Programme on HIV and AIDS
• Department/Division: Strategic Information Department
• Project Type/Output: Dataset, The AI is intended to be used to improve data validation
process and to enrich the dataset on AIDS.
• Projected End Year: 2022
• Project updates: Project has started in September 2021 with an inception phase where
the case scenarios and the applied methodologies are described, and the AI summarizer
104
content is defined. The inceptions phase IDs completed by end October and the next
step is to implement it for the selected datasets. Several pilot runs of AI searches to fill
specific data gaps. The searches returned many results but nonewere useful for real-time
UNAIDS
tracking of the HIV epidemic. As per the results of these results, it was deemed that certain
technology was not fit for addressing the needs of data validation.
• Data Source: Aggregate data on health and HIV related issues.
• Link to data: https://aidsinfo.unaids.org
• Technology/Platform: Apache Solr or Elastic Search, coupled with a language model, e.g.
BERT or GPT-2. Other technologies applicable for retrieving information from documents
are AI-models for text (OCR, optical character recognition) extraction and table extraction.
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Well-Being ;
SDG 10 – Reduced Inequalities
• Contact information: Sonia Arias Garcia (ariasgarcias@unaids.org)

SDG 3, 10
3. Relevant Links
https://www.unaids.org/en
Contact Information
Ali Feizzadeh (feizzadeha@unaids.org)
105
United Nations Conference on Trade and Development
Project 1: Data for development

• Project Description: At a time when digital connectivity defines our lives, the importance
of data and AI in shaping the trajectory of global development cannot be understated. To
harness the benefits of AI and address its associated risks, robust governance mechanisms
are essential for the international community. Preparations are currently underway
for the 2024 United Nations Summit of the Future. These preparations, along with
negotiations for the Global Digital Compact (GDC), aim at accelerating the development
of international frameworks for AI governance. In parallel, existing international initiatives
like the World Summit on Information Society (WSIS) are under review to assess their
successes and shortcomings in governing digital technologies. Other regional initiatives
like G7 Hiroshima Process, EU AI Act, and OECD AI principles have recently taken shape
with a promise of regulatory spill-over effects across the globe. In the core of these AI
policy efforts are data and data governance.
At its twenty-six session held in March 2023, the United Nations Commission on Science
and Technology for Development (CSTD) selected “Data for Development” as one
of its priority themes for the 2023–2024 intersessional period. The secretariat of the
Commission convened an intersessional panel in Lisbon from 6 to 7 November 2023
to better understand this theme and assist the Commission in its deliberations at its
twenty-seventh session. The theme “Data for Development” will also be discussed at the
forthcoming Annual Session of CSTD from 15 to 19 April 2024. The UNCTAD publication
‘Data for Development’ will be released in late April 2024. This publication will provide
comprehensive insights into the potential benefits and challenges associated with data
for development and data governance. The publication is based on the CSTD panel’s
findings and recommendations, country case studies contributed by CSTD member
states, international organizations and other stakeholders.
While data-enabled technologies offer tremendous opportunities ranging from scientific
advancement to the creation of new business models, improper management of data and
associated technologies can exacerbate digital divides, create market monopolies, and
pose threats to human rights.
In response to these challenges, and to advance data governance for development,
consistent with the imperatives of multilateralism, multistakeholder approach and
multidisciplinary consideration of data, the UNCTAD publication ‘Data for Development’
and the associated report of the United Nations Secretary-General proposed the following
seven data governance principles:
(1) foundation in human rights;

(2) treating data in context;
(3) balancing risks and innovation;
(4) people empowerment;
106
(5) multilayered approach in data governance;

(6) multi-stakeholder inclusivity;
UNCTAD
(7) youth inclusion for future orientation.
• Department/Division: Technology, Innovation and Knowledge Development Branch/

Division of Technology and Logistics
• Project Type/Output: Publication
• Project Status: To be completed
• Project Updates: Project to be completed.
• Project Domain: Telecommunications; Inequalities; Sustainable development;
Technological gaps
• Data Source: Contributions from member states and the UN System.
• Publicly available data: Yes.
Growth; SDG 9 – Industry, Innovation, and Infrastructure; SDG 10 – Reduced Inequalities;
SDG 17 – Partnerships for the Goals.
• Links and Multimedia: https://unctad.org/system/files/information-document/CSTD2023
-2024_Issues01_data_en.pdf;
• Contact information: Mr. Angel González Sanz, Chief, Technology, Innovation and
Knowledge Development Branch/Division of Technology and Logistics, Division on
Technology and Logistics (angel.gonzalez-sanz@un.org)
Project 2: Panel discussion on “Protecting information integrity in the age of Artificial

Intelligence (AI)” at Twenty-sixth annual session of the United Nations Commission
on Science and Technology for Development (CSTD), 29 March 2023
• Project Description: Purposeful information manipulation (disinformation) conducted
by malicious or unethical actors has been recognised globally as a major threat to public
safety and security, as well as democratic stability of nations. At the same time, the
tendency for online censorship and surveillance to be used as a way to address real or
presumed information manipulation threats puts in danger fundamental human rights,
especially, freedom of expression online, while appropriate measures should be taken
to avoid illegal content on the internet.
This challenge is now aggravated by the maturing of AI technologies. Generative AI, e.g.
large language models made popular by Chat GPT, voice generation and voice cloning
and other technologies, can now be used by malicious and unethical actors to conduct
information manipulation campaigns. The omnipresence of deep-fakes and loss of trust in
information in the public space that the new AI capabilities portend can make restricting
online freedom seem a necessary price to pay for protecting stability of societies.
However, both information disorder and suppression of online freedom are avoidable. A
holistic approach focusing on information integrity and relying on multifaceted policies
can support a free, open, safe and secure online environment resilient to the negative
impacts of information manipulation, even in the new technological environment. Just as
international cooperation is necessary to find new solutions to the challenges presented
by AI, it is equally important to share more broadly the already accumulated knowledge
and best practices on addressing information manipulation.
The event aims to provide solutions on protecting information integrity by inviting
speakers to address the following questions: What are the implications of the evolving
AI capabilities for the vision of an online public space that is open, free, but also safe and
secure? How can governments and societies protect information integrity and address
107
AI-powered information manipulation with effective, rights compatible policies? What is

the role of the existing, “traditional” policy tools against information manipulation in the
new technological environment?
• Project Type/Output: Meeting
• Project Updates: Project was completed.
Technological gaps
• Data Source: N/A.
• Publicly available data: N/A.
SDG 17 – Partnerships for the Goals
• Links and Multimedia: https://unctad.org/meeting/26th-cstd-side-event-panel-discussion
-protecting-information-integrity-age-artificial;
• Contact information: Mr. Angel González Sanz, Chief, Technology, Innovation and
Knowledge Development Branch/Division of Technology and Logistics, Division on
Project 3: Side-event on “The multifaceted implications of AI” at Twenty-sixth

annual session of the United Nations Commission on Science and Technology for
Development (CSTD), 29 March 2023
• Project Description: To provide a brief overview of the potential multifaceted implications
of the application of AI and possible solutions among policy makers, academia, private
sector, civil society and international organizations. In the absence of an effective
international normative framework, the future of AI is evolving without much input from
most developing countries. Currently, most AI frameworks and regulation are developed
by the Global North or self-regulation by digital platforms that control the data used
to feed AI models and algorithms. So, what are the real alternatives? The event invited
speakers to address the following questions: The power of AI systems is growing very
fast. Knowledge and creative jobs were thought to be less exposed to AI disruption? Is
this still true? Will any occupation escape the impact of AI? Are our societies ready to deal
with the of AI, for jobs, for human rights, for civic and political participation, for ethics?
Can education help make AI more likely to enhance workers’ abilities rather than replace
them? What needs to change in education policies and practices for this to happen? Given
that AI does not currently reflect most of the world, is it ethical or equitable to allow it to
have an increasing role in managing global systems? Is there a riks that AI development
will escape the regulatory power of governments? Can an open innovation approach
for AI offer an alternative model which is more amenable to the ethical and equitable
imperatives that matter for developing countries? How shall the rights issue in relation to
AI and emerging technologies be addressed? How can international cooperation help
solve these issues?
• Project Type/Output: Meeting
108

UNCTAD
• Project Updates: Project was completed.
Technological gaps
• Data Source: N/A.
• Publicly available data: N/A.
• Links and Multimedia: https://unctad.org/system/files/information-document/Concepte
_note%26Programme_AI_26CSTD_side_event29Mar23PM.pdf
• Contact Information: Mr. Angel González Sanz, Chief, Technology, Innovation and
Knowledge Development Branch/Division of Technology and Logistics , Division on
2. Relevant Sustainable Development Goals

SDGs 8, 9, 10 and 17
3. Relevant Links
www.unctad.org
Contact Information
Mr. Angel González Sanz, Chief, Technology, Innovation and Knowledge Development
Branch/Division of Technology and Logistics , Division on Technology and Logistics (angel.

gonzalez-sanz@un.org)
109
United Nations Department of Economic and Social Affairs
Project 1: Use of Complex Network Mappings Development for Digital Government

• Project Description: The United Nations E-Government Survey is produced every two
years and is a flagship publication of UN DESA. It presents the ranking of e-government
development across 193 United Nations Member States by assessing e-government
development according to a quantitative composite index based on Online Service,
Telecommunication Infrastructure, and Human Capital Indices. It provides decision makers
with information that enables them to identify their areas of strengths and challenges, as
well as suggested options on how best to move ahead.
In 2022, UN DESA hired a consultant to conduct a pilot study using the science of complex
systems to expand the analysis of factors affecting countries’ e-government development
beyond income level and test a complex network analysis model to address possible
inequalities and biases adherent to rankings and find as yet unidentified similarities
and differences between the Member States. The following section provides details on
methodology of complex networks model used for the UN DESA pilot study conducted
by Roberto Bellotti, Professor in Applied Physics and Director of the Physics Department
of the University of Bari, Italy. More detailed information on the study and its findings is
available on UNDESA Egovknowledge base.
The data set used for the analysis consisted of 305 World Development Indicators
(WDIs) relating to health, economy, society and environment and 214 SDG indicators
characterizing the general development level of each Member State. The SDG indicators
have acquired a decisive role in the characterization of the 193 UN Member States as they
allow to represent the general conditions for development of each nation in a multifaceted
way, providing additional and complementary information to the WDIs, especially in
strategic areas relevant to EGDI: access to electricity, mobile networks coverage, and the
number of fixed internet broadband subscriptions, to name a few. Selection of indicators
followed the criteria of data availability, consistency, and non-redundancy. The reference
year for data is 2020, with missing values filled up with data from 2019 and 2018 to
represent a snapshot of the current situation.
• Entity Name: Digital government Branch.
• Department/Division: DESA – Division for Public Institutions and Digital Government
• Project Type/Output: Report/Academic paper/Software tool/Seminar/meeting
• Project End Year: Continuous
• Project Domain: This is a cross-cutting project covering most fields, as it is about effective
Governments and Public Institutions, Digital transformational and Digital Development
for sustainable development.
110
• Data Source:
o World Development indicators (WDI) – World Bank Group
UNDESA
o E-government Development Index (EGDI) – UN-DESA
o Link to data: https://publicadministration.un.org/egovkb/en-us/Data-Center; https://
databank.worldbank.org/source/world-development-indicators

• Technology/Platform: Python: libraries on machine learning and complex network analysis
Growth, 9 – Industry, Innovation and Infrastructure, 10 – Reduced Inequalities, and 16 –
Peace, Justice and Strong Institutions
o Academia: University of Bari – Italy - Department of Physic Scientific Partner
• Links: https://publicadministration.un.org/en/Research/UN-e-Government-Surveys;
https://www.nature.com/articles/s41598-020-74964-3 -Scientific publication as starting
point for this pilot project.
• Lessons Learned: Grouping the countries into four development clusters using complex
network analysis enables to reinterpret the level of e-government development of the UN
Member States and their E-Government Development Index (EGDI) ranking considering
their starting conditions, as well as their mutual similarities and differences. By comparing
the countries EGDI values both within the same cluster, and between different clusters
it is possible to identify top-of the-class countries, whose performance goes beyond the
expectations based on their development status, and room-for-improvement countries,
that have the potential to reach their cluster peers in the EGDI ranking by increasing their
efforts. For top-of-the-class countries their EGDI values are above the 75th percentile of
the cluster they belong to, and, at the same time, they are above the 25th percentile of
at least one development cluster above. By the same token, for room-for-improvement
countries the EGDI values are below the 25th percentile of the cluster they belong to and,
at the same time, are below the 75th percentile of at least one developed cluster below.
The study also identified the benchmark countries, regarded as the best cases compared
to the rest of the world, and trailing countries, which would need specific support to
improve their condition in areas relevant for EDGI ranking. Benchmark countries are
characterized by an EGDI values above the 75th percentile of the distribution within
cluster I, while trailing countries have EGDI values falling below the 25th percentile of
the distribution within cluster IV. l.
• Contact information: Vincenzo Aquaro (aquaro@un.org) & Deniz Susar (susar@un.org)
Project 2: Fast-evolving technologies in e-government: Government Platforms,

Artificial Intelligence and people
Chapter 5 - The Future of Digital Government: Trends, Insights and Conclusions, within
the United Nations E-Government Survey 2022 discusses evolving technologies and new
approaches in digital government. It covers cognitive government, agile and adaptive
government, seamless government.
• Project Type (Status): Other (Other)

• Project Domain: UN E-Government Survey 2022
• AI Approach: Publication
• Related SDGs: All SDGs, specifically SDG 16 Peace, Justice and Strong Institutions
• Membership or Secretariat-driven: Both
• Project Website (links): https://publicadministration.un.org/egovkb
111
• Resources/Skills: The need for data is nothing new but the ways in which data are created
and used have changed dramatically in recent years, bolstered by the revolution in data
technologies and the proliferation of applications of different types and forms of data,
including small and big data, real-time data and geospatial data. The current COVID-19
pandemic also reinforces the centrality of data -- how governments and businesses
handle data, as it turns out, is a crucial part of their pandemic response. Learn more about
open government data development, policy and institutional trends on government data
sharing, exchange and interoperability, as well as data security, privacy and ethics; and
recommendations on national data leadership and data governance framework.
• Technology: Publication
• Challenges: Developing indicators to measure how AI is used in public administration
• Contacts: Vincenzo Aquaro (aquaro@un.org) & Deniz Susar (susar@un.org)
Project 3: National and Local Surveys of Digital Government Branch
To prepare for the production of United Nations E-Government Surveys, the Digital Government
Branch conducts a national survey called Member States Questionnaire (MSQ) at the national
level, and a city level survey called Local Government Questionnaire (LGQ) at city level. Both
surveys include questions on artificial intelligence;
National Level
• Does the national e-government strategy make specific reference to the use of new/
emerging technologies such as artificial Intelligence (AI), robotics, blockchains, 5G and
Internet of Things (IoT)? *
Local Level
• Do you have a specific city/municipality strategy on the following new technologies?

(e.g., Artificial Intelligence, IoT, Blockchain, Smart City, 5G, Virtual/Augmented Reality,
Robotics, 3D Printing)
• Does your city/municipality strategy plans for the use of big data analytics in their decision-
making processes? Please explain further and provide links.
• Project Type (Status): Other (Other)
• Project Domain: UN E-Government Survey 2022
• AI Approach: Publication
• Related SDGs: All SDGs, specifically SDG 16 Peace, Justice and Strong Institutions
• Membership or Secretariat-driven: Both
• Project Website (links): https://publicadministration.un.org/egovkb
• Technology: Publication / Data
• Challenges: Developing indicators to measure how AI is used in public administration
• Contacts: Vincenzo Aquaro (aquaro@un.org) & Deniz Susar (susar@un.org)
Project 4: TFM findings on the impacts of rapid technological change on the SDGs
New and rapidly changing technologies, such as artificial intelligence, robotics and other
automation technologies hold great promise for making accelerated progress towards the
Sustainable Development Goals, but also pose formidable challenges in all of the SDG
dimensions. Against this background, the UN General Assembly has called upon the TFM in
112
repeated resolutions to present their updated findings to the Annual Multi-stakeholder Forum
on Science, Technology and Innovation for the SDGs.
UNDESA
• Project Type (Status): Research/Study paper (Recurring event)
• Project Domain: New technologies, including AI
• AI Approach/Activity: Findings are crowdsourced from TFM partners and scientific and
technological communities, through calls for inputs (policy briefs and research papers),
leveraging institutional networks, university partnerships and meetings. In particular, a
series of UN expert group meetings on AI since 2016 have provided a convergent series
of general policy recommendations, upon which recommendations for specific issues
elaborate.
• Datasets: Scientific data provided by contributors and volunteers. Database under
development
• Related SDGs: SDG 17 Partnerships for the Goals
• Project Partners: DESA, IATT, 10-Member Group, TFM partners, Universities
• Project Website (links): https://sustainabledevelopment.un.org/tfm
• Resources/Skills: Mainly volunteer work; knowledge of technologies developments,
sustainable development models and pathways.
• Technology: UN platform
• Challenges: Key challenges have been the vast scope of the exercise, limited resources,
and large expectations. However, a series of lessons-learnt have been identified and
important support provided to various reports. Present work in 2020 focuses on the
environmental impacts of AI.
• Contacts: Mr Richard A Roehrl, Senior Economic Affairs Officer (roehrl@un.org)
Project 5: Exploring the impacts of new Internet applications and AI on the global
energy system
New Internet applications and especially AI technologies have become a rapidly increasing
source of energy demand but have also greatly shaped the opportunities for smart and
cleaner energy systems. This project reviews what is known and what might be potential policy
responses to these trends in the future.
• Project Type (Status): Research/Study paper (Ongoing)

• Project Domain: AI and Energy
• AI Approach/Activity: Desk study and expert surveys
• Datasets: (Under development)
• Related SDGs: SDG 7 Affordable and Clean Energy, SDG 9 Industry, Innovation and
Infrastructure, SDG 17 Partnerships for the Goals
• Project Partners: DESA, IATT, 10-Member Group, Other experts
• Project Website (links): https://sustainabledevelopment.un.org/tfm
• Resources/Skills: Expert knowledge, volunteer work, and scientific networking skills.
UNDESA 38 United Nations Activities on Artificial Intelligence (AI)
• Challenges: A key challenge has been the identification of work that exists in
fragmented forms in various disciplines and both in academia and private sector. Hence,
interdisciplinary expert surveys are key to their identification. Furthermore, a common
technical terminology is needed.
113
Project 6: IEEE/UN Event series
Training and outreach event on technology, policy, ethics and engagement of AI and other
new technologies.
• Project Type (Status): Event (Concept note)

• Project Domain: New technologies, including AI: technology, policy, ethics and
engagement
• AI Approach/Activity: Webinar series
• Datasets: IEEE datasets
• Project Partners: DESA, IEEE, TFM partners
• Resources/Skills: Expert knowledge.
Project 7: Long-term AI and technology scenarios for the SDGs
Long-term technology scenarios are routinely used to explore feasible technology pathways
to tackle big global challenges, such as climate change and biodiversity. While an increasing
number of them assume significant new opportunities due to AI, most of them do not make
any effort to quantify these effects in both positive and negative terms. This initiative aims to
explicitly account for AI and potential future AI technology developments based on existing
technology development data. It also provides inputs for the mandated discussions of long-
term future scenarios and the impact of current trends in the high-level segment of ECOSOC
each year.
• Project Type (Status): Research/Study paper (Recurring event)

• Project Domain: AI scenarios
• AI Approach/Activity: Scenario analysis
• Datasets: Various scientific and technological data sources
• Related SDGs: SDG 7 Affordable and Clean Energy, SDG 13 Climate Action, 17
• Project Partners: DESA, TFM partners
• Project Website (links): https://undocs.org/e/2020/60
• Resources/Skills: Scenario analysis, technology change, AI techs
• Technology: Various scenario models
• Challenges: A challenge is the linking to the national level and national level AI scenarios
(where they exist)
Project 8: Guidebook on AI ethics for government and development practitioners
While there are hundreds of publications and proposed AI ethics frameworks and codes of
conduct by scientific and engineering communities, as well as an UNESCO initiative on AI
ethics, little practical United Nations Activities on Artificial Intelligence (AI) 39 guidance exists
for governments and development practitioners, especially guidance that is fully based on
a balanced scientific and technological understanding. The guidebook aims to fill this gap.
• Project Type (Status): Report (Ongoing)

• Project Domain: AI ethics
114
• AI Approach/Activity: Collaborative product developed by academics working on AI

ethics with practical experience
• Related SDGs: SDG 9 Affordable and Clean Energy, SDG 16 Peace, Justice and Strong
UNDESA
Institutions, SDG 17 Partnerships for the Goals
• Project Partners: DESA, TFM partners
• Challenges: A key challenge is the translation of technical specificities into practical, easy
understandable guidance for practitioners.
Project 9: TFM online platform
The TFM online platform was mandated to provide a single-entry point for technology
information.

• Project Domain: Online platform for information on technologies and SDG knowledge
• AI Approach/Activity: Gateway to networks of curated SDG-related technologies and
knowledge from UN and non-UN resources
• Related SDGs: All SDGs
• Project Partners: DESA, OICT, UNCTAD, 10-Member Group, and an increasing number
of other partners (see website)
• Project Website (links): https://tfm2030connect.un.org/
• Challenges: Further development of content and operational support work.
• Contacts: Ms Stephanie Rambler, Sustainable Development Officer (rambler@un.org)
Project 10: Guidebook to resources on AI strategies (supplement to the IATT

Guidebook on STI roadmaps for the SDGs)
While there is an increasing number of AI strategies and an exponentially increasing number

of publications on AI, government officials and development practitioners alike could benefit
from a trusted, curated and annotated list of written resources on the various aspects of AI.
• Project Type (Status): Report (Ongoing)

• Project Domain: Curated listing of AI publications
• AI Approach/Activity: Curated and annotated list of publications on the various aspects
of AI, in support of STI roadmaps for the SDGs
• Project Partners: DESA, IATT
• Challenges: Curation and selection of most important publications and other resources
on the various aspects of AI strategies.
• Contacts: Mr Wei Liu, Sustainable Development Officer (liuw@un.org)
Project 11: FAO-UNSD project using satellite data and farm surveys to estimate crop
statistics
The project aims to identify crops, map crop areas and estimate crop yield using satellite data
and farm surveys.
• Project Type (Status): Software product (Development)

115
• AI Approach/Activity: Supervised Machine Learning uses random forests and support

vector machines – Datasets: Satellite data and Farm surveys
• Related SDGs: SDG target 2.4
• Project Partners: FAO
• Membership/Secretariat-driven: Driven by UNSD and FAO
• Resources/Skills: We work with highly skilled data scientists, statisticians and computer
engineers of international and national statistical agencies.
• Technology: UN Global Platform, https://marketplace.officialstatistics.org/earth
-observation
• Challenges: While providing service through a Cloud-based environment, the biggest
challenge is still making the tools and applications useful to national statistical agencies
in developing countries by lowering the entry level of required technological knowledge.
• Contacts: Mr Ronald Jansen, Chief of Data Innovation and Capacity Branch, Statistics
Division (jansen1@un.org)
Project 12: Estimating Port Calls using AIS vessel tracking data
The project aims to identify ships which are entering and leaving a port (by vessel type) using
AIS vessel tracking data AIS data are real-time data of ship positioning. This is obtained as a
global feed.
• Project Type (Status): Software product (Development)

• Project Domain: Maritime Transport
• AI Approach/Activity: Supervised Machine Learning uses random forests to estimate
vessel types and carrying capacities
• Datasets: AIS vessel tracking data (https://c omtrade. un. org/d ata/c ache/
AISdashboardMethodology.pdf)
• Related SDGs: SDG target 9.1
• Project Partners: UNCTAD, University of Oxford, ONS (UK)
• Membership/Secretariat-driven: Driven by UNSD – Project Website (links): https://
marketplace.officialstatistics.org/ais-weekly-port-calls;
• Resources/Skills: We work with highly skilled data scientists, statisticians and computer
engineers of international and national statistical agencies, and research institutes.
• Challenges: UN Global Platform, see https://marketplace.officialstatistics.org/ais-data.
• Contacts: Mr Markie Muryawan, Chief of Trade Statistics Section, Statistics Division
(Muryawan@un.org, +1 212 963 3083) United Nations Activities on Artificial Intelligence
(AI) 41
Project 13: LinkedSDG
A demo app that automatically extracts key concepts related to sustainable development from
text documents and links them to the most relevant sustainable development goals, targets,
indicators and series.
• Project Type (Status): Software product (Full-fledged development)

• Project Domain: SDG ontology – AI Approach/Activity: This uses Semantic Web
technologies and ontologies, which is a subfield of AI and Computer Science research
• Datasets: Sustainable Development Goals Taxonomy (http://metadata.un.org/sdg/?lang
=en)
• Related SDGs: All SDGs
• Project Partners: DESA - Division for Sustainable Development Goals
116
• Membership/Secretariat-driven: Driven by DESA – Project Website (links): http://linkedsdg

.apps.officialstatistics.org/#/; https://sustainabledevelopment.un.org/LinkedSDGs/about
• Resources/Skills: Statisticians and computer engineers of DESA
UNDESA
• Technology: UN Global Platform
• Contacts: Mr Luis Gonzalez Morales, Chief of Web Development and Data Visualization
Section, Statistics Division (gonzalezmorales@un.org, +1 212 963 0692)

All SDGs
3. Relevant Links
https://www.un.org/development/desa/en/
Contact information
Vincenzo Aquaro (aquaro@un.org)
117
United Nations Development Programme
Project 1: Increasing AI capability

• Project Description: From big data to machine learning, digital technologies are
transforming our world. In mid- 2019, UNDP launched its firstDigital Strategy to
enhance its support to governments in adapting to this rapidly-changing environment
- including by building digital capacity within the Organization. The strategy seeks to
increase understanding of digital technologies and how they can be used to achieve the
Sustainable Development Goals, as well as the risks and trade-offs that come with them.
AI is a key element of the strategy, as a tool for equitable and accessible digital responses
and UNDP has been nurturing AI projects as well as investing in the data foundations
needed to harness its potential. In parallel, UNDP is working on ways to manage the
ethical challenges that arise from deploying AI in an international development context.
This work is coordinated by the Chief Digital Office, in close collaboration with alll UNDP
bureaux and offices.
There are currently several evolving projects with a focus on increasing AI ability internally
and externally, as well as assessing AI readiness of member states and safeguarding
internal products.
AI readiness tool for governments: Building on UNDP’s Digital Readiness Assessment,
CDO has developed a readiness assessment focused specifically on AI - The Artificial
Intelligence Readiness Assessment (AIRA), which was successfully launched in Rwanda,
Colombia, and Sri Lanka in 2023. .The assessment comprises a comprehensive set of
tools that allow governments to get an overview of the AI landscape and assess their
level of AI readiness across various sectors. The framework is focused on the dual roles
of governments as 1) facilitators of technological advancement and 2) users of AI in the
public sector. Critically, this assessment also prioritises ethical considerations surrounding
AI use. The assessment highlights key elements necessary for the development and
implementation of ethical AI, including policies, infrastructure and skills and is suitable to
all AI maturity levels. These aspects are important for countries to consider as AI-powered
technologies are implemented at population scale to help meet national priorities and
achieve the Sustainable Development Goals.
• Establishment of the UNDP AI Working Group: To navigate the transformative impact of AI
on UNDP's internal operations and member state support, the UNDP AI Working Group
was officially formed in December 2023 with the key role of overseeing responsible AI use,
streamlining internal applications, and developing approaches for legal compliance, risk
management, and transparency. The Working Group is led by the Secretariat, consisting
of the Chief Digital Officer (CDO), the Chief Technology Officer (ITM) and the Deputy
Assistant Administrator and Deputy Director (GPN).
• Department/Division: Chief Digital Office
• Project Type/Output: Digital Readiness Assessment tool, Working Group
118
• Project Start Year: 2021, 2023 respectively

• Data Source: open and closed government data, survey data, UNDP project data
• Reported as part of 2022 Compendium on UN AI Activities? The AI Readiness tool - yes,
UNDP
the rest - no
• Related Sustainable Development Goals (SDGs): SDG 5 - Gender Equality, SDG 8 -
Decent Work and Economic Growth, SDG 10 - Reduced inequalities, SDG 16 – Peace,
Justice, and Strong Institutions, 17 - Partnerships for the Goals
o Government: AI readiness assessment will be made available to all governments.

Rwanda, Colombia, and Sri Lanka governments were partners in 2023.
o www.digitalstrategy.undp.org
o https://feature.undp.org/making-ai-work-for-us/
o https://www.undp.org/blog/are-countries-ready-ai-how-they-can-ensure-ethical-and
-responsible-adoption
• Contact information: Gayan Peiris (gayan.peiris@undp.org)
Project 2: SURGE Data Hub

• Project Description: The SDH helps UNDP Country Offices provide governments with
human-centered data for evidence-based decision-making that ensures no one is left
behind in crisis response and recovery. The Hub ensures decision-makers have access to
evidence and insights that reveal the true cost and impact that fragility and shocks have
on people’s lives.
SDH was created in 2020 building on previous individual UNDP initiatives such as the
HBDA and the Digital Socio-Economic Impact Assessment (SEIA) tools, to provide
a comprehensive, structured, and institutionalized solution in UNDP (and partners)
for end-to-end digital assessments in crisis. The Crisis Bureau developed these tools,
making use of available technologies and building on the existing assessment offer in
the humanitarian and development sector.
SDH seeks to facilitate rapid end-to-end digital assessments remotely for Country Offices
in crisis all over the globe—in time zones that are convenient and in languages they
understand. Since its inception we have supported over 45 countries across five regions in
partnership with national counterparts to carry out digital assessments for decisions about
recovery programming (e.g., debris and waste management, emergency employment
etc.) and advanced data analysis to address underlying causes of vulnerability and fragility.
SDH is powered through the UNDP Crisis Bureau Country Support Management Team
(CSMT). To roll out the tools in a fast-paced and effective fashion, CSMT established a
Remote Support Unit (RSU) that provides targeted technical support with local language
and context knowledge on planning and implementing digital assessments. It includes
experts in Information Management, Economics, Statistics, Data analysis, and Systems.
In addition, SDH has developed a training programme to strengthen UNDP practitioners’
skills in digital assessments to enhance COs analytical capabilities. We seek to invest in
people and nurture a community of 800+ practitioners to generate internal transformation
and bring about change with national counterparts through a corporate culture of data
and evidence-based decision-making.
To uphold our shared ambition under Agenda 2030, SDH offers a bold new approach
to knowing what’s going on in the world through data, ensuring vulnerable populations
aren’t overlooked in policy-making decisions after crisis.
• Department/Division: Crisis Bureau
• Project Type/Output: Report; Data set; Software tool
119

• (Projected) End Year: Undefined
• Project Domain: Crisis
• Data Source: Primary data collection through digital assessments (and surveys).
o HBDA collects digital and georeferenced data on damages, for public and private
buildings, and captures socio-economic indicators on Health, Food Security, Education,
and WASH at the household level. HBDA have also been expanded through additional
modules that collect other data, such as Micro/Informal SME, impact in inventories
and sales, available coping mechanisms or support received, and other aspects of
business needs.
o Digital SEIA allows impact assessment of the crisis on vulnerable households and
micro, small and medium enterprises. In most of the cases, the focus is on the impacts
of COVID-19. Households’ assessments can target specific population groups within
a country (e.g. refugees and IDPs, informal settlements, women and girls). MSMEs
assessments allow for analysis of small business, the socio-economic situation of the
workers, and can include gender, risk reduction, and other dimensions for a tailored
intervention.
o The tools can be adapted to focus on specific categories, particularly by modifying
the Household section and adding, wherever relevant, issues of concern pertaining to
vulnerable groups, economic sectors, or geographical areas. Through innovative tools
and open-source software, customized assessments can dig deeper into a thematic
area or specific indicators for both holistic and segmented views of impact.
o The tools also lend themselves to the integration of innovative solutions and aggregation
of secondary data analysis and resources, including big data (i.e. sentiment analysis),
satellite imagery, and machine learning. For instance, assessments have been carried
out with social media trend analysis and sentiment analysis from Twitter posts. Other
advanced analyses have also been carried using satellite data and machine learning
algorithms.For instance, following the Russian invasion in Ukraine, SDH assessed the
humanitarian needs of affected communities using ACLED data and machine learning
techniques.

• Project Updates:
o Household Building Damage Assessment (HBDA), piloting in Philippines, working

with the Disaster Risk Reduction & Recovery for Building Resilience at the Asia Pacific
Regional Hub, the SDH and the UNDP Philippines CO. The goal is to harmonize the
Early Recovery Needs Assessment (ERNA), Rapid Damage Need Assessment (RDNA),
the HBDA and iPDNA, on indicators glossary, data exchange and data analyses. The
need to integrate different assessments implemented after a crisis is crucial to avoiding
overlap and maximizing efficiency to address challenges in the field.
o Another new initiative is the interagency Global Data Access Initiative (GDAI). This
project is coordinated from the UNDP’s Chief Digital Office (CDO), with the technical
lead of the SDH. The project is in the design phase, and in collaboration with UN
Global Pulse, World Food Programme and McKinsey, will add value by leveraging data
on risk reduction, risk management, response and recovery. The project is expected
to have the first pilots of Minimum Scalable Product (MSP) in 2022.
o In 2023, SDH introduced a novel tool called Rapid Post-Disaster Damage Assessment
(RAPiDA), which combines the capabilities of satellite imagery and artificial intelligence
with primary data to enable effective response efforts. RAPiDA has already been
deployed in four countries, namely Syria, Libya, Afghanistan, and Gaza.
o In Syria, SDH initially utilised satellite imagery to estimate the volume of debris
in the aftermath of the 2023 earthquakes, enabling a swift response to the crisis.
120
Currently, ongoing research employs nighttime satellite data, normalized difference

vegetation index (NDVI), and machine learning to evaluate the short-term impact of
the earthquake on economic activities and food security.
UNDP
o In Libya, RAPiDA played a crucial role in assessing the consequences of recent floods
on communities by analyzing satellite data and other relevant secondary information.
This assessment informed the design of early recovery programs by the UNDP country
office in Libya.
o In Afghanistan, satellite imagery was leveraged to develop a robust multistrata geo-
sampling strategy, facilitating a household and building damage assessment (HBDA)
carried out following the recent earthquake.
o In the case of Gaza, SDH relied on nighttime data and machine learning techniques to
estimate the early impact of the conflict on economic activities and macroeconomic
indicators.
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 10 – Reduced

Inequalities; SDG 16 – Peace, Justice, and Strong Institutions; SDG 17 – Partnerships for
the Goals
• Partnership(s)/Collaborator(s): All assessments can be conducted across the globe – in all
UNDP regions – with access to additional data sources at scale. To support the process,
SDH engages partners at the local and global levels, including government authorities,
NGOs/CSOs, private sector organizations and UN agencies.
o UN Partners: SDH has worked and will continue working to build strong synergies with
relevant actors in the UN system for instance with the UN Big Data Working group,
WFP, UN Global Pulse, UNSSC, DPO, UNOOSA and the OCHA-UNDP Connecting
Business Initiative
o Government: By design, SDH works to support national authorities, INSTAT offices,
civil protection and emergency units, ministries, and works in collaboration with
partners at country and local level.
o NGO: REACH-IMPACT Initiatives, ACLED (Armed Conflict Location & Event Data
Project)
- Academia: LSE (London School of Economics, OPHI (Oxford Poverty and Human
Development Institute), QCRI (Qatar Computing Research Institute)
- Private Sector: ESRI
o Civil Society: Engineers Without Borders
• Links and Multimedia
o Twitter
o Medium
o Website
o Blogs:
- Suriname’s Indigenous community “continue to hope and believe” despite

devastation from COVID-19
- #Data4Development
- How human-centred data is helping make evidence-based decisions post-Easter
Floods
- From disaster to recovery: this is the 360 reconstructions of the archipelago of San
Andrés, Providencia and Santa Catalina
o Videos:
- Household and Building Damage Assessment in Equatorial Guinea
121
- Paris Peace Forum - Digital SEIAs
o Images:
- SURGE Data Hub Images
• Lesson Learned
o The project has underscored the importance of building internal and external capacity
for digital end-to-end digital assessment and data literacy across organizations.
SDH has designed a Household and Building Damage Assessment (HBDA) training
programme and has already certified more than 300 UNDP staff and partners.
o Partnerships for development and implementation are also important. This includes
the need for digital infrastructure, digital frameworks as well as data-sharing protocols
with external partners, and further building on synergies with other initiatives in the
UN system.
o Further action is needed to create an enabling environment for digital innovation, in
terms of appropriate resources, operating frameworks, and partnerships development.
Additional support is needed for developing innovative partnerships with the private
sector on data, digital, and infrastructure development. This could be achieved by
fostering engagement with external partners, supporting collaboration models with
appropriate resources, attracting investment, and actively supporting risk mitigation.
o UNDP COs should be further empowered, through mechanisms and support that
builds their capacity to inform governments about UNDP’s digital offer, to use existing
data and analyses, and to offer their owned datasets and analyses. This could include
a common regulated framework, clear protocols of data exchange and the necessary
digital infrastructure.
o UNDP could also be better positioned - playing an active role in clusters, and among
other agencies, to make sure data can be one of drivers toward HDP Nexus.
• Contact information: Fabjan Lashi, Project Manager (Fabjan.lashi@undp.org)
Project 3: Artificial Intelligence for Development Analytics (AIDA)

• Project Description: The Artificial Intelligence for Development Analytics (AIDA) is
a system developed to search inside more than 6,000 UNDP Evaluation Reports and
generate insights based on sentiment analysis and topic modelling. It provides 4 search
types, including by each Sustainable Development Goal.
• Department/Division: Independent Evaluation Office
• Project Type/Output: System with a website portal
• Project Status: Ongoing. Phase 1 completed in January 2022, and phase 2 completed in
September 2023.
• Project Domain: Evaluation evidence
• Data Source: UNDP Evaluation Reports from centralised and decentralised evaluations
• Link to Data: http://erc.undp.org
• Technology/Platform: Amazon AWS with hosting provided by UNICC
• Related Sustainable Development Goals (SDGs): Project covers work related to all SDGs.
o Collaboration with UNDP ITM, UNICC and Amazon Web Services.
• Relevant links and Multimedia: https://aida.undp.org
122
• Contact information: Gonzalo Gomez (gonzalo.gomez@undp.org), Anish Pradhan (anish.

pradhan@undp.org)
UNDP
Project 4: iVerify
• Project Description: Building on UNDP’s long standing support for democratic
governance, iVerify is part of the next generation of tools to assist states in carrying out
free and fair elections in this era of misinformation.
The tool uses open-source machine learning to track misinformation and AI to detect hate
speech. It leverages Crowdtangle to conduct daily monitoring of significant conversations
happening on Facebook and runs posts through an open algorithm, Detoxify, to
determine whether there might be toxic content. When such content is flagged, it gets
sent to Meedan Check, an open collaborative media annotation platform that uses a
‘humans in the loop’ approach to combine machine learning with human oversight.
Trained experts from a number of organizations can use the platform to send potentially
toxic content and/or misinformation - using email, an online form, Facebook Messenger,
WhatsApp and Telegram - to local partner organizations who can verify the content as
true, false or somewhere in between.
AI increases efficiency by matching flagged content to similar posts, and reports are
published automatically to a Wordpress website for access and amplification - with a
simple traffic light system that enables users to see at a glance whether a claim is true, half
true, unproven, misleading, false, toxic or later retracted. An analytics dashboard further
supports quick analysis, agile monitoring and evaluation.
iVerify is built like a comprehensive support package around the provision of digital tools
as well as the expert technical support to activate the digital innovations. The project
has been deployed in Zambia at the occasion of the August 2021 general elections.
The team is actively working at evaluating the impacts of that pilot phase while planning
for the continuation of the activities in a sustainable manner. In addition, the initiative is
currently in a roll-out phase in Honduras ahead of the November presidential elections.
Other countries for future planning include: Zimbabwe, Liberia, Mali and Kenya.
• Department/Division: EC-UNDP Joint Task Force on Electoral Assistance, Bureau for
Policy & Programming and Chief Digital Office
• Project Type/Output: Software tool and dataset
• Project Domain: Governance and democratic process. While it is currently applied with
a focus on elections, the tool has many potential applications – from preventing human
rights abuses by flagging hate speech to encouraging vaccine take-up by combating
misinformation.
• Data Source: Social Media Data, information submitted via tip lines
• Link to Data: https://www.crowdtangle.com/
• Data Publicly available: Yes in anonymised form
• Technology/Platform: Crowdtangle, Detoxify
• Related Sustainable Development Goals (SDGs): SDG 4 Quality Education; SDG 5 –
Gender Equality; SDG 16 – Peace, Justice, and Strong Institutions; SDG 17 – Partnerships
for the Goals
o UN Partners: UN Development Programme, UN International Computing Centre

o Government: European Union, USAID, UK aid, Irish aid, Germany, Sweden
o Private Sector: Crowdtangle, Meedan
o Civil Society: Various organizations, where in each country these is a different partner
123
o Academia: UNITEC Honduras (Honduras application)
o https://www.undp.org/digital/iverify
o https://digitalpublicgoods.net/about/
o https://www.undp.org/policy-centre/oslo/publications/guidance-implementation
-undp-iverify
o https://www.undp.org/policy-centre/oslo/publications/undp-iverify-monitoring-and
-evaluation-framework
o https://sl.i-verify.org/
o Prioritise preparation time

Getting things ready well before the start date is crucial for setting up a tool like iVerify
successfully. This process involves developing and getting the system up and running,
training the individuals who'll be using it, and starting an awareness campaign. During
the planning phase, it's also important to make sure there's enough time to get the
system up to speed. This preparation time is vital for iVerify. It allows the system to
work faster and more efficiently and helps to build its reputation. This all means iVerify
can make a real difference in reducing the amount of misinformation, disinformation,
and offensive speech being spread.
o Set achievable goals
The process of implementing iVerify requires a careful balance between ambitions,
time constraints, and available funding. An assessment mission is essential in guiding
the best method of allocating resources in an impactful, efficient, and sustainable way.
It's also crucial to understand the ongoing efforts required to maintain the system after
the primary roll-out.
o Ensure local presence for an efficient implementation follow-up
During the early stages of implementation, having an iVerify team on the ground
or a dedicated focal person is fundamental for smooth and swift coordination. This
arrangement helps tackle any immediate challenges and keeps the implementation
process on track.
o Consider refresher trainings
Training individuals on how to use iVerify is certainly a key step, but it's just as important
to keep the learning process going. Just as you would revisit a recipe to perfect a
dish or repeat a drill to improve at sport, going back to the training helps keep the
knowledge fresh and builds user confidence.
o Maintain the online-offline balance
It's important to note the significant role offline mediums, such as radio, continue to
play in disseminating information, especially in developing countries. Despite the
increasing digitalisation, many communities still rely on traditional media channels
for their daily news and updates.
This reality presents an opportunity for tools like iVerify. While a strong online presence
is undoubtedly advantageous, ensuring accessibility through offline channels, like
radio, becomes equally crucial. This strategy allows us to extend iVerify's reach
beyond internet users, thereby creating a more comprehensive network for countering
misinformation and ensuring that our tool benefits a larger, more diverse audience.
o Invest in raising awareness is critical
Awareness plays a critical role in the success of any initiative. Without it, even the
most effective tool may go unnoticed and unused, rendering all development
and implementation efforts ineffective. In the context of iVerify, a comprehensive
communication campaign is instrumental to the success of this initiative. Such a
124
campaign can dramatically amplify iVerify's impact and uptake, positioning it as

a go-to resource for fact-checked news within the community. Ultimately, a well-
planned communication strategy can transform iVerify from a mere tool into a widely
recognized and trusted resource for countering misinformation.
UNDP
o Invest in core functionality
Like any digital tool, continuous tweaking, debugging, and feature enhancements
based on user feedback are all integral to the long-term effectiveness of tools such
as iVerify. This ongoing improvement and maintenance process helps the tool stay
relevant and valuable in identifying potential misinformation, disinformation, and hate
speech. However, such work requires not just time, but also consistent investment.
Often, funding is secured for specific deployments and might overlook the need for
continual investment in the tool's core functionality. Hence, it's important to develop
a funding strategy that recognizes and accommodates the need for regular updates
and improvements.
• Contact information: Gabriel Van Oppen (gabriel.van.oppen@undp.org) Saré Knoope

(sare.knoope@undp.org )
Project 5: eMonitor+
• Project Description: eMonitor+ is a set of digital tools, with advanced artificial intelligence
capabilities, which is used to expand national, regional, and global stakeholder capacities
to analyse and address a range of challenges, including but not limited to toxic
communication, hate speech, disinformation, and gender-based violence. eMonitor+
collects and analyses thousands of pieces of online content daily, from platforms like
Facebook, Instagram, Twitter, and YouTube, as well as website content and offline media.  
The system integrates external tools and solutions into UNDP-built software. Nonetheless,
human monitors lead its operation, utilising the suite's automated functions to streamline
otherwise time-consuming tasks. This fusion of human oversight and AI-driven analytics
results in a more cost-effective, time-efficient, and comprehensive monitoring process.
Every eMonitor+ deployment is accompanied by comprehensive technical support,
enabling national stakeholders to develop and execute information integrity strategies
that are evidence-based and tailored to address the unique challenges of their contexts.
Outcomes can include advancing long-term digital resilience, particularly in the form
of innovative policies, encouraging whole-of-society responses, and enhancing media
literacy and public awareness.
eMonitor+ has supported national partners in several countries, analysing more than 2
million pieces of online content, and is being used by more than 100 national monitors
and fact-checkers. The system has contributed to fostering a national culture of digital
media analysis and the measurement of trends in information pollution issues, which were
previously not researched. Additionally, the support enhanced the capacities of partners
and built national alliances to promote information integrity. eMonitor+ has contributed
to strengthening public resilience and media literacy through communicating results to
the public.
• Department/Division: BPPS
• Project End Year: -
• Project Domain: Information integrity - Democratic Institutions and Process
• Data Source: Social media
• Technology/Platform: 1- ReactJS 2- Elasticsearch. 3- Docker 4-Node.js. 5- Python 6. Multer
7. Nodemailer.
125

Reduced Inequalities, SDG 16 - Peace, Justice & Strong Institutions
• Partnership(s)/Collaborator(s): Government, Media Institutions, Civil Society, Academia,
and Electoral Management Bodies
• Links and Multimedia: https://www.undp.org/arab-states/emonitor
• Contact information: Osama Aljaber (osama.aljaber@undp.org)
Project 6: Horizon Scanner (HS)
The Horizon Scanner (HS) is an off-the-shelf software suite which brings a paradigm shift by
incorporating built-in automation, empowering users to identify and assess emerging risks
and opportunities. Identified problems and how the Horizon Scanner suit addresses them:
The following are the identified problems that the Horizon Scanner has been designed and
built to solve, along with the corresponding solutions.
1. Fragmented Data Landscape:
- Solution: The platform consolidates all available data into a comprehensive, categorized,
and prioritized repository. It establishes a one-stop-shop, future-proofing data accessibility
for the organization.
2. Costly and Time-Consuming Data Collection:
- Solution: The built-in feature for data requests, automated questioning protocols, and
innovative tools like API integration, wizard features, and drop boxes streamline communication
between organizations. The platform provides proxy indicators developed from available data,
and recommends data collection when necessary, resulting in a significant 65% reduction in
time and costs based on previous trials.
3. Conflicting Data Sources:
- Solution: The platform centralizes all data sources, offering an additional reliability matrix and
triangulation according to organizational guidelines, ensuring data consistency and coherence.
4. Enhancing Data Literacy:
- Solution: The user-friendly design (UX) of the platform ensures seamless usage, supplemented
by training initiatives, fostering enhanced data literacy among staff.
5. Financial Constraints in Data Positioning:
- Solution: The platform overcomes financial constraints by deploying analytics power through
algorithm development and AI-backed mechanisms. It harnesses the potential of existing data,
providing new insights and optimizing data collection requests for strategic positioning.
Key Product Features:
• Customization: Tailor the system to your specific needs by seamlessly incorporating on-
demand features.
• Interrogatable Nature: Dive deep into analysis and customization with the software's user-
friendly and easily interrogatable with existing ERP systems.
• Real-world Applicability: Tested through two integration experiments and five extensive
pilots in Sudan (3 to 9 months each), the software has evolved based on user feedback.
126
Insights from these trials have guided the completion and successful launch of the pilot
phase for expansion in Egypt.
Strategic Collaborations:
UNDP
In partnership with UNDP's Regional Bureau for Arab States (RBAS), the software is to go
through its second integrability test with eMonitor+, an AI-powered elections monitoring and
fake news detection system. This collaboration enhances the software's capabilities, solidifying
its position as a comprehensive solution for horizon scanning and risk management.
Current Phase (scale-up):
Localization of the software and the development of on-demand features are currently
underway in Egypt in collaboration with UNDP partners, including the government. Our
software employs a data-driven approach to ensure relevance and up-to-date functionality,
actively supporting horizon scanning efforts in the country. This phase reflects our dedication
to practical implementation and the software's real-world impact.
In Conclusion:
This innovative software stands as a powerful tool, offering a data-driven solution for proactive
horizon scanning. Leveraging ready-available segmented data, this analytics software generates
insightful analyses, significantly cutting costs and time in data collection. The utilization of
segmented data leads to a remarkable reduction in overhead costs for data collection, up to
65%. Additionally, the software's automated opportunities and threats detection streamlines
processes, enabling new projects and interventions. With its adaptability, interrogatable nature,
and successful integrations, it serves as a valuable asset for organizations aiming to enhance
their decision-making processes.
• Horizon Scanner (Sudan - SHS) Demo: https://d rive. google. com/f ile/d /
19hn7rjmH67O4agSevj-HA7xNbiCcpARo/view
Horizon Scanner (Egypt - EHS): Egypt Horizon Scanner (EHS) Implementation Handbook
https://drive.google.com/file/d/1s60zDBoMMFUkj0eOlNQwAlgaj3S1FbOj/view
Egypt Horizon Scanner (EHS) Custom Features Design https://drive.google.com/file/d/1s60

zDBoMMFUkj0eOlNQwAlgaj3S1FbOj/view
• Presentation: SHS Algorithm AI Presentation1

• Contact information: Wigdan Seedahmed wigdan.seedahmed@undp.org OR
wigdan229@gmail.com (preferred)
SDGs: Goal 9 & Goal 17 (for both Sudan and Egypt versions). Plus Goal 16 (for Sudan Horizon
Scanner due to an on-demand customizable feature)
Project 7: POPP ChatBot

• Project Description: Leveraging ChatGPT 4.0 to create a virtual agent to answer questions
about UNDP policies and procedures.
• Department/Division: ITM and BPC/BMS
• Project Type/Output: ChatBot embedded in Intranet
• Project Start Year: Q3/2023
127
• Project End Year: Q1/2024

• Project Domain: Operations
• Data Source: POPP web sites
• Technology/Platform: Microsoft Azure
• Related Sustainable Development Goals (SDGs): SDG 9 - Industry, Innovation, &
Infrastructure
• Partnership(s)/Collaborator(s): Microsoft
• Links and Multimedia: popp.undp.org
• Contact information: Diwen Xu (diwen.xu@undp.org)
Project 8: Revamping of the UNOSSC Galaxy and introduction of an AI digital assistant

• Project Description: UNOSSC and CDO are revamping the South -South Galaxy repository
of solutions on SSTrC. As part of the initiative, AI will be used to assist end-users to
navigate, extract insights and access valuable resources on the platform related to SSTrC.
• Department/Division: United Nations Office for South-South Cooperation (UNOSSC)
• Project Type/Output: Digital dynamic repository of SSTrC solutions
• Project Domain: Cross-cutting
• Data Source: Multiple datasets
• Technology/Platform: TBC
• Related Sustainable Development Goals (SDGs): SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): Various partnerships with member states, UN agencies,
think tanks, academia and private sector.
• Links and Multimedia: https://unsouthsouth.org/library/south-south-galaxy/
• Contact information: Dumitru Vasilescu, Research Specialist, UNOSSC dumitru.vasilescu@
unossc.org
Project 9: Viva Implementation

• Project Description: As part of UNDP's continued efforts in transforming our digital
workplace experience, Microsoft (MS) Viva Suite was enabled across the organisation.
The MS Viva Suite, including Viva Connections, Viva Insights, Viva Engage, Viva Topics,
Viva Learning and Viva Goals, will ensure all personnel are connected to their colleagues
and have personal productivity insights, expert knowledge, strategic priorities and key
learning resources available at their fingertips. All those components are backed up by
Microsoft AI engines.
• Department/Division: ITM/BMS, OHR/BMS and SDGI/BPPS
• Project Type/Output: SharePoint, Teams, Viva Suite
• Project Domain: Operations and Management
128
• Data Source: M365 datasets

• Technology/Platform: Microsoft 365
UNDP
Infrastructure, SDG 12 - Responsible Consumption & Production, SDG 17 - Partnerships
for the Goals
• Partnership(s)/Collaborator(s): Microsoft
• Links and Multimedia: https://undp.sharepoint.com/teams/ITM/SitePages/DWP
-MSvivasuite.aspx (internal)
• Contact information: Diwen Xu (diwen.xu@undp.org), Daniel Tshin (daniel.tshin@undp
.org)
Project 10: ABADEI Chatbot project -- Afghanistan Country Office

• Project Description: ABADEI project implemented by Afghanistan Country Office collects
a vast amount of data from multiple sources, stores in a SQL database and visualises
data on a web-based portal. The ABADEI portal is soon expected to introduce a chatbot
powered by Azure OpenAI where users could ask questions about the data.
• Department/Division: AFG/RBAP and ITM/BMS
• Project Type/Output: Chatbot
• Project Domain: Humanitarian crisis response
• Data Source: Azure SQL
• Technology/Platform: Microsoft Azure/OpenAI/ChatGPT
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 2 - Zero
Hunger, SDG 3 - Good Health & Well-being, SDG 10 - Reduced Inequalities, SDG 11 -
Sustainable Cities & Communities
• Partnership(s)/Collaborator(s): Afghanistan Government
• Links and Multimedia: https://www.undp.org/afghanistan
• Contact information: Arvind Kumar (arvind.kumar@undp.org)
Project 11: Entity extraction and text formulation for emails

• Project Description: The project will allow OAI investigators to go through identified email
account messages to flag important information related to investigations.
• Department/Division: OAI and ITM/BMS
• Project Domain: Operations
• Data Source: Outlook Messages
• Technology/Platform: Microsoft Azure/ChatGPT
129

Infrastructure
• Partnership(s)/Collaborator(s): N/A
• Links and Multimedia: N/A
• Contact information: Daniel Samuel (daniel.samuel@undp.org)
Project 12: GeoAI for coastline erosion and slums detection

• Project Description: The project is an experiment of GeoAI in two fields of work:
1) Coastline erosion: was tested over multiple years on the Liberia shore and Gaza
shoreline.
2) Slums detection: was created for the Philippines by a private sector through the RBAP
regional hub which we took over to corporately host it and test it. slums detection was
tested over Cairo, Egypt
• Department/Division: GIS/DATA/ITM/BMS
• Project Type/Output: GIS
• Project Start Year: Implemented in 2022 to 2023.
• Project End Year: 2023 but with ongoing finetuning and testing for improving the machine
learning algorithm.
• Project Domain: Coastline erosion and slums detection
• Data Source: Satellite Images with ground truth data for validation
• Technology/Platform: Python script, Azure Hosting, Integration with ArcGIS Enterprise
Server
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 13 - Climate
Action
• Partnership(s)/Collaborator(s): Academia
• Links and Multimedia: https://undp.sharepoint.com/:p
:/r/sites/SatelliteImageryAnalysis/
Shared%20Documents/Projects/Coastline%20Tool/Coastline_Tool_presentation_to
_end_user.pptx?d=wfe1abf2577d743c897cf2951b2f2f4ac&csf=1&web=1&e=PznZ48
(internal)
• Contact information: Tala Hussein (tala.hussein@undp.org)
Project 13: AI for Sustainable Development

• Project Description: This project proposes an AI-powered platform designed to bridge
the information gap and empower the region. This project is expected to result in accurate
monitoring, data-driven insights, an early warning system, forecasting capabilities, and
scenario analysis, all aimed at driving sustainable development in the Arab Region. It is
poised to address the region's data deficiency challenge comprehensively and contribute
significantly to achieving the Sustainable Development Goals.
• Department/Division: RBAS
130
• Project Domain: Optimising Agricultural and Environmental Resilience, Ensuring Climate

Adaptation, Improving water resources management, Fostering Social Cohesion,
Reducing inequalities
UNDP
• Data Source: Satellite imagery, weather data, social media data, news reports, and more
• Technology/Platform: Google Cloud, Python
• Partnership(s)/Collaborator(s): Multiple partners, including Office of the Minister of AI in
the UAE, MBZUAI, TII, and more to be confirmed.
• Links and Multimedia: Panel discussion at COP28 that convened main partners: https://
www.undp.org/arab-states/press-releases/uaes-minister-state-artificial-intelligence-undp
-and-e-discussed-ways-harness-power-ai-sustainable-development
• Contact information: Dany Wazen (dany.wazen@undp.org) and Yasmine Hamdar
(yasmine.hamdar@undp.org)
Project 14: UN Volunteers Description of Assignment qualitative rating

• Project Description: UN Volunteers Description of Assignment qualitative rating. For
each request for a UN Volunteer of assignment, a Description of Assignment (DoA) is
produced. As it contains lots of text and description, and is critical to identify suitable
talent, we will deploy an AI solution to give a qualitative scoring to this DoA.
• Department/Division: ICTS/UNV/BMS
• Project Type/Output: Feature within the UNV Recruitment tool
• Project Status: Not initiated
• Project Start Year: Q2 2024
• Project End Year: Q1 2025
• Project Domain: HR
• Data Source: UNV recruitment data
• Technology/Platform: TBC
Reduced Inequalities, SDG 16 - Peace, Justice & Strong Institutions
• Partnership(s)/Collaborator(s): None
• Links and Multimedia: Not applicable yet
• Contact information: Frederic LeMaistre (frederic.lemaistre@unv.org)
Project 15: Chat GPT AI application for Advanced Search in Project Information
Management System, PIMS+
• Project Description: By applying the Chat GPT technology, PIMS+ is aiming to provide a
more in-depth, communication based data search/analysis function to the user this year.
• Department/Division: BPPS/VF oversight unit
• Project Type/Output: Dataset is the product documents and other projects documents
uploaded in PIMS+ platform
• Project Start Year: Q1-2 in 2024 for the preliminary version to be published in PIMS+
website.
131
• Project Domain: System and Data

• Data Source: Data & documents entered by each project team
• Technology/Platform: Chat GPT4
• Related Sustainable Development Goals (SDGs): SDG 13 - Climate Action
• Partnership(s)/Collaborator(s): None
• Links and Multimedia: pims.undp.org (internal)
• Contact information: Bonjung Goo (bonjung.goo@undp.org)
Project 16: Thematic tagging of UNDP projects

• Project Description: Using supervised machine learning algorithms to tag UNDP projects
based on project description text. Modelling based on pre-tagged training data sets with
verification and fine-tuning by iterative training and model-adjusting.
• Department/Division: BPPS/Effectiveness/Results
• Project Type/Output: Excel data sets, STATA script (SVM)
• Project Start Year: 2019 and is on-going with adjustment of models for existing themes
and expansion of themes (social protection, SP outputs, SDGs, etc.)
• Project Domain: Various thematic areas (rule of law, climate, disaster, digital, anti-racism,
women's political participation, etc)
• Data Source: Project description data extract from Quantum (not public). Training
completed by thematic subject experts.
• Technology/Platform: Local offline PC with potential for Azure virtual machine (cost
implications).
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 5 - Gender
Equality, SDG 10 - Reduced Inequalities, SDG 11 - Sustainable Cities & Communities,
SDG 13 - Climate Action, SDG 16 - Peace, Justice & Strong Institutions
• Partnership(s)/Collaborator(s): Thematic experts from various BPPS/CB units.
• Links and Multimedia: Portfolio Analytics Dashboard (internal Power BI report)
• Contact information: Xiaodong Cai (xiaodong.cai@undp.org), Huiyan Du (huiyan.du@
undp.org)
Project 17: Sentiment analysis of UNDP-tagged tweets

• Project Description: With help from UN Pulse API, we extract tweets from the X platform
(previously known as Tweeter) for the past 6 months, where UNDP was tagged. Then
three off-the-shelf python sentiment analysis packages are applied on the tweets with an
automated/calculated final assessment based on the agreement among the results from
the three packages. Initial testing showed accuracy above 85% within 1500+ human-
labelled tweets. The results are visualised in an internal Power BI report.
• Project Type/Output: Excel data sets, Python packages
132
• Project Domain: Social media

• Data Source: UN Pulse Tweeter API
UNDP
• Technology/Platform: Local offline PC with potential of full online automation within
Power BI
• Partnership(s)/Collaborator(s): UN Pulse
• Links and Multimedia: Internal Power BI report
undp.org)
Project 18: Mapping UNDP projects against SDG concepts and indicators
• Project Description: We use the UN LinkedSDGs API to analyze UNDP projects and map
them against SDG concepts and indicators. The API employs natural language processing
(NLP) techniques to identify key concepts based on UNBIS and EuroVoc thesauruses and
their association with specific SDGs. For each project, we extract the top 5 key concepts
and the top 3 SDG goals most closely related to these concepts. Then we integrate the
extracted concepts and related SDG goals into UNDP's project portfolio dashboard,
which allows for easy categorization and filtering of projects based on their alignment
with specific SDGs and concepts.
• Project Type/Output: Excel data sets, Python
• Project Domain: SDGs
• Data Source: UNDP project database
• Technology/Platform: Python, Power BI
• Links and Multimedia: Internal Power BI report
undp.org)
Project 19: Automated Scoring System for UNV Volunteer Descriptions of Assignment
(DoA)
• Project Description: United Nations Volunteers (UNV) performs the full recruitment cycle
for thousands of volunteers for various UN agencies for different kinds of contract types.
These volunteer assignments cover a range of specifications, encapsulated within job
descriptions that include both structured and unstructured data. However, the variability
in the quality of these job descriptions drafted by different UN agencies necessitates an
automated scoring system to rank them based on quality.
• Department/Division: United Nations Volunteers (UNV)
• Project Type/Output: Dataset and Software service
133

• Project Domain: Human Resource analytics
• Data Source: UNV Descriptions of Assignment database
• Technology/Platform: Potentially: Azure OpenAI or Open source model hosted on Azure.
Industry, Innovation, & Infrastructure, SDG 17 - Partnerships for the Goals
• Contact information: Ashok Hariharan (ashok.hariharan@unv.org)
Project 20: SDG Push Diagnostics (National Priorities)

• Project Description: The National Priorities feature of the SDG Push Diagnostics enables
users to upload and analyse policy documents with respect to 17 Goals. Documents such
as Voluntary National Reviews (VNRs) and National Development Plans can be used to
identify priorities of the government that can be mapped to the SDGs. A neural network
algorithm helps to discover which SDGs feature most prominently as a priority. The SDG
Push Diagnostics platform already provides results for more than 120 VNRs. Link: https://
sdgdiagnostics.data.undp.org/AFG/current-priorities
• Department/Division: Bureau for Policy and Programme Support / SDG Integration
• Project Type/Output: Web-based software tool
• Data Source: Documents, e.g., publications, reports, policy briefs and so on.
• Technology/Platform: Python, Microsoft Azure.
• Partnership(s)/Collaborator(s): National governments in more than 100 countries.
• Links and Multimedia: https://sdgdiagnostics.data.undp.org/AFG/current-priorities
• Contact information: Babatunde Abidoye (babatunde.abidoye@undp.org) or Mykola
Skrynnyk (mykola.skrynnyk@undp.org)
Project 21: DFx AI Chat

• Project Description: UNDP AI as a Service (AIaaS) is a collection of services that leverage
the latest advancements in Machine Learning (ML), Natural Language Processing (NLP)
and Generative Artificial Intelligence (GenAI) to help UNDP staff tap into the wealth of
data for better development policies. It consists of DFx AI Chat, a flagship multi-purpose
GenAI platform built upon Azure OpenAI, and a set of micro-services we use to build
custom products that enable you to extract, analyse and summarise knowledge from
non-traditional data sources including webpages, documents, images, and social media
interactions at scale.
• Project Type/Output: Web-based software tools, platform, APIs.
134

UNDP
• Data Source: Any text data, including but not limited to project documents, reports,
publications etc.
• Technology/Platform: Python, Azure OpenAI Service, Azure Cognitive Services and other
Microsoft Azure services.
• Partnership(s)/Collaborator(s): Government and private sector.
• Links and Multimedia: https://chat.data.undp.org and https://data.undp.org/what-we
-do/ai-as-service
Project 22: NDC Analytics App

• Project Description: Nationally determined contributions (NDCs) are at the heart of the
Paris Agreement and the achievement of its long-term goals. NDCs embody efforts by
each country to reduce national emissions and adapt to the impacts of climate change.
The NDC Analytics App is an AI-powered application that allows users to search through
120 Nationally Determined Contributions (NDCs) to find information of interest. The
application supports search queries in English, French and Spanish.
• Project Type/Output: A web-based application
• Project Domain: Environment, energy and climate
• Data Source: Documents from NDC Registry.
• Technology/Platform: Python
• Related Sustainable Development Goals (SDGs): SDG 7 - Affordable & Clean Energy, SDG
11 - Sustainable Cities & Communities, SDG 13 - Climate Action, SDG 17 - Partnerships
for the Goals
• Partnership(s)/Collaborator(s): Governments.
• Links and Multimedia: https://ndc-analytics-app.azurewebsites.net
Project 23: Gender Social Media Monitoring Tool

• Project Description: The Gender Social Media Monitoring Tool is a pilot initiative that
leverages artificial intelligence technologies to track conversations on social media
networks across categories relevant to gender-responsive policymaking: education,
politics, reproductive rights, work, and violence. Supporting over 100 languages, the
tool detects harmful content and hate speech against women and girls, an indicator of
online gender violence and backlash against women’s rights and gender equality. It can
135
also provide policymakers, civil society and practitioners with new data and insights to
take action to address these backlashes.
• Project Type/Output: A web-based tool
• Project Domain: Gender
• Data Source: Social media data
• Technology/Platform: Python, Twitter API and Databricks.
Reduced Inequalities, SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): Government, private sector and other UN agencies.
• Links and Multimedia: https://data.undp.org/insights/gender-social-media-monitoring
• Contact information: Babatunde Abidoye (babatunde.abidoye@undp.org) or Raquel
Lagunas (raquel.lagunas@undp.org)
Project 24: Identifying electricity access gaps at a hyperlocal level

• Project Description: UNDP, in partnership with University of Michigan, has used high-
resolution satellite imagery to complement household derived information on electricity
access at subnational level for more than 100 countries. By monitoring and comparing
how strong the night-time light of a specific settlement is in relation to other areas around
the globe, the approach uses a statistical model to calculate the likelihood of a human
settlement to be electrified. Unlike household surveys, satellite data allows us to rapidly
capture the evolution, quality and frequency of electrification. Based on this approach,
the number of people that lack access to reliable electricity services surpasses 1 billion –
a much larger figure than the official count of 675 million lacking electricity access. This
suggests that not all those classified as having access actually benefit from electricity.
• Project Type/Output: A web-base tool and a dataset
• Technology/Platform: Python and Microsoft Azure
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 7 - Affordable
& Clean Energy,
• Partnership(s)/Collaborator(s): Academia and private sector.
• Links and Multimedia: https://data.undp.org/insights/achieving-universal-electricity
-access/electricity-access-from-space
• Contact information: Babatunde Abidoye (babatunde.abidoye@undp.org) or Riad
Meddeb (riad.meddeb@undp.org)
136
Project 25: SIDS Data Platform

• Project Description: UNDP’s integrated approach supports Small Island Developing
States to accelerate transformative development based on three pillars: Climate Action,
UNDP
Blue Economy, and Digital Transformation.
• Project Type/Output: A web-based tool
• Data Source: Indicator data and geospatial data.
• Technology/Platform: JavaScript and Python.
• Partnership(s)/Collaborator(s): Governments.
• Links and Multimedia: https://sids.data.undp.org
• Contact information: Babatunde Abidoye (babatunde.abidoye@undp.org) or Oksana
Leshchenko (oksana.leshchenko@undp.org)
Project 26: Risk Anticipation Data Hub

• Project Description: The project aims to transform an existing UNDP Crisis Bureau internal
data platform (Crisis Risk Dashboard) into a digital public good by providing a one-
stop data warehouse combining a large set of public datasets to facilitate integrated
multidimensional data analytics. A key feature of this project will be a Generative AI
LLM front-end that will enable analysts to interact with the data warehouse using natural
language queries. This innovative feature will empower analysts to extract insights
without relying on traditional business intelligence tools, enhancing rapid risk analysis
and reducing barriers to entry and costs.
• Data Source: Multiple sources of mostly tabular data
• Technology/Platform: TBD
Infrastructure, SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): CRAF'd funded project
• Links and Multimedia: TBD
• Contact information: Shouryadipta Sarkar (shouryadipta.sarkar@undp.org)
137
Project 27: Digital Social Vulnerability Index (DSVI)

• Project Description: The Digital Social Vulnerability Index is a cutting-edge digital solution
for vulnerability assessments to monitor and understand the exact location, distribution
and underlying drivers of social vulnerabilities. While previous vulnerability measures
would require conducting timely and costly surveys, the DSVI provides a higher resolution
and improved representation of a country’s social vulnerability beyond administrative
boundaries. Moreover, compared with previous instruments, the DSVI is the first tool of
its kind to incorporate a much more comprehensive SV analysis by integrating numerous
data sources and indices into one.
Tailored to the specific needs of UNDP, government agencies and other development
actors, DSVI can provide effective digital SV analyses with an implementation time
frame of two to three months per country. It seeks to help UNDP national or regional
offices to improve their understanding of local vulnerabilities, thereby facilitating the
adoption of more targeted and coordinated interventions that build stronger community
resilience. DSVI offers a set of outputs which help to deliver the key messages and data
to its target audiences. These elements start with raw datasets and scientific methods
for calculation and end with training sessions, maps, reports or digital infrastructure to
visualise the findings. First, DSVI offers high-accuracy SV scores. SV scores are calculated
by an automated data science pipeline which gathers high-quality and freely available raw
data from USAID, the UN and scientific resources. Using GIS and machine learning, DSVI
generates high-resolution vulnerability maps. These maps are a new, more technologically
advanced addition to the long tradition of vulnerability mapping.
The DSVI has been implemented in 6 countries: Tajikistan, Burkina Faso, Niger, Ethiopia,
Kenya and Somalia.
• Department/Division: BPPS/ICPSD/SDG AI Lab
• Project Type/Output: Digital Social Vulnerability Index (DSVI) calculations, Interactive
Social Vulnerability Maps, Analytical Briefs and Trainings (General and Technical)
• Project Domain: Resilience
• Data Source: USAID Demographic and Health Surveys (DHS), satellite imagery, open-
source data
• Technology/Platform: Python, GIS, Machine Learning
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 6 - Clean
Water & Sanitation, SDG 8 - Decent Work and Economic Growth, SDG 9 - Industry,
Innovation, & Infrastructure, SDG 10 - Reduced Inequalities, SDG 11 - Sustainable Cities
& Communities, SDG 13 - Climate Action, SDG 16 - Peace, Justice & Strong Institutions,
SDG 17 - Partnerships for the Goals
o UNDP Country Offices (Tajikistan, Niger, Burkina Faso, Niger and UNDP Resilience
Hub)
o National stakeholders (Ministries and Statistics Agencies)
o UNDP Disaster Risk Reduction and Recovery for Building Resilience Team (DRT)
o Policy makers and decision makers in selected countries
o UN Online Volunteer Data Scientists
• Links and Multimedia: https://www.undp.org/policy-centre/istanbul/news/iicpsd-undp

-tajikistan-hosted-training-digital-social-vulnerability-index-dsvi
138
• Contact information: Gokhan Dikmener (gokhan.dikmener@undp.org )
Project 28: Frontier Technology Radar for Disaster Risk Reduction (FTR4DRR)
UNDP
• Project Description: Frontier Technology Radar for Disaster Risk Reduction (FTR4DRR) is
an online tool for the monitoring and tracking of existing and emerging digital solutions
for anticipatory risk, risk reduction, and crisis recovery. Developed through a partnership
between UNDP’s Disaster Risk Reduction and Recovery for Building Resilience Team
(DRT), ICPSD SDG AI Lab, and Connecting Business Initiative (CBi), the FTR4DRR aims to
effectively support UNDP and other development stakeholders in navigating fast-growing
digital solutions for disaster risk reduction and management. It enables an improved
understanding of the technological landscape among development stakeholders and
encourages the adoption of proven solutions while furthering UNDP as a thought leader
on the use of innovative technologies in disaster contexts.Since the launch of FTR4DRR in
October 2022, SDG AI Lab has continued developing the tool. In addition, the partners
have hosted three outreach webinars that introduced how frontier technologies are used
in Disaster Risk Reduction (DRR).
• Project Type/Output: Online web platform
• Project Domain: Resilience
• Data Source: UNDP publications, open sources, online volunteers crowdsourcing
• Technology/Platform: Typescript, React, sass, D3.js
Infrastructure, SDG 11 - Sustainable Cities & Communities, SDG 13 - Climate Action,
SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): UNDP Disaster Risk Reduction and Recovery for Building
Resilience Team (DRT) and Connecting Business Initiative (CBi)
• Links and Multimedia: https://drrtechradar.org/#/
Project 29: Crisis Risk Dashboard: Predictive Analytics and Multi-dimensional Analysis
• Project Description: The Risk Anticipation Hub is developing predictive analytics
components to the Crisis Risk Dashboards and related products. This includes:
1. Better anticipation of conflicts & impacts of prevention
We leverage the most recent advances in machine learning to obtain conflict forecasts and
conflict prevention impact estimates to improve risk anticipation. For this we combine the
world-leading conflict forecasting systems in a new data product that will enable forward-
looking risk analysis and decision-support, including a prioritisation of countries for prevention
activities and sub-national forecasts of conflict risks.
2. Conflict forecasting using news data
We utilise GDELT’s high frequency, text-based data to capture the “chatter” of what is happening
in local contexts to see if it is predictive of conflict. We limited the geographic scope to the
139
Horn of Africa and aggregated to each country’s admin level 1 and to a 1-2 week time interval.
With an escalation of ACLED fatalities as the outcome to predict, we utilise article document
embeddings as the main predictors, and conflict history and some static features (population
density, etc.) as additional features.
3. Multi-dimensional analysis
We utilise AI and machine learning-based technologies for in-depth multi-dimensional analyses.

The types of analyses include:
* Natural language processing and text analysis - Given a body of texts, we can find the topics
and representative keywords, commonly used terms and phrases, and sentiments and emotions
reflected in the content. We can also categorise texts into predefined groups to find trends,
and summarise and distil text into its most essential content. This includes analysis of social
media data.
* Hotspot and spatial analysis - Given data on events or phenomena, hotspot analysis can find
patterns and uncover relationships between spatial information. We can identify geographic
areas with significantly high (or low) levels of incidents for further investigation.
* Outlier and anomaly detection - Given data on events or measurements, various algorithms
and methods can be used to find unexpected values that differ significantly from the norm or
usual trend. This can alert us to further monitoring or investigation.
* Time series analysis - Given data recorded over time at regular intervals, we can identify
underlying trends and detect seasonality and repeating patterns to better understand a context.
Department/Division: Crisis Bureau
• Project Type/Output: Software tool, dashboard

• Project Start Year: Ongoing
• Data Source: Multiple open source datasets
• Technology/Platform: Azure services, Databricks, Python
Infrastructure, SDG 16 - Peace, Justice & Strong Institutions, SDG 17 - Partnerships for
the Goals
• Partnership(s)/Collaborator(s): Several.
• Contact information: Sun-joo Lee (sun-joo.lee@undp.org )
Project 30: Fair Digital Finance SupTech

• Project Description: UNDP Morocco's Accelerator, Morocco Central Bank, and ICPSD
SDG AI Lab partnered to analyse consumer sentiments to identify vulnerable social
groups' needs. To do so the partners used social listening to amplify the collective voice
of consumers and highlight the concerns of the most marginalised people. ICPSD SDG
AI Lab has used Machine Learning (ML) and Natural Language (NLP) techniques to collect
140
and analyse data from social media platforms like Twitter, Facebook, and Instagram.
The developed Supervisory Technology (SupTech) solution offers insights on financial
stability, inclusion, and market conduct. ICPSD SDG AI Lab established an advanced
data analysis system for two languages and developed a web application. This SupTech
UNDP
solution can analyse daily social media posts in Arabic and French. Moreover, the SupTech
web application aggregates the analysis into PDF-ready reports and visualisations. The
users can listen to specific topics or data sources and produce visualised reports.
This project supports the country’s digital transformation and experiments with digital
technologies for financial supervision and innovative digital finance. It aims to identify
vulnerable social groups' needs and amplify the collective voice of consumers and
highlight the concerns of the most marginalised people. The developed Supervisory
Technology (SupTech) solution offers insights on financial stability, inclusion, and market
conduct. The supervisory technologies will also advance Morocco’s digital transformation,
while generating insight into the local digital economy.
• Project Type/Output: Data Analysis Pipeline, Web Platform and reports
• Project Domain: Economic Growth, FinTech
• Data Source: Social media platforms (Twitter, Facebook, Instagram)
• Technology/Platform: Python, Django
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, SDG 8 - Decent
Work and Economic Growth, SDG 9 - Industry, Innovation, & Infrastructure, SDG 10 -
Reduced Inequalities, SDG 16 - Peace, Justice & Strong Institutions, SDG 17 - Partnerships
for the Goals
• Partnership(s)/Collaborator(s): UNDP Morocco Accelerator Lab, Morocco Central Bank
• Links and Multimedia: https://www.undp.org/fr/morocco/blog/ais-next-frontier-making
-finance-more-sdgs-aligned
Project 31: OSDG

• Project Description: The OSDG is a partnership between PPMI and UNDP Istanbul
International Centre for Private Sector in Development (ICPSD) SDG AI Lab. The partners
joined forces to streamline efforts by consolidating existing research and to foster
transparency by making its methodology and code publicly accessible. The partnership’s
flagship product is OSDG, a free, fully open-source, and user-friendly tool that classifies
text and publications into the different SDGs. This innovative tool is designed to assist
various stakeholders, including governments, companies, development organisations,
universities, research funders, and individual researchers, in aligning their activities with
the Sustainable Development Goals (SDGs). The OSDG has been recognized as one of
Top 100 AI Solutions for SDGs by UNESCO IRCAI.
Under the OSDG initiative, the partners have established the largest broadest possible
consultation on the SDGs monitoring. To do so, the OSDG team has mobilised over 2500
online volunteers from 130 countries, which includes 900+ UN Online Volunteers. Online
volunteer citizen scientists assess the relevance of various text excerpts to SDGs, bringing
together the knowledge and expertise to better understand the nature of SDGs. With the
volunteers’ contributions PPMI and ICPSD SDG AI Lab released nine quarterly versions
of the OSDG Community Dataset with a total of 7,901 views and almost 5000 downloads
141
from over 40 countries. The dataset has been used in academic research (15+ academic
publications) and several hackathons.
• Project Type/Output: Online SDGs classification tool, API, OSDG Community Dataset
• Project Domain: Sustainable Development
• Data Source: Original Dataset
• Related Sustainable Development Goals (SDGs): SDG 4 - Quality Education, SDG 9 -
Industry, Innovation, & Infrastructure, SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): European Research Center - PPMI
• Links and Multimedia: https://osdg.ai/ , https://www.undp.org/news/osdg-initiative
-recognized-top-100-ai-projects-advancing-sustainable-development-goals
Project 32: AI for Tourism

• Project Description: Tourism stands as a priority for sustainable development in both
Malawi and Zanzibar, Tanzania. However, both regions encounter challenges such as
inadequate infrastructure, low-quality service, and ineffective marketing, leading to a lower
ranking in the Travel & Tourism Competitiveness Index compared to their counterparts.
Addressing these issues involves employing AI for data gathering and analysis to identify
trends and insights, facilitating informed decision-making and improving tourism
development, visitor experiences, and revenue generation.
Ai for Tourism Project offers a path to a greater understanding of the tourist industry
using advanced digital technologies such as Artificial Intelligence, Natural Language
Processing, GIS, and others. SDG AI Lab will develop an interactive web platform with
real-time data to help businesses understand the most popular destinations, preferred
types of accommodations, and activities that tourists enjoy. By using this information to
make informed decisions about their services and offerings, tourism businesses will be
able to improve customer satisfaction and generate more revenue.
Additionally, for predictive analytics the project uses AI algorithms to analyse past
customer behaviour data and forecast future trends, such as popular destinations and
services in high demand. Supported with this information, tourism businesses will be
able to better plan and prepare for the future, effectively meeting customer needs and
expectations. Overall, the partnership supports rising tourist destinations with innovative
open-source solutions, which can be adopted by many countries.
• Project Type/Output: Data Pipeline, Web Platform
• Project Domain: Economic Growth
• Data Source: Social Media Data
142
• Related Sustainable Development Goals (SDGs): SDG 1 - No Poverty, 8 - Decent Work and
Economic Growth,SDG 9 - Industry, Innovation, & Infrastructure, SDG 17 - Partnerships
for the Goals
UNDP
• Partnership(s)/Collaborator(s): UNDP Malawi Accelerator Lab, UNDP Tanzania Accelerator
Lab
Project 33: Illegal Dumpsite Detection Guatemala

• Project Description: In Guatemala, the implementation of participatory actions in the
integrated management of solid waste at the municipal level is listed as one of the
strategic goals of the National Development Plan and the Country Priorities. The Ministry
of Environment and Natural Resources (MARN) and UNDP Guatemala are focusing on
the establishment of baselines and indicators in areas of biological importance in fluvial
and marine-coastal zones, such as the Motagua river (the largest in the country). The
river carries a high volume of waste to the Atlantic Ocean that has created a geopolitical
crisis with the neighbouring country Honduras. Illegal waste dumping undermines
efforts to protect the environment and contaminates soil, water resources, oceans, and
ecosystems. It poses serious environmental and health risks. It is necessary to identify
the illegal dumping areas and initiate action that would prevent further environmental
and health risks. Proper identification and monitoring of illegal waste dumping sites
with methods such as regular inspection, reporting, and on-field surveillance require a
considerable amount of budget on a local and national level. Remote sensing and other
digital technologies offer an alternative approach to identify and monitor these areas in
a timely and cost-effective manner.
UNDP ICPSD SDG AI Lab partnered with UNDP Guatemala Accelerator Lab for identifying
the illegal waste dumping sites, particularly alongside the river basins. To do so, SDG
AI Lab utilised cutting-edge tools and techniques like machine learning and artificial
intelligence combined with advanced GIS and remote sensing techniques. The project
is planned for further scaling up and handover to national authorities. The obtained data
will complement official statistics, particularly the National Information System for Climate
Change. SDG AI Lab combined publicly available high-resolution satellite images, very
high-resolution satellite images originating from MAXAR, Google Earth and Sentinel
Hub (under ESA sponsorship) with confirmed illegal dumpsites’ locations. The obtained
data was used to train the machine learning and deep learning models that help in the
detection of new locations and estimation of volume and structure of the waste. After the
calculations and visualisation efforts are completed, the partners will conduct trainings to
localise the capacity and ensure the sustainability of the initiative.
• Project Type/Output: Image classification model for detecting illegal dumping sites and
Technical Report
• Technology/Platform: Python, Google Maps Static API, GIS, Remote Sensing, Computer
Vision
• Related Sustainable Development Goals (SDGs): SDG 6 - Clean Water & Sanitation, SDG
9 - Industry, Innovation, & Infrastructure, SDG 11 - Sustainable Cities & Communities,
143
SDG 13 - Climate Action, SDG 14 - Life Below Water, SDG 15 - Life on land, SDG 17 -
• Partnership(s)/Collaborator(s): UNDP Guatemala Accelerator Lab, Ministry of Environment
and Natural Resources (MARN), Polytechnic University of Milan
Project 34: AI-powered Data Management and Analysis for SDG Finance
• Project Description: Integrated National Financing Frameworks (INFFs) help countries
finance their national sustainable development objectives and the Sustainable
Development Goals (SDGs). Through INFFs, countries develop a strategy to mobilise and
align financing with all dimensions of sustainability, broaden participation in the design,
delivery and monitoring of financing policies, and manage risk. Globally, there are over
85 countries currently using INFFs to build a more sustainable financial architecture at
the national level. With an increasing number of countries moving forward in the INFF
process, the INFF Facility team identified the need for an efficient system to manage,
collect, and analyse data related to the reforms being put forward as part of countries’
INFF processes, both as part of their financing strategies and development finance
assessments.
INFF has partnered with ICPSD SDG AI Lab to obtain strategic insights into national-
level financing solutions and help understand how different countries are proposing to
enhance the mobilisation and alignment of financing with the SDGs. Moreover, it helps
identify trends and inform strategic planning by INFF Facility partners and decision-
making around the supply of relevant technical assistance. To visualise the findings, the
Lab has prepared an interactive PowerBI dashboard. The partners also have worked on
creating a comprehensive and standardised taxonomy on various financial categories.
The established taxonomy supported taxonomy-based analysis, which allowed careful
validation of the findings. Moreover, the taxonomy allowed the creation of an invaluable
labelled dataset for future machine-learning applications.
• Project Type/Output: Data pipeline, PowerBI dashboard, taxonomy
• Project Domain: SDG Finance
• Data Source: INFF Reports
• Technology/Platform: Python, PowerBI
• Related Sustainable Development Goals (SDGs): SDG 16 - Peace, Justice & Strong
Institutions, SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): Integrated National Financing Framework (INFF)
Project 35: INFORM-Warning

• Project Description: INFORM is a multi-stakeholder forum for developing shared,
quantitative analysis relevant to humanitarian crises and disasters. INFORM is developing
a suite of quantitative, analytical products to support decision-making at different stages
of the crisis management cycle.
144
INFORM-Warning seeks to make a contribution to the better use of data to anticipate,

prevent, and respond to climate and complex risks in fragile and crisis-affected settings.
The objective of INFORM-Warning is to present reliable, quantified, multi- hazard
information that warns about risk trends, forecasts, scenarios and events that could lead
UNDP
to crisis impacts in the next 12 months and can be easily used to support decisions on
preparedness and anticipatory action. The INFORM Warning system will be unique as
an open source aggregator of existing early warning data sources that simplifies and
interprets quantitative data to understand early warning trends.
• Project Type/Output: Interactive website, models, scientific paper
• Project Domain: Climate, crisis
• Data Source: Multiple open source datasets
• Technology/Platform: N/A
• Related Sustainable Development Goals (SDGs): SDG 13 - Climate Action, SDG 16 -
Peace, Justice & Strong Institutions, SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): OCHA, JRC, WFP, UNICEF, UNHCR, FAO, IOM, WHO,
IFRC, ACAPS, EU, OECD, FCDO
• Contact information: Sun-joo Lee (sun-joo.lee@undp.org )
Project 36: SDG interlinkages report

• Project Description: Development of a report in 2024 that will look into the interlinkages
between SDG16 and SDG13. The research team plans to test an AI tool for research.
• Department/Division: UNDP Global Policy Centre for Governance
• Project Type/Output: Research to be facilitated by an AI tool
• Project Domain: Governance
• Data Source: Not known yet
• Related Sustainable Development Goals (SDGs): SDG 13 - Climate Action, SDG 16 -
Peace, Justice & Strong Institutions
• Partnership(s)/Collaborator(s): UNDP GPCG is working with IDOS to jointly develop this
publication.
• Contact information: Julia Kercher (julia.kercher@undp.org)
Project 37: UN “HR Policies Interagency” Gen AI Chat Website

• Project Description: To design & build production grade AI chatbot website based on
Azure OpenAI for HR
145
Inter Agency to enable them to automate time consuming activities related to

administration of
HR Policies. The Chat Website provides the ability to benchmark responses against the
complex HR policies of different UN entities.
The provision of the UN “HR Policies Interagency” Gen AI Chat Website to UNDP and 12
other UN entities is managed by United Nations International Computing Centre (UNICC).
UNDP is part of the HR Network that has outsourced the development, hosting, and
support of this solution to UNICC.
• Department/Division: BMS/Office of Human Resources
• Project Type/Output: Chatbot Website
• Project Domain: HR
• Data Source: HR policies of various UN entities
• Technology/Platform: Azure Open AI
• Reported as part of 2022 Compendium on UN AI Activities?
• Related Sustainable Development Goals (SDGs): SDG 17 - Partnerships for the Goals
• Partnership(s)/Collaborator(s): Collaboration among 13 UN entities
• Contact information: Yuichi Kawamoto, yuichi.kawamoto@undp.org
Project 38: NBSAP Target Similarity Assessment

• Project Description: The Convention on Biological Diversity (CBD) aims to guide action to
conserve biodiversity, promote the sustainable use of its components, and ensure the fair
and equitable sharing of the benefits arising from the use of genetic resources. National
Biodiversity Strategies and Action Plans (NBSAPs) are the main policy instrument for
implementation of the CBD at the national level and contain national biodiversity targets
that contribute towards the Convention. At the 15th CBD Conference of Parties (COP15)
in 2022, 196 nations adopted the Kunming-Montreal Global Biodiversity Framework to
put nature on a path to recovery by 2030 and achieve harmony with nature by 2050.
Parties are initiating processes to align the national biodiversity targets within their
NBSAPs through an inclusive whole-of-government and whole-of-society approach by the
COP16. To improve alignment of NBSAP alignment with the Kunming-Montreal Global
Biodiversity Framework, it is essential to identify which elements of the global goals and
targets can be more effectively incorporated into national targets to reduce gaps.
The GEF Early Action Support Project is exploring new applications of Artificial Intelligence
to support governments to more quickly advance in uncovering patterns and gaps in
national biodiversity target alignment using a whole-of-government and whole-of-society
approach. The goal is to provide individual assessments for 69-support countries on the
similarity between a country’s national biodiversity targets and the four goals and 23
targets of the Kunming-Montreal Global Biodiversity Framework. The results offer detailed
guidance that can inform country-led processes to strengthen alignment between national
biodiversity targets with those in the Kunming-Montreal Global Biodiversity Framework
and support NBSAP revision or update. Developed through the application of GPT-4,
these customized assessments have the potential to:
o Enable stakeholder engagement to convene around an informed neutral starting

point;
o Support more dynamic and effective engagement by providing a common starting
place for advanced conservation among diverse stakeholders;
146
o Expedite the alignment processes so that resources can be conserved for other
aspects of NBSAP review, update, and implementation;
o Support activities towards the GEF Global Biodiversity Framework Early Action Support
UNDP
Project; and
o Provide useful information towards CBD reporting requirements in advance of COP16.
• Department/Division: UNDP Nature for Development, GEF Global Biodiversity Framework

Early Action Support Project
• Project Type/Output: Customized policy summaries and in-depth analysis for countries
• Project Domain: Biodiversity
• Data Source: National biodiversity targets are found via the CBD website: https://www
.cbd.int/nbsap/targets/
• Technology/Platform: Azure OpenAI Service
• Related Sustainable Development Goals (SDGs): 1. No Poverty, 2. Zero Hunger, 5. Gender
Equality, 6. Clean Water & Sanitation, 7. Affordable & Clean Energy, 8. Decent Work and
Economic Growth, 10. Reduced Inequalities, 11. Sustainable Cities & Communities, 12.
Responsible Consumption & Production, 13. Climate Action, 14. Life Below Water, 15.
Life on land
• Partnership(s)/Collaborator(s): To date, feedback and recommendations on how to
improve the assessments have been provided by the CBD, UNDP Chief Digital Office,
UNDP Information and Technology Management, UNDP country offices in Cambodia,
Costa Rica, Lebanon, Nepal, and the National Commission for Knowledge and Use of
Biodiversity of Mexico.
• Links and Multimedia: https://www.youtube.com/watch?v=9
UV3yFRrX4M&f eature=y outu
.be https://www.learningfornature.org/wp-content/uploads/2020/07/Target-Similarity
-Assessment-Tool-Presentation-GM-final.pdf
• Contact information: Lea Phillips (lea.phillips@undp.org), Christina Supples (christina.
supples@undp.org), Nicole DeSantis (nicole.desantis@undp.org)
Project 39: Brazil Justica 4.0
By July 2023, the “Justiça 4.0” Program launched four new Artificial Intelligence models for
classifying court cases, contributing to the reduction of trial time and the number of appeals
in the Brazilian courts. All models were developed in partnership with Federal Universities as
follows:
• PEDRO (Court Precedents Extraction and Discovery Platform): Developed in partnership

with the University of Brasília (UnB), the model infers, from the initial petitions, the qualified
precedents of the Federal Supreme Court (STF) and the Superior Court of Justice (STJ),
also showing the degree of similarity with the analyzed case.
• KAIROS (K-Means Clustering Similarity for Legal Documents): Groups processes using
an unsupervised machine learning method and can be used to manage the collection
of processes and classify them by subject, branch, among others. It is the result of a
partnership with the Federal University of Goiás (UFG).
• ALICIA (Linguistic Analysis for Automated Classification): Developed in partnership with
the University of Fortaleza (Unifor), it promotes assignment of subjects based on Unified
Procedural Tables.
147
• ANA (Automated Standards Analysis): Model trained for retrieving legal texts and
associated jurisprudential precedents, developed with Unifor.
• Contact information: moema.freire@undp.org
Project 40: Argentina AI - Argentine Intelligence
The project advances AI for the development agenda. It comprises a series of initiatives to
foster a national dialogue on AI and its potential for development and raising awareness about
it among the population and relevant stakeholders.
The portfolio of initiatives includes the identification of the ecosystem of referents and researchers
in the country; the exploration of uses, good practices, risks, and opportunities of this tool;
the analysis of global and regional regulations to formulate guidelines and recommendations
for the country; and the use of AI as a tool to tackle political polarisation in the context of the
country's 40th anniversary of democracy.
Contact information: lorena.moscovich@undp.org

All 17 SDGs
3. Relevant links
https://www.undp.org/
Contact Information
General: Natalie Samarasinghe (natalie.samarasinghe@undp.org)
China: Yuwei Dai (yuwei.dai@undp.org)
148
United Nations Department of Political and Peacebuilding

Affairs and Department of Peace Operations
UNDPPA
Project 1: AI -powered large-scale synchronist dialogues

• Project Description: Complementing other efforts to address the challenges of
systematically involving public voices in the specifics of peace negotiations, the Innovation
Cell has been exploring the use of Artificial Intelligence (AI) for mediators and actors
to hold real-time consultations with a large group of individuals in local dialects and
languages, allowing for analyses and segmentation based on demographic interests.  
• Department/Division: DPPA – Policy and Mediation Division – Innovation Cell
• Project Type/Output: Software tool, Design Methodology
• Project Domain: Human Rights, Gender, Justice
• Data Source: Anonymous demographic data and insights from participants to the AI-
enabled digital dialogue.  
• Technology/Platform: Interface with external software.  
o Following the successful pilot of the Artificial Intelligence (AI)-assisted large-scale

public dialogues in June 2020 with the Office of the Special Envoy of the Secretary-
General for Yemen (OSESGY) on challenges and opportunities brought on by
COVID-19 pandemic on humanitarian and economic issues, the DPPA Innovation Cell
continued its support of large-scale digital dialogues for inclusive peacemaking. For
instance, towards the end of 2020, the Innovation Cell supported UN Support Mission
in Libya (UNSMIL) in designing and deploying the methodology in Libyan dialect on
two separate occasions related to the UN-led Libyan Political Dialogue Forum (LPDF).
The first digital dialogue, which took place on 16 October with 1000 Libyan youth,
aimed to help set the agenda for a subsequent meeting with youth groups under
the LPDF. Based on the experience and results of the October dialogue, UNSMIL
chose to apply the methodology for a consultation with the wider Libyan public (1500
individuals) on political, military, and economic issues in early November 2020. In
continuing support of UNSMIL, the Innovation Cell developed an online dialogue
platform (Alhiwar.ly) and prepared for a nationwide poll using Computer-Assisted
Telephone Interviews (CATI).
o In 2021, the Innovation Cell advanced partnerships with field missions, such as most
recently with the United Nations Assistance Mission for Iraq (UNAMI) as part of its larger
efforts in providing technical support to the upcoming early elections in October. This
149
ongoing initiative with UNAMI also incorporates behavioral insight to test dialogue
engagement nudges.
o Additionally, the Innovation Cell completed the building of five different dialect
dictionaries (or corpora) – Yemeni, Libyan, Iraqi, Palestinian and Sudanese Arabic.
This represents a major step forward in training computers to understand what is
being said in different dialects. These corpora can now be leveraged to support the
relevant UN Special Political Missions (SPMs) and presences conduct more AI-enabled
dialogues, in addition to helping them parse and better understand radio, TV, and
online content in the aforementioned dialects.

Infrastructure; SDG 16 – Peace, Justice, and Strong Institutions
• Partnership(s)/Collaborator(s):  
o Private Sector: Remesh.AI
o https://futuringpeace.org/ai-for-peacemaking.html
o https://osesgy.unmissions.org/cutting-edge-tech-service-inclusive-peace-yemen
o https://unsmil.unmissions.org/asrsg-williams-conducts-digital-dialogue-1000-libyans
o https://iraq.un.org/en/144266-srsg-jeanine-hennis-plasschaert-conducts-first-digital
-dialogue-iraqi-voters
• Lessons Learned: In addition to substantial lessons learned, such as the development of

the discussion guide and importance of baseline study suitable to the context, the issue
of trust in AI, security and data gathering continues to be a recurring theme. Although
the insights gathered by the tool itself are anonymous, participants in the dialogue have
expressed a general sense of declining trust in digital tools, especially when discussing
formal political processes. The Innovation Cell continues to consider more confidence-
building measures, including by working through partnership with local groups.  
• Contact information: Martin Waehlisch (martin.waehlisch@un.org), Daanish Masood
(masoodd@un.org)
Project 2: Exploring the Potential and Risks of Generative Artificial Intelligence in

the UN System
• Project Description: As Generative Artificial Intelligence technologies progress, they offer
numerous advantages, such as improved knowledge management, enhanced translation
capabilities, and efficient content generation. Especially with the release of AI applications
such as ChatGPT, Perplexity.ai, and Microsoft Bing for general use, Generative AI
continues to permeate the tools and applications used in the UN’s daily work. To foster
greater awareness and capabilities among UN staff regarding both the potential benefits
and risks of this technology, DPPA’s Innovation Cell organizes AI Innovation Workshops,
including prompt engineering labs and other creative ideation sessions.
• Department/Division: DPPA – Policy and Mediation Division – Innovation Cell
• Project Type/Output: AI Innovation Workshops
• Project Domain: Generative AI
• Data Source: Insights from participants from across the UN system
• Technology/Platform: ChatGPT, Perplexity.ai, Casper.AI
150
o As an example, on 3 May 2023, DPPA’s Innovation Cell, in collaboration with EOSG's
UNDPPA
Data Strategy Team and the United Nations Innovation Network (UNIN), organized a
brainstorming workshop on the potential and risks of Generative Artificial Intelligence
in the UN System with over 40 participants from across the organization. The discussion
revealed that the main areas where AI is already applied in the UN system include: 1)
content creation and editing, 2) learning and analysis, 3) communications, 4) accessibility
and assistive technologies, and 5) task automation and efficiency. In particular, AI is
being tested for processing massive amounts of sitreps and producing automated
summaries, used for SWOT analysis, assessing project timelines, and much more.
With regard to future potential, the group highlighted a number of ideas in similar
categories, such as insight extraction from planning and reporting documents,
developing chatbots for knowledge bases, predictive modeling, and AI for auditing.
The group also discussed concerns regarding data security and organizational
challenges and outlined several potential priorities, such as tailoring Generative AI
application to the UN Secretariat.

Infrastructure; SDG 16 – Peace, Justice, and Strong Institutions
• Partnership(s)/Collaborator(s): EOSG's Data Strategy Team and the United Nations
Innovation Network (UNIN)
o https://futuringpeace.org/project/innovation-workshop-on-artificial-intelligence
o https://f uturingpeace. org/p roject/g enerative- artificial- intelligence- and- its
-implications
• Lessons Learned: Advancing Generative AI applications tailored to the UN Secretariat

requires different levels of collaboration between various parts of the UN. This emphasizes
the importance of training and sensitizing management and staff regarding the potential
and the risks of AI. Participants welcomed the workshops as providing an important forum
for discussion and expressed interest in further developing and maintaining it to facilitate
exchange and practice amongst different parts of the UN in the area of AI.
• Contact information: Martin Waehlisch (martin.waehlisch@un.org), Min Ji Song (min.
song@un.org)

SDG 9 and 16
3. Relevant Links
www.dppa.un.org
Contact Information
Martin Waehlisch (martin.waehlisch@un.org), Daanish Masood (masoodd@un.org), Min Ji

Song (min.song@un.org)
151
United Nations Economic Commission for Europe
Project 1: UK Office of National Statistics (ONS) – UNECE Machine Learning Group

2022
• Project Description: national and international statistical organisations produce official
statistics that affects important policies and decisions for the society and economy. ML
holds a great potential for statistical organisations to harness new data sources (e.g.,
big data) and make their business more efficient, allowing them to provide better data
services. The ONS-UNECE Machine Learning Group 2022 (ML 2022) is an initiative
targeted at the official statistics community, providing a platform for the global statistical
community to develop research, build skills and share common challenges and solutions
on ML developments and applications. It builds on the momentum of the 2019-2020
UNECE HLG-MOS Machine Learning Project and the ML Group 2021. The initiative was
launched in January 2022, is led by the ONS Data Science Campus and the UNECE. ML
2022 consists of theme groups focusing on following topics text classification, modelling,
imagery analysis, quality of training data, model re-training, web-scrapping data and IT
infrastructure.
• Department/Division: Statistics
• Project Type/Output: Community of expertise, research collaboration
• Project Domain: Statistics
• Data Source: Survey data, register data, automatic identification sysem (AIS) data, web-
scraped data, satellite imagery data (LandSat, Sentinel), etc.
• Data Publicly available: No
• Project Updates: the results from the ML Project 2019-20 and ML Group 2021 have been
compiled as an UNECE publication “Machine Learning for Official Statistics”: https://
unece.org/statistics/publications/machine-learning-official-statistics
o Government: ONS (United Kingdom Office for National Statistics) Data Science
Campus is the main coordinating partner of the ML2022 Group, but theme groups
are led by different organisations such as Statistics Netherlands, Statistics Flanders,
Statistics Sweden, Statistics Ireland and Norwegian School of Economics and Center
for Applied Research; the group itself consists of more than 380 members from various
national and international statistical organizations
• Relevant Links and Multimedia: https://statswiki.unece.org/display/ML/Machine+

Learning+Group+2022
152
• Contact information: InKyung Choi (choii@un.org)
Project 2: Functional Requirements for Automated Vehicles
UNECE
• Project Description: The Group dealing on Functional Requirements for Automated
Vehicles is led by China (MIIT), Germany (KBA) and the United States of America (NHTSA).
It reports to the Working Party on Automated/Autonomous and Connected vehicles of
UNECE. It works on safety requirements covering all Automated Driving Systems (ADS)
configurations (SAE Levels 3-5). These activities at intergovernmental level form a novel
initiative aimed at harmonizing globally automated vehicles regulations and creating a
more productive environment for innovation.
• Work assumptions: to Improve road transport, Performance-based, Technology-neutral,
Measurable, Feasible, and Socially acceptable.
• Entity Name: WP.29/GRVA
• Project Type/Output: Seminar/meeting; Policy Framework; Intergovernmental working
group dealing with technical regulations for vehicles.
• Project Status: Ongoing, initial deliverables already in force.
• End Year: 2024
• Project Domain: Automotive
• Technology/Platform: performance based, technology neutral
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Wellbeing;
SDG 9 – Industry, Innovation and Infrastructure; SDG 11 – Sustainable Cities and
Communities
o UN Partners: All UN Agencies may participate

o Government: All UN member States
o Private Sector: Represented through NGO with ECOSOC accreditation
o Civil Society: Represented through NGO with ECOSOC accreditation
o Academia: In support of the national delegations or upon invitation
o Press release on adoption of the Framework Document on Automated vehicles:

https://unece.org/media/press/1610
o Press release on adoption of the Level 3 UN Regulation No. 157: https://unece.org/
media/press/348314
o Press release on extending UN Regulation No. 157 to trucks and coaches:
https://unece.org/media/press/362551
o Press release on adoption of amendment extending automated driving to up to 130
km/h: https://unece.org/media/press/368227
o Meetings and documents: https://unece.org/transport/vehicle-regulations/working
-party-automatedautonomous-and-connected-vehicles-introduction
• Contact information: Francois E. Guichard (francois.guichard@un.org)
Project 3: Validation Method for Automated Driving

• Project Description: The Group on Validation Method for Automated Driving is led by
Canada (Transport Canada), The Netherlands (RDW) and Japan (NTSEL). It reports to the
Working Party on Automated/Autonomous and Connected vehicles of UNECE. It includes
153
four subgroups on (a) Scenarios, (b) Simulation, (c) Audit and Monitoring, and (d) Track
Test and Real-world Test.
It deals with the validations methods leading to the demonstration of a robust design and
validation process based on a systems-engineering approach with the goal of designing
automated driving systems free of unreasonable safety risks and ensuring compliance
with road traffic regulations [..]. Design and validation methods should include a hazard
analysis and safety risk assessment for Automated Driving System (ADS), for the Object
and Event Detection and Response, but also for the overall vehicle design into which
it is being integrated and when applicable, for the broader transportation ecosystem.
Design and validation methods should demonstrate the behavioural competencies
an Automated/autonomous vehicle would be expected to perform during a normal
operation, the performance during crash avoidance situations and the performance of
fall-back strategies. Test approaches may include a combination of simulation, test track,
and on-road testing.
• Entity Name: WP.29/GRVA
• Project Type/Output: Policy Framework; Seminar/meeting; Intergovernmental working
group dealing with technical regulations for vehicles.
• Project Domain: Automotive
• Technology/Platform: performance based, technology neutral
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Well-
Being; SDG 9 – Industry, Innovation, and Infrastructure; SDG 11 – Sustainable Cities and
Communities
o UN Partners: All UN Agencies may participate.

o Government: All UN member States
o Private Sector: Represented through NGO with ECOSOC accreditation
o Civil Society: Represented through NGO with ECOSOC accreditation
o Academia: In support of the national delegations or upon invitation
• Relevant Links and Multimedia: https://unece.org/reference-documents-0

• Contact information: Francois E. Guichard (francois.guichard@un.org)
Project 4: Task Force on Digitalization in Energy

• Project Description: Technologies facilitating new market opportunities: digital innovations
– tools, technologies and processes (such as Artificial Intelligence, Blockchain, Machine
Learning, Advanced Data Analytics, Internet-of-Things, Big Data, Cloud Computing,
Sensors, Automation, 3D Printing, Robotics, etc.), are inspiring energy suppliers,
transmission and distribution companies, and demand sectors (buildings, industry,
transport) to establish new business models allowing to generate, deliver and consume
energy in a more sustainable fashion. These innovative technologies are providing new
opportunities to businesses by changing the way how interaction happens, optimizing
processes, enhancing flexibilities, and improving efficiencies.
The Task Force on Digitalization in Energy was established by the Committee on
Sustainable Energy in 2020 with the mandate for the period of 2021-2022, which was
further extended for 2023-2024 with a possibility of extension. The Task Force on
Digitalization in Energy reports to the Group of Experts on Energy Efficiency. The Task
Force provides a platform for cross-industry experts from the energy sector (including
154
from the other subsidiary bodies of the Committee on Sustainable Energy) and digital
innovations to develop a unified voice on digitalization in energy. It critically explores
the landscape of stakeholders through a constructive dialogue (including assessment
UNECE
of opportunities, challenges, risks, and trade-offs) to understand the interaction in the
digitalized energy system and bringing consensus about the approach that should be
considered for shaping the future of the energy system. Activities of the Task Force
also include aggregating and reviewing the existing national policy initiatives as well as
harmonizing the information produced by other key national and international bodies, in
order to better assist policymakers and other stakeholders in UNECE region to provide
evidence-based direction to achieve the higher levels of efficiency in the energy system.
• Department/Division: UNECE Sustainable Energy Division
• Project Type/Output: Conference(multiple meetings on a theme), papers
• Data Source: Publicly available data sources.
Communities; SDG 12 – Responsible Consumption and Production; SDG 17 – Partnership
for the Goals
o UN Partners: United Nations Development Programme, United Nations Economic

and Social Commission for Western Asia, UNEP-CCC Copenhagen Climate Centre
o Government: Swiss Federal Office of Energy, Ministry of the Environment (France),
Government of the United Kingdom, National Energy Efficiency Centre (Russian
Federation)
o Private Sector: The Energy Authority, Energy Policy Group, Sustainable Decisions, North
American Electric Reliability Corporation, DigiTransfo Expertise, Grid Singularity, Irys,
Google X
o Academia: International Energy Research Centre, Vector Institute, Harvard Kennedy
School of Government, University of Bayreuth, Institute for Energy Efficiency in
Production (EEP), Green Growth Knowledge Partnership (Green Industry Platform)
• Relevant Links and Multimedia: https://unece.org/sustainable-energy/energy-efficiency/

digitalization-energy
o https://unece.org/sites/default/files/2020-12/GEEE-7.2020.INF_.3.pdf
o https://unece.org/sed/documents/2021/06/working-documents/improving-efficiency
-buildings-through-digitalization
o https://u nece. org/s ed/d ocuments/2 022/0 7/s ession- documents/d igitalization
-accelerating-electricity-system
o https://u nece. org/s ed/d ocuments/2 022/0 7/s ession- documents/a ddressing
-behavioural-barriers-energy-digitalization
o https://unece.org/sed/documents/2022/07/informal-documents/policy-discussion
-challenges-big-data-and-analytics-driven
o https://unece.org/sed/documents/2023/08/working-documents/key-considerations
-and-solutions-ensure-cyber-resiliency-0
o https://unece.org/sed/documents/2023/08/working-documents/improving-efficiency
-and-reliability-energy-systems-means-1
155
o https://unece.org/sustainable-energy/publications/digitalization-energy-case-study
-grid-edge-management-reference
o https://unece.org/sustainable-energy/publications/digitalization-energy-case-study
-cyber-resilience-critical-energy
• Contact information: Igor LITVINYUK (litvinyuk@un.org)
Project 5: Improving efficiency and reliability of energy systems by means of Big

Data analytics
• Project Description: The energy sector has experienced a shift towards disruptive
trends such as decarbonization, decentralization and digitalization, fuelling the energy
transition that creates a major impact on the utility industry worldwide. These ambitions
drive the imperative for Artificial Intelligence in general and more specifically, Data
Analytics. Deployment of Big Data analytics, machine learning, and Artificial Intelligence
in utilities and energy providers is growing at a rate that may outpace the maturity of the
organizations. In fact, organizations may have already engaged in advanced algorithmic
deployment. Yet, without a strategy, organizations may not have developed workable
plans for the curation of data, training datasets, or analytics results. This document is
the result of the collaborative work of the product of the Group of Experts on Cleaner
Electricity Systems and the Group of Experts on Energy Efficiency, conducted on the
platform of the Task Force on Digitalization in Energy. It is an expansion of the unofficial
document GEEE-9/2022/INF.3 entitled “Policy discussion – Challenges of Big Data and
analytics-driven demand-side management” that contains the original set of questions
posed for further consideration by the Task Force on Digitalization in Energy in
collaboration with the subsidiary bodies of the Committee on Sustainable Energy. .
• Project Type/Output: Official document
• Project Status: Complete (presented at the tenthsession of the Group of Experts on
Energy Efficiency)
• Project Domain: Energy; Telecommunications
• Related Sustainable Development Goals (SDGs) SDG 7 – Affordable and Clean Energy;
SDG 9 – Industry, Innovation and Infrastructure; SDG 10 – Reduced Inequalities; SDG
11 – Sustainable Cities and Communities; SDG 12 – Responsible Consumption and
Production; SDG 17 –Partnerships for the Goals
o UN Partners: UNEP-CCC Copenhagen Climate Centre,

o Government: Swiss Federal Office of Energy
o Private Sector: The Energy Authority, Energy Policy Group, DigiTransfo Expertise, Grid
Singularity, Irys
o Academia: N/A
• Relevant Links and Multimedia: https://unece.org/sed/documents/2023/08/working

-documents/improving-efficiency-and-reliability-energy-systems-means-1
156
Project 6: Key considerations and solutions to ensure cyber resiliency in the smart
integrated energy systems
UNECE
• Project Description: Digitalization, as the application of digital technologies and business
models for existing processes, is gaining more attention as a way to support and
complement the energy transition. However, integration of different energy sources and
interconnection of various energy system components which constitute smart integrated
energy systems, involve exchange of large amounts of data that increases the exposure
to cybersecurity risks. This document was developed on the platform of the Task Force
on Digitalization in Energy, by the Group of Experts on Energy Efficiency and the Group
of Experts on Cleaner Electricity Systems in line with their respective Work Plans for 2022-
2023 and in recognition of the growing information security threats amid increasing
digitalization of the energy system. The Task Force on Digitalization in Energy further
recognizes the importance of collaboration across all subsidiary bodies of the Committee
on Sustainable Energy in the effort to address aspects specific to various elements of the
energy value chain. The document contains an overview of the subject, identifies, and
categorizes the types of cyberattacks on the exposed energy system components and
their possible consequences, and concludes with a set of managerial- and technical-level
measures and policy recommendations to mitigate the cybersecurity risks.
• Project Type/Output: Official document
• Project Status: Complete (presented at the tenth session of the Group of Experts on
Energy Efficiency)

o Private Sector: Sustainable Decisions, The Energy Authority, Energy Policy Group,
DigiTransfo Expertise
o Academia: N/A
• Relevant Links and Multimedia: https://unece.org/sed/documents/2023/08/working

-documents/key-considerations-and-solutions-ensure-cyber-resiliency-0
Project 7: Development of national case studies on digitalization in energy

• Project Description: Development of national case studies on digitalization in energy
that have significant potential of replication, feeding into a compendium of case studies
(publication) in 2024.
• Project Type/Output: Compendium of national case studies (minimum five case studies)
• Project Status: In progress (two case studies published)
157


o Private Sector: Energy Policy Group, Grid Singularity, The Energy Authority
o Academia: N/A
o Case Study on "Grid Edge Management Reference Architecture and Policy

Recommendations for Interoperability and Resilience": https://unece.org/sustainable
-energy/publications/digitalization-energy-case-study-grid-edge-management
-reference
o Case study on “Cyber Resilience of Critical Energy Infrastructure”: https://unece.org/
sustainable-energy/publications/digitalization-energy-case-study-cyber-resilience
-critical-energy
Project 8: Platform for Resilient Energy Systems – chatenergy.app

• Project Description: Countries across the ECE region are in high need of tools to make
informed decisions on how to design and build resilient energy systems (i.e. ones where
energy makes an optimal contribution to a country’s social, economic, and environmental
development and that is able to withstand and recover quickly from any unanticipated
shocks and reflects potential impacts of climate change on energy resources in its planning
and operations). With a mission to mainstream the resiliency concept and systems-
thinking approach in national policies UNECE is shaping a Platform for Resilient Energy
Systems. Its objective is to provide member States and the energy expert community with
an advanced AI-based tool that allows to navigate through a secure, comprehensive and
authoritative knowledge base built by UNECE and partnering organizations (EIB, ESCAP,
IAEA, IEA, IRENA, OSCE, WMO, etc.), producing user-friendly inputs and insights for
informed decisions on how to create more resilient energy systems. The demo-version
of the Platform is called chatenergy.app and has been developed by the team of prof.
M. Leippold from University of Zurich. UNECE is mobilizing funds to a fully operational
AI-enabled query & response platform which can leverage a library composed of energy
documents from authoritative sources and provide real-time, summarized insights, enable
UNECE and the wider community to spot knowledge gaps and trends, provide support
for UNECE member states and beyond in drafting future energy policies and guidelines.
From a knowledge repository with a smart search engine, robust data and documents
library of vetted sources, UNECE seeks to move towards a more advanced scenario
building tool, which will allow to see different options for decision-makers and make
data understandable and actionable. The AI model will be specifically crafted with vetted
data and tailored for the unique requests regarding energy systems resilience.
• Project Type/Output: Software Tool
158
• Project Status: ongoing (launched at the thirty-second session of the Committee on

Sustainable Energy)
UNECE
• Project Domain: Energy; Technology and Applications;
• Data Source: Publicly available data sources from partnering organizations.
Communities; SDG 13 – Climate Action; SDG 17 – Partnerships for the Goals
o EIB, IEA, IRENA, OSCE

o UN Partners: UNICC, IAEA, WMO, ITU, ESCAP, ECLAC
o Academia: University of Zurich
• Relevant Links and Multimedia: https://unece.org/climate-change/press/unece-and

-partners-develop-ai-powered-platform-help-build-resilient-energy
• Contact information: Iva BRKIC (iva.brkic@un.org )
Project 9: Use of artificial intelligence in trade facilitation

• Project Description: Artificial intelligence (AI) is shifting global value chains and
international trade pattern. Artificial intelligence (AI) has a transformative effect on
international trade. Specific applications in multiple areas such as analytics and cognitive
services are diminishing trade barriers.
AI undoubtedly will underpin productivity growth, economic growth and create new
opportunities in facilitating trade
The purpose of this project is to look at AI’s role in trade facilitation in the context of
UN/CEFACT’s mandates and create whitepapers that focus on how AI can be used to
facilitate trade processes and key issues that need to be looked into while leveraging AI
capabilities in collecting, processing, analyzing data and extracting inferences from the
data
Potentially, this work could also provide guidance to data providers, application
developers and technology adopters
• Project scope
The project scope is to define and create white papers on the best practices in
implementing secure data flow during cross border trade with a view to examining:
o How AI technology could be used to facilitate trade and related processes in

international supply chain including study of areas such as data privacy, AI based
trade policies, use of AI in e-Commerce and payments
o How existing UN/CEFACT deliverables could be used in AI applications
o Possible changes to existing UN/CEFACT deliverables, or new deliverables, that could
be considered in order to support AI trade facilitation applications
All of the above will be examined from the perspective of UN/CEFACT’s mandates in
order to provide input to the Bureau, Programme Development Areas and Domains on
Possible future work and a possible common approach
Application developers as a potential new user group for UN/CEFACT standards
• Project deliverables
o Deliverable 1: A white paper on technical aspects of AI and its relation to UN/CEFACT

deliverables
159
o Deliverable 2: A business case/process oriented whitepaper on how AI technology

could be used to facilitate cross border paperless trade
• Department/Division: The United Nations Centre for Trade Facilitation and Electronic
Business (UN/CEFACT)
• Project Type/Output: White Paper
• Project Start Year:2021
• End Year: 2022
• Project Domain: Trade
• Data Source: Initial contributions include existing descriptions and technical specifications
for the UN/CEFACT:
o AGAT document “Artificial Intelligence Demystified"

o Core Components Library (CCL);
o Business Requirement Specifications (BRSs),
o Requirement Specification Mappings (RSMs) and
o Reference Data Models (RDMs) as well as
o already published material on AI technology and implementations,
o Blockchain work undertaken by UN/CEFACT

• Related Sustainable Development Goals (SDGs): SDG 17 – Partnerships for the Goals
o Private Sector: UN/CEFACT Experts
• Relevant Links and Multimedia: https://uncefact.unece.org/display/uncefactpublic/Use+

of+Artificial+Intelligence+in+Trade+Facilitation
• Contact information: (UNECE Secretariat uncefact@un.org)
Project 10: United for Smart Sustainable Cities (U4SSC)

• Project Description: In 2016, UNECE and the International Telecommunication Union
(ITU) established the UN global initiative United for Smart Sustainable Cities (U4SSC),
which currently involves 16 UN bodies. U4SSC is a global platform for smart cities
stakeholders, which advocates for public policies to encourage the use of ICT to facilitate
the transition to smart sustainable cities. The initiative aims to: Generate guidelines,
policies and frameworks for the integration of ICTs into urban operations, based on the
SDGs, international standards and urban key performance indicators (KPIs); and help
streamline smart sustainable cities action plans and establish best practices with feasible
targets that urban development stakeholders are encouraged to meet. The topics of this
phase of U4SSC are: circular cities, financing smart sustainable cities projects, blockchain
in cities, artificial intelligence in cities, sensing technologies and Internet of Things in cities.
The initiative delivers policy guidelines and training materials through the work on specific
outputs elaborated via regular e-meetings and physically gathers once per year. In 2017,
the U4SSC stakeholders also elaborated a set of Key Performance Indicators (KPIs) for
smart sustainable cities which includes 92 indicators (core and advanced) divided in the 3
dimensions of sustainable development: economy, environment, and society and culture.
The indicators are fully aligned with the Sustainable Development Goals (SDGs) and serve
as a tool for evidence-based decision making, progress monitoring and achieving the
SDGs at the local level. They are being implemented by 50 cities of different sizes and
development worldwide.
160
• Department/Division: Housing and Land Management, Forests, Land and Housing

Division
• Project Type/Output: Multi-agency partnership
UNECE
• Project Domain: Sustainable urban development
Communities; SDG 17 – Partnership for the Goals
o UN Partners: UNECE, ITU, UN Habitat, CBD, ECLAC, FAO, UNDP, UNECA, UNESCO,
UNEP, UNEP-FI, UNFCCC, UNIDO, UNOP, UNU-EGOV, UN-Women and WMO
• Relevant Links and Multimedia: https://unece.org/housing/sustainable-smart-cities;

https://twitter.com/UNECEHLM; https://www.facebook.com/UNECE.Housing/
• Contact information: Gulnara Roll (gulnara.roll@un.org)
Project 11: Regulatory compliance of products with embedded artificial intelligence

or other digital technologies
• Project Description: Products with embedded technologies present specific challenges
to their conformity assessment and market surveillance. Regulators need to ensure that
the products are safe and secure (for consumers, for the environment, for public health,
for the culture…) while not stifling innovation. Agencies may approach such products
differently, resulting in technical regulations which may not be harmonized between
countries and may not even be harmonized within a same country. This would in turn
create problems for traders as they may need to adapt their product for multiple markets.
An additional challenge for regulators is the possibility for these types of products to
evolve in time, either because of their own learning or because of distant updates coming
from a remote source potentially outside of the country. Market surveillance authorities
need to ensure that these products remain compliant throughout their lifecycle and do
not receive updates which may render them non-conformant to regulations.
The UNECE-WP6 Recommendation L proposes a policy framework document to
encourage harmonization of such technical regulations through a voluntary mechanism
called a common regulatory arrangement (CRA). The project team will then work on a
“declaration” which countries could ratify on this topic.
• Department/Division: Economic Cooperation and Trade Division, Working Party on
Regulatory Cooperation and Standardization Policies (WP.6)
• Project Type/Output: Policy Framework (a “common regulatory arrangement” as per
UNECE-WP.6 Recommendation L. Declaration that countries could ratify on this topic),
White Paper
• Project End Year: 2024+
• Project Domain: Trade
• Data Source: Concerns technical regulations on products. Will address international
standards and conventions to which such products should remain compliant (see the
white paper already completed).
161
• Project updates:
o Project team have already delivered the White Paper with the base principles, available
at: https://unece.org/trade/documents/2023/12/session-documents/regulatory
-compliance-products-embedded-artificial
o Further information on project advancement available at: https://unece.org/trade/
wp6/digital-regulation-goods-artificial-intelligence
• Related Sustainable Development Goals (SDGs): SDG 5 – Gender Equality, SDG 8 –

Decent Work and Economic Growth, SDG 9 – Industry, Innovation, and Infrastructure,
SDG 10 – Reduced Inequalities, SDG 12 – Responsible Consumption and Production,
SDG 16 – Peace, Justice, and Strong Institutions, SDG 17 – Partnerships for the Goals
• Partnership(s)/Collaborator(s): Note that this project is under UNECE WP.6. All projects
within WP.6 are open to all experts and experts participate in their own capacity without
representing any countries, organizations or specific interests. The resulting deliverables
are presented to UN Member States for endorsement.
o https://unece.org/trade/wp6/digital-regulation-goods-artificial-intelligence
o https://unece.org/trade/wp6/Projects#accordion_4
• Contact information: Lance Thompson (regulatory.cooperation@un.org)

All the SDGs
3. Relevant links
https://www.unece.org/info/ece-homepage.html
Contact Information
Thomas Croll-Knight, Information Unit, United Nations Economic Commission for Europe
(thomas.croll-knight@un.org)
162
United Nations Environment Programme
UNEP
Project 1: Artificial Intelligence tool for valuing the contributions of nature (ARIES
for SEEA)
• Project Description: The project aims to accelerate implementation of the new ground-
breaking standard for valuing the contributions of nature that was adopted by the UN
Statistical Commission in March 2021, the System of Environmental-Economic Accounting
(SEEA) Ecosystem Accounting.
Artificial Intelligence for Environment & Sustainability (ARIES), developed by researchers
at the Basque Centre for Climate Change (BC3), is an integrated, open-source modelling
platform for environmental sustainability, where researchers from across the globe can
add their own data and models to web-based repositories.
Using ARIES technology, the ARIES for SEEA Explorer application allows users anywhere
in the world to produce rapid, standardized, scalable and customizable ecosystem
accounts for their area of interest that are consistent with the SEEA Ecosystem Accounting
framework. ARIES for SEEA is available on the UN Global Platform, a cloud-service
platform supporting international collaboration in the development of official statistics
using new data sources and innovative methods.
The ARIES for SEEA platform is currently being piloted in Botswana, Ghana, Uganda,
Rwanda, Senegal and South Africa, following a workshop in Kigali in July 2022.
• Data Source: freely available global remote-sensing derived data
• Data Publicly available: Yes
• Technology/Platform: ARIES technology; semantics and machine reasoning to automate
data and model integration
• Related Sustainable Development Goals (SDGs): SDG 6 – Clean Water and Sanitation;
SDG 13 – Climate Action; SDG 14 – Life Below Water; SDG 15 – Life on Land
o The ARIES for SEEA Explorer was developed by the Basque Center for Climate
Change (BC3) under the EU-funded Natural Capital Accounting and Valuation of
Ecosystem Services (NCAVES) project, which is jointly implemented by the Statistics
Division of the UN Department of Economic and Social Affairs and UNEP.
• Relevant Links and Multimedia: https://seea.un.org/content/aries-for-seea
163
• Lessons Learned: Functionalities for the ARIES for SEEA explorer will be continuously
improved and expanded
• Contact information: William Speller (william.speller@un.org)
Project 2: Adaptation to climate change in sub-Saharan African humanitarian

situations
• Project Description: The project harnesses artificial intelligence to investigate past
environmental change around selected humanitarian hotspots of displacement in
Burundi, Chad and Sudan and future projections in the face of the climate crisis to
inform climate change adaptation measures and anticipatory action for integration in
humanitarian programming.
• Entity Name: UNEP/OCHA Joint Environment Unit
• Project Type/Output: Report; Academic paper; Seminar/meeting
• Project Start Year: Sept 2018
• Project End Year: April 2021
• Domain: Environment; Humanitarian
• Data Source: Population data, satellite imagery and IPCC climate projections
• Publicly Available Data : No
• Technology/Platform: GIS, MapX
• Related Sustainable Development Goals (SDGs): SDG 13 – Climate Action; SDG 17 –
o UN Partners: WFP, UNHCR, UNICEF, OCHA

o Government: Governments of Burundi, Chad and Sudan
o Academia: University of Ghadarif, Sudan
• Relevant Links and Multimedia: https://eecentre.org/2019/05/01/iki-project/

• Lessons Learned: The project was implemented during the COVID-19 pandemic,
with artificial intelligence proving an effective solution to continue delivering despite
movement restrictions that hampered fieldwork and other project activities requiring
presence
• Contact information: Margherita Fanchiotti (fanchiotti@un.org)
Project 3: Identifying the potential applications of Artificial Intelligence for Disaster

Management.
• Project Description:
Background
Rapid advances in science and practices of Artificial Intelligence (AI), Robotics, Drones and
Internet of Things (IoT) is expected to change all aspects of human society in the coming
decades. Social scientists call this collective change the “Fourth Industrial Revolution”
(4IR). There are many possibilities for introducing these technologies into the domain of
disaster response, especially for environmental disaster response. In addition to helping
mitigate the consequences of disasters, the technologies of the fourth industrial revolution
also help prepare for them.
The proposed project is aimed at the current state of play on the applications of 4IR
technologies for disaster management and identifies opportunities and partners to
promote these technologies for environmental emergency response in the future.
164
For instance, in October 2020, during the AI for Good Summit organized by ITU, UNEP
led a session on Artificial Intelligence for Natural Disaster Management. In addition to
UNEP and ITU, the session had speakers form Google, Government of India and Monash
University. With over 2000 registrations, it was the most well attended session at the AI
UNEP
for Good summit. This led ITU, UNEP and WMO to form a partnership to take this work
forward.
• Scope of Work
o Conduct a comprehensive literature review of the developments in the field of AI,

Robotics and IoT to identify those with potential applications in Environmental Disaster
Response
o Identify technologies which has been developed and employed, even at pilot
phase, for emergency response and prepare a technology forecast for its potential
applications in environmental emergency response
o Identify research and technology partners in academia, research laboratories and
industries who are currently working on the area of applications of 4IR technologies
and form a network for information exchange and potential collaborations
o Prepare training programme on the potential applications of 4IR technologies in
environmental emergency response
• Department/Division: Ecosystems Division, Resilience to Disasters and Conflicts Global

Support Branch.
• Project Type/Output: Report, Seminar/meeting, Conference
• Domain: Environment; Health; Telecommunications; Weather; Technology
• Data Source: We are currently using open source information and qualitative data.
• Data Publicly Available: Yes
Infrastructure; SDG 13 – Climate Action; SDG 17 – Partnerships for the Goals
o UN Partners: International Telecommunications Union (ITU), World

Meteorological Organisation (WMO)
o Webinar series promotion
- UNEP Website
- Other materials
o Video recordings of Modern Technologies for Disaster Management Webinar Series
- Robotics for Disaster management – video

- Artificial Intelligence for Disaster Management – video
• Lesson Learned:
o Future work:
- Continue to identify research and technology partners in academia, research

laboratories and industries who are currently working on the area of applications
165
of 4IR technologies and form a network for information exchange and potential
collaborations
- Prepare training programme on the potential applications of 4IR technologies in
environmental emergency response and deliver such trainings.
• Contact information: Muralee Thummarukudy (muralee.thummarukudy@un.org) and

Paula Padrino Vilela (paula.padrinovilela1@un.org)
Project 4: Global Partnership on Marine Litter (GPML) Recommender

• Project Description: The Global Partnership on Marine Litter (GPML) Digital Platform is
a multi-stakeholder platform that compiles different resources, connects stakeholders, and
integrates data to guide action. Its vision is to be the go-to open source multistakeholder
Platform that successfully connects and informs all actors working towards addressing the
global problem of marine litter and plastic pollution.
The Platform allows stakeholders to get an overview of the different initiatives, find
relevant resources and connect with other stakeholders in the field. To encourage and
make it easier for stakeholders to connect, UNEP would like to implement a matchmaking
system that automatically matches stakeholders based on their interests, skills, and other
relevant information.
Three levels of recommender systems are under consideration.
o Level 1 is a Content-based Recommender Engine. This will match an individual to

an individual and an entity to an entity. For individuals, user profile data is used. For
entities, the data on representative sector, country, geo-coverage type, entity, seeking,
offerings, and bookmarks fields are used. For both, the Geo-coverage is used to
further rank/order recommendations.
o Level 2 is a Content-based Recommender Engine Using Natural Language Processing.
Apart from the predefined fields used in the Level 1 Recommender, this model uses
the “About Yourself” and document uploads i.e., “Curriculum Vitae” features for
matching stakeholders. These fields are natural language data; hence NLP techniques
are beneficial for the design of the recommender engine. Matches are prioritized by
geo-coverage and offerings of the stakeholders.
o Level 3 is a Content-Based Recommender Boosted by Collaborative Filtering
(Interaction Data). This is a hybrid recommender model that treats collaborative
information (interaction data) as additional feature data of the Level 1 or Level 2
recommenders. At this level, the K-means clustering algorithm is used to combine
content-based and collaborative filters. In addition to profile data used by the Level
1/Level 2 recommenders, the system will take in interaction data prioritized as follows;
Time spent on resources. Clicks on a link or resource, and Stakeholder browsing
history.
With more users signing up to the platform, there will be more interaction
data being recorded and stored. The Recommender will not only perform Individual-
Individual matching, but also incorporate Entity-Entity and Entity-Individual matching.
• Department/Division: Ecosystems Division
• Data Source: User Profile Data + innovative data sources for the data hub
• User profile data is collected when individuals or entities sign up to the GPML Digital
Platform.
166
• This data includes information such as what resources a user is offering and what they are
seeking from other users on the GPML digital platform as well as their geo-coverage type
if a user signs up as an entity.
UNEP
o Jupyter Notebook: The Jupyter Notebook App is a server-client application

that allows editing and running notebook documents (documents produced by
the Jupyter Notebook App, which contain both computer code (e.g., python) and
rich text elements (paragraph, equations, figures, links, etc.…)) via a web browser.
The Jupyter Notebook App can be executed on a local desktop requiring no internet
access (as described in this document) or can be installed on a remote server and
accessed through the internet.
o Python: Python is an interpreted high-level general-purpose programming language.
o Natural Language Toolkit (NLTK): a suite of libraries and programs for symbolic and
statistical natural language processing for English written in the Python programming
language.

• Related Sustainable Development Goals (SDGs): SDG 6 – Clean Water and Sanitation;
SDG 11 – Sustainable Cities and Communities; SDG 12 – Responsible Consumption and
Production; SDG 14 – Life Below Water; SDG 15 – Life on Land; SDG 17 – Partnership for
the Goals
o UN Partners: United Nations Office of ICT

o Civil Society: AKVO Foundation
o Digital Platform: https://digital.gpmarinelitter.org/about-us

o GPML: https://www.gpmarinelitter.org/
o LinkedIn: https://www.linkedin.com/company/global-partnership-on-marine-litter/
o Youtube: https://www.youtube.com/channel/UCoWXFwDeoD4c9GoXzFdm9Bg
o https://digital.gpmarinelitter.org/GPML-One-pager-19.08i.pdf
• Lesson Learned: The GPML Recommender project has just started, lessons learned yet
to be obtained.
• Contact Information: Heidi Savelli-Soderberg (heidi.savelli@un.org)
Project 5: Using Machine Learning to Make Government Spending Greener

• Project Description: The project seeks to show how machine learning (ML) models can
help policy makers and researchers design data-driven policies that most efficiently
and effectively allocate scarce government resources at home and abroad to maximize
inclusive and sustainable prosperity and development.
For policy- and decision-makers in many countries, one of the key impediments to
designing well-targeted green transition policies is a lack of data and intelligence on the
causal chains from a policy to its impact on society, the economy, and the environment.
It is difficult to manage and prioritize green spending if you can’t measure its e effects.
Properly trained machine learning (ML) models can enable rapid, quantitative predictions
of policy impacts.
This exploratory research venture between the UN Environment Program (UNEP)
and UNCTAD showcases how machine learning has the potential to transform the
measurement of policy impacts on Sustainable Development Goals (SDGs), Nationally
167
Determined Contributions (NDCs) and National Biodiversity Strategies and Action Plans
(NBSAPs) and enable targeted and efficient decision making for underpinning green and
inclusive transitions.
Illustrative models and analyses were created for 6 different countries: Zambia, Haiti, DRC,
Solomon Islands, Liberia, Madagaskar
• Project Type/Output: Report; Software tool
• Data Source: Data features are collected from: The ODA by sector, Aid activities targeting
Global Environmental Objectives, and The Global Forest Watch
• Technology/Platform: Trained and evaluated five different types of ML models: Artificial
Neural Networks (ANNs), Long Short-Term Memory Networks (LSTMs), Gradient Boost
Models (GBMs), Ordinary Least Squares Regression (OLS) models, and Random Forest
(RF) models.
Growth; SDG 9 – Industry, Innovation and Infrastructure; SDG 12 – Responsible
Consumption and Production; SDG 13 – Climate Action; SDG 15 – Life on Land; SDG
17 – Partnership for the Goals
• Partnership(s)/Collaborator(s) :
o UN Partners: UN Environment Program (UNEP) and UNCTAD
• Relevant Links and Multimedia: https://greenfiscalpolicy.org/blog/using-machine

-learning-to-make-government-spending-greener/
• Lesson Learned: A more comprehensive pilot study in collaboration with a national
government using historical budget data would serve as a valuable next step in exploring
this innovative method of transforming public finance decision-making.
• Contact Information: Himanshu Sharma (himanshu.sharma@un.org)
Project 6: SDG-meter
• Project Description: Textual data (policies, project reports, speeches, academic papers,
news) that can provide insights on progress, obstacles, and opportunities on SDGs is
available digitally. However, such data is not mapped nor scored according to the 17
Goals or its 169 targets. Valuable knowledge remains siloed with overall progress difficult
to capture and made known. UNEP’s ICT Unit of the Economy and Industry Division has
been developing a ML/NLP tool since mid-2020 to identify and score concepts. This
digital good, named "SDG-meter", is already able to identify and score the relevance of
the 17 SDGs within any text, with similar accuracy to a human expert.
The SDG-Meter, a web platform based on one of the most sophisticated deep learning
- natural language processing algorithms (BERT), can analyze any text to rate its relation
to each of the 17 SDGs .
• Department/Division Economy Division, ICT Unit
• Project Type/Output: Web application
• Project Status: In progress
• Data Source: Textual data, in English, with two entries: texts from IISD website (limited to
a maximum of 512 words) + associated labels (SDGs) as chosen by experts.
• In total, the database contains 2242 labeled texts (1247 from "News", 677 from "Guest
Article" and 318 from "Policy Brief)
168

• Technology/Platform: BERT (Bidirectional Encoder Representations from
Transformers), PyTorch, Python, jupyter Notebook.
UNEP
o Current work focuses on the search for an algorithm that is just as powerful as BERT but
has the ability to analyze texts without word limitations. To do so, we are in the process
of forming a consortium of experts around the problem of automatic classification
of SDGs in order to acquire more information on the latest advances in the field.
Moreover, via the chair established with the engineering school ISEP, we had the
opportunity to realize a hackathon allowing to gather engineering students around our
problem of word limitations. The study and the continuation of the various proposed
solutions are in progress.

o Academia: ISEP (Institut supérieur d’électronique de Paris)
• Relevant Links and Multimedia: Pilot - http://62.160.8.100

• Lesson Learned:
o Challenges:
- Limitation of the size of the text to process with the Google BERT algorithm (512
words or 2 pages)
o Future work:
- Experiment new deep learning algorithm in order to avoid input text length
limitation
- The tool can be further developed to analyze texts in bulk, from web databases,
and provide insights from collections of texts. Ideally this can be included in
different CMS as a plugin/module, streamlining the process.
• Contact Information: Robert Rodriguez (robert.rodriguez@un.org), Jade Guisiano (jade.

guisiano@etu.sorbonne-universite.fr) and Jonathas De Mello (Jonathas.demello@un.org)
Project 7: Predictive analytics to support governments in transforming food systems

(TEEBAgriFood)
• Project Description: This project catalyzes the use of predictive analytical tools to support
governments in transforming food systems faster, better and at a greater scale than
before.
Predictive analytics and scenario modeling allow decision makers to compare future
policy intervention options in the food system, illuminating the full impacts of a policy
decision on natural, social, human and produced capital. This enables strategic foresight
and anticipatory decision-making that values instead of discounts the future.
Modeling techniques and tools are increasingly becoming more efficient through
machine learning and the use of Earth Observations. Models applied for such exercise
can range from ‘snapshot’ biophysical ecosystem service models (such as InVEST) up to
the modeling of dynamic biophysical and socio-economic systems including dynamic
relationships between conflicting or compounding government policies (for example
agriculture subsidies and environmental regulations).
Applications are initiated in Brazil, China, India, Indonesia, Mexico, Thailand, Malaysia,
Colombia, Kenya, Tanzania, Thailand and Mexico.
169
• Project Type/Output: Report; Policy Framework

• Data Source: freely available global remote-sensing derived data; national and subnational
government data; biophysical data at the watershed level
• Contact Information: Salman Hussain (salman.hussain@u n. org), Tomas
Declercq (declercq2@un.org)
Project 8: UNEP International Methane Emissions Observatory (IMEO)
IMEO is operationalizing a Methane Alert Response System (MARS) to support governments

and businesses reducing methane emissions as necessary to meet the objective of the Paris
Agreement and as promised in the framework of the Global Methane Pledge. Machine learning
is used together with expert judgement to detect major methane emission sources based on
satellite imagery and sensor data.
• Project Type/Output: Monitoring platform

• Project End Year: N/A
• Project Domain: Climate change / Methane
government and site data;
• Contact Information: Manfredi Caltagirone <manfredi.caltagirone@un.org>
Project 9: Freshwater Ecosystems Explorer
Uses machine learning to identify areas of surface water from satellite imagery at a global level.
It produces an annual dataset of surface water extent as a formal reporting indicator for SDG
6.6.1. This is an important precedent as it is the only globally derived satellite-based indicator in
the SDG framework that member states can use in their national reporting. SDG 6.6.1 platform
is a partnership between UNEP, Google Earth Engine and the EU Joint Research Center.

• Project Domain: Freshwater ecosystems
• Contact Information: Stuart Crane <stuart.crane@un.org>
170
Project 10: Counter Measures II Project
This innovative project has been working with communities along Asian rivers including the
UNEP
Mekong and Ganges Rivers to map where and how plastic waste is entering the waterways.
Citizen science is a key part of this effort, alongside drone imaging, machine learning and GIS
algorithms. Using the project app, communities can identify plastic waste hotspots to help find
how plastic is leaking into rivers.

• Project Domain: Plastic waste
• Contact Information: Kakuko Nagatani-Yoshida <nagatani-yoshida@un.org>
Project 11: Environment-GPT
A refined large language model for information aggregation and generation of high-integrity
scientific information on the environment. EnvironmentGPT will be trained using authoritative
data sources including from the IPCC, IPBES, GEO, and scientific papers and reports about
the climate and interconnected topics such as ecosystems, biodiversity, air quality, and energy.
The vision for EnvironmentGPT is to develop a digital public good that will distil the world’s
environmental science to support the work of a) Environmental managers/practitioners and
decision-makers in government and business to easily access and use authoritative science; b)
Scientists and specialists to deliver environmental assessments more rapidly and track progress
against environmental commitments more easily.
• Project Type/Output: chatbot platform

• Project Domain: environmental knowledge
• Data Source: UNEP vetted environmental science
• Contact Information: Clara Wegenast <clara.wegenast@un.org>
Project 12: Sustainable Consumption and Production Hotspot Analysis Tool (SCP-
HAT)
SCP-HAT is deploying AI to enable the tracing of environmental pressures and impacts along
the supply chain of the goods and services consumed within a given country. It can identify
the hot spot areas of unsustainable production and consumption in order to support setting
priorities in national sustainable product and consumption policies and transition plans. The tool
was jointly developed by the Life Cycle Initiative, the One Planet network and the International
Resource Panel, in collaboration with the University of Vienna of Economics and Business, the
171
CSIRO and the University of Sydney. The current version of SCP-HAT uses relatively simple
predictive text techniques to derive automated reports from the information within the tool;
more sophisticated AI could be used to help the user navigate / extract further information.

• Project Domain: sustainable consumption
• Data Source: GLORIA database
• Contact Information: Llorenc Mila I Canals <llorenc.milaicanals@un.org>

All the SDGs
3. Relevant links
https://www.unep.org/
Contact Information
Saiful Ridwan, Chief, Enterprise Solutions, Corporate Services Division, UN Environment

Programme (saiful.ridwan@un.org)
Sally Radwan, Chief Digital Officer, UN Environment Programme, (golestan.radwan@un.org)
172
United Nations Educational, Scientific and

Cultural Organization
UNESCO
Project 1: Defending Human Rights in an Age of Artificial Intelligence

• Project Description: UNESCO and UNITAR jointly launched a new microlearning course
on AI and Human Rights for youths aged 16 to 24. The course breaks down complex
concepts about AI through activities built around our daily technology interactions.
Through interactive exercises, users will identify and engage with practical examples of
uses of AI, which are problematic from a human rights perspective. In turn, they learn
about the implications of AI for freedom of expression, the right to privacy and the right
to equality.
The course is originally available in English, French, Spanish, Russian and Chinese.
• Department/Division: Sector for Communication and Information
• Project Type/Output: Other: Online course
• Project Update:
o Translated online course into 20 languages including Bosnian, Catalan, Polish, Italian
etc.
• Project Domain: Education; Human Rights (Freedom of Expression, Right to Privacy, Right
to Equality); Artificial Intelligence
• Related Sustainable Development Goals SDGs: SDG 4 – Quality Education, SDG 17 –
o UNITAR (Co-partner)
o SALTO (co-partner)
o https://en.unesco.org/news/join-unesco-and-unitars-ai-and-human-rights-course
o https://www.edapp.com/course/defending-human-rights-in-the-age-of-artificial
-intelligence-2
• Contact information: Prateek Sibal (p.sibal@unesco.org)
173
Project 2: AI and the Rule of Law

• Project Description: Based on the findings of the AI Needs Assessment Survey in Africa
and another survey of over 1200 judicial operators in 100 countries, UNESCO has
launched a global MOOC on AI and the Rule of Law to strengthen capacities of judicial
operators in the use of AI in the administration of justice, while addressing the human
rights and legal implications of the use of AI with respect to bias, discrimination, privacy,
freedom of expression among others. The course is structured around six introductory
modules that unpack AI's application and impact in the judiciary.
The course is available in 7 languages and taught by 20 speakers - including sitting judges
from Supreme Courts, Human Rights Courts, Legal Experts and Technology experts. The
experts teaching the course come from India, Senegal, Kenya, Netherlands, United States,
Chile, Brazil to China.
Over 5900judicial operators from 142countries have enrolled in the course until2023.
The project continues to support capacity building for judicial operators through online
and in person trainings on AI and the Rule of Law.
• Project Type/Output: Capacity Building through - Massive Open Online Course (MOOC),
online and in person trainings and webinars
o Follow up trainings are being held in-person and online to further build capacities
on AI and the Rule of Law, reaching over 140 judicial operators from Qatar, The
Gambia, Congo- Brazzaville, Nepal, Ghana, Togo and Costa Rica. The latter comprised
a regional training of all Judges from the Inter-American Court of Human Rights.
o Two interregional trainings conducted in partnership with SADA reaching over 350
judicial operators from 15 countries in Africa.
o Webinar series on Artificial Intelligence and the Rule of Law
o Collaborating with the Future Society in hosting the 5th edition of the Athens
Roundtable of AI and the Rule of Law.
• Project Domain: Artificial Intelligence

• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education, SDG 16 –
Peace, Justice, and Strong Institutions; SDG 17 – Partnership for the Goals
o CETIC.br
o IEEE
o The Future Society
o National Judicial College
• Links and Multimedia: AI and the Rule of Law: Capacity Building for Judicial Systems
(unesco.org)
• Contact information: Cédric Wachholz (c.wachholz@unesco.org), Prateek Sibal (p.sibal@
unesco.org), Guilherme Canela De Souza Godoi (g.godoi@unesco.org)
174
Project 3: Multistakeholder AI Development – 10 building blocks for inclusive policy

design
UNESCO
• Project Description: With the Innovation for Policy Foundation (i4Policy), UNESCO
facilitated community consultations and developed a report to advise governments on
inclusive and multi-stakeholder-driven processes for developing AI policies. Through an
iterative series of multi-stakeholder learning and co-creation workshops, i4Policy and
UNESCO developed a report on Multistakeholder AI Development, including existing
examples of inclusive national approaches to AI and Digital Policy Development.
In the report, UNESCO and i4Policy distill 10 essential lessons for policymakers to harness
the collective intelligence of communities and ensure that the process of creating and
implementing public policy is inclusive and multi-stakeholder driven.
It leverages AI and innovation community networks in developing countries to inform the
development of global protocols on AI policy development process.
• Project Type/Output: Policy Framework/Seminar/meeting
• Project Updates: Mozfest 2022 (held in Mar 2022): Workshop on building a research
agenda and community to foster Global South engagement in global AI and data
governance
• Project Domain: Human Rights; Artificial Intelligence
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education; SDG
5 – Gender Equality; SDG 8 – Decent Work and Economic Growth; SDG 9 – Industry,
Innovation and Infrastructure; SDG 10 – Reduced Inequality; SDG 13 – Climate Action;
SDG 16 – Peace, Justice, And Strong Institutions; and SDG – 17 Partnership for the Goals
• Partnership(s)/Collaborator(s): Innovation for Policy Foundation (Co-partner)
• Links and Multimedia: https://events.unesco.org/event/?id=4180394255&lang=1033
unesco.org)
Project 4: UNESCO's first graphic novel on Artificial Intelligence

In the framework of UNESCO's work to harness emerging technology for sustainable
development, this graphic novel for young adults explores the impact of Artificial
Intelligence on humankind. The comic strip is designed to provide basic AI education to
all, showcasing the opportunities and challenges of technology in regard to freedom of
expression and human rights and ethics, along with other issues related to AI, including
AI and Sustainable Development and AI Professions.
The publication goes beyond theoretical understanding of the topic and shows the
everyday applications of AI, their origins, challenges and how these challenges can be
addressed. Additionally, the publication will further elaborate AI in relation to the SDGs
and UNESCO's ROAM principles.
In this original work, examples will respect geographical diversity and gender distribution
by showing one story based in Asia, Africa, Oceania and South America. Half of these
stories feature women as the main character. The screenwriters and graphic designers
are from different world regions, ensuring inclusion.
The comic strip will be published in September 2022. It has also been translated into
Kiswahili and Malagasy to enhance multilingualism in cyberspace especially in Africa, and
is due to be launched on World Kiswahili Day (7 July) 2024.
175
• Project Type/Output: Graphic Novel

• Project Domain: Environment; Gender; Human Rights; Justice; Artificial Intelligence
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education, SDG 5 –
Gender Equality, SDG 17 – Partnership for the Goals
o Katherine Evans (Writer and AI Expert)

o Bimot Media (Illustration and Publication)
o Sylvia Adongo (Kiswahili Translation)
o https://www.unesco.org/en/articles/inside-ai-algorithmic-adventure
o https://unesdoc.unesco.org/ark:/48223/pf0000382456
o Contact information: Cédric Wachholz (c.wachholz@unesco.org)Prateek Sibal
(p.sibal@unesco.org)
Project 5: Globalpolicy.ai Portal

• Project Description: The platform provides policy and decision makers with data, research,
use cases, and best practices in the field of AI policy and facilitates access to relevant AI-
related resources from International Organizations. It serves as a platform that highlights
cooperation between International Organizations and makes information readily available
on one centralized portal related to AI Governance. Key partners include the Council of
Europe, the European Commission, the European Union Agency for Fundamental Rights,
the Inter-American Development Bank, the Organisation for Economic Co-operation and
Development (OECD), the United Nations (UN), and the World Bank Group.
• Project Type/Output: A neutral, authoritative portal where citizens and stakeholders can
access up-to-date, accurate information on global AI policy initiatives.
• Project End Year: Continuous
• Project Domain: Artificial Intelligence - Artificial Intelligence will disrupt/impact almost all
industries and society.
• Technology/Platform: Wordpress
Infrastructure; SDG 17 – Partnership for the Goals
o UNESCO
o Council of Europe
o The European Commission
o The European Union Agency for Fundamental Rights
o Inter-American Development Bank
o Organisation for Economic Co-operation and Development (OECD)
o United Nations (UN)
176
o World Bank Group
UNESCO
o https://globalpolicy.ai/en/
o https://www.facebook.com/unesco/videos/3005127219743840/
• Contact Information: Cédric Wachholz (c.wachholz@unesco.org), Prateek Sibal (p.sibal@

unesco.org)
Project 6: Missing Links in AI Governance

• Project Description: The publication will provide world leaders, policy-makers, and civil
society members with perspectives that will be critical to face the immense task they're
presented with: to ensure the development of AI reaches its full potential in accordance
with democratic values and fundamental rights and freedoms. The magnitude of
this challenge requires a collaborative effort that transcends disciplinary barriers and
geographical borders. This publication brings together academics, civil society
representatives, artists and innovators to help us shift the conversation from what we
already know to what we have yet to render visible to ensure AI technologies leave no
one behind.
• Department/Division : Sector for Communication and Information
• Project Type/Output: Academic journal
o Downloaded more than 3619 times and consulted 18991 times
• Project Domain: Artificial Intelligence – Artificial Intelligence will disrupt/impact almost all
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education; SDG 5
– Gender Equality; SDG 8 –Decent Work and Economic Growth; SDG 10 – Reduced
Inequality; SDG 13 – Climate Action; SDG 16 – Peace, Justice, And Strong Institutions;
SDG 17 Partnership for the Goals
o Academia: MILA – Quebec Artificial Intelligence Institute

unesco.org)
Project 7: AI and Gender Equality

• Project Description: UNESCO, the Organisation for Economic Co-operation and
Development (OECD), the Inter-American Development Bank (IADB) and Cambridge
University have developed a joint report concerning AI's effects on women's working
lives.
The report is intended to raise general awareness of the disruptions of AI during the
workforce lifecycle from a gender perspective.
This descriptive publication is targeted broadly at the general public and decision-makers
across sectors including public sectors, private sectors and academia, and provides
policymakers with an introduction to issues regarding gender and AI, including applied
use cases to consider in AI Programmes and Policy Development
177
It has been published in March 2022, in English, Spanish and French.

• Project update:
o International Women's Day (8 Mar 2022): Launch of OECD-IDB Joint Publication on

"The Effects of AI on the Working Lives of Women"
o A new policy brief on gender bias in Large Language Models will be published in
March 2024 with the International Research centre on AI (IRCAI), Slovenia

• Technology/Platform: Report, on UNESCO web page
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education; SDG 5
– Gender Equality; SDG 8 –Decent Work and Economic Growth; SDG 10 – Reduced
Inequality; SDG 17 – Partnership for the Goals
o UNESCO (Data Collection and Project Management)

o Organisation for Economic Co-operation and Development (Data Collection and
Project Management)
o Inter-American Development Bank (Data Collection and Project Management)
o The Minderoo Centre for Technology and Democracy, University of Cambridge
o https://www.unesco.org/en/articles/what-are-effects-ai-working-lives-women-global
-experts-weigh
• Contact information: Cédric Wachholz (c.wachholz@unesco.org), Vanessa Dreier

(v.dreier@unesco.org); Prateek Sibal (p.sibal@unesco.org)
Project 8: AI and Digital Transformation Competency Framework for Civil Servants

• Project Description: The challenges of digital era governance require a new set of skills
and competencies from civil servants, ICT ministries and digital units in government.
The UN Broadband Commission's Working Group on AI Capacity Building, which
UNESCO co-chairs with Nokia, is developing a Digital Transformation and Artificial
Intelligence Competency Framework for Civil Servants. Through a series of interviews
with regional policymakers, collection and assessment of good practices globally, and
global consultations, the Working Group has developed evidence-based competency
domains, complementary attitudes and recommendations, in a Digital Transformation
and AI competency framework, for public sector duty bearers.
As an open resource, the Competency Framework is context-sensitive and adaptable for
use to support capacity building for civil services across ICT Ministries and Digital Units
in governments.
• Project Type/Output: Broadband Commission Report
178

UNESCO
o Piloting and implementing Competency Framework on AI and Digital Transformation

for Civil Servants with the Rwanda Development Board (Feb-Apr 2024)
o Disseminated knowledge on the Competency Framework on AI and Digital
Transformation for Civil Servants for more than 430 stakeholders, via 20 events.
o Developed and piloted the assessment methodology for the Competency Framework
on AI and Digital Transformation for Civil Servants with approximately 70 policymakers
from more than 40 countries.
o Published a policy brief and summary of framework with translations into Arabic,
Chinese, English, French, Spanish and Russian.
o Developed short videos for outreach to civil servants.  
o There are ongoing projects in the preliminary phase in collaboration with over 10
countries.

• Technology/Platform: Publication
Infrastructure; SDG10 – Reduced inequalities, SDG 16 – Peace, Justice, And Strong
Institutions; SDG 17 – Partnership for the Goals
o UN Broadband Commission
o UNESCO
o Nokia
o Open Knowledge Foundation
o https://www.broadbandcommission.org/working-groups/ai-capacity-building/

unesco.org)
Project 9: The Athens Roundtable

• Project Description: On 1-2 Dec 2022, the fourth edition of The Athens Roundtable will
be co-hosted by UNESCO to advance and widen the global dialogue on the enforcement
of AI policies and regulations, coordinated efforts toward AI standards and benchmarks,
and the adoption of AI in alignment with human rights and democratic values.
The Athens Roundtable is an annual multi-stakeholder conference seeking to advance the
sound development of policies, educational initiatives and evidence-based instruments,
to enable the trustworthy adoption of AI in legal systems, the practice of law, and related
regulatory compliance.
• Project Type/Output: Conference
• Project Domain: Artificial Intelligence – Artificial Intelligence will disrupt/impact almost all
179
• Technology/Platform: Hybrid format

• Related Sustainable Development Goals (SDGs): SDG10 – Reduced inequalities, SDG
16 – Peace, Justice, And Strong Institutions; SDG 17 – Partnership for the Goals
o The Future Society

o European Parliament
o IEEE
o Council of Europe
o https://www.aiathens.org/dialogue/fourth-edition

unesco.org)
Project 10: AI and the Futures of Learning project (based on the Teaching Artificial
Intelligence at School project)
• Project Description: To support Member States to harness AI for education, with equity
and inclusion as guiding principles, UNESCO initiated this project on effective use of AI
in the future of learning underpinned by three enablers: (1) needs-driven AI-enabled
futures of learning: the project will reveal emerging use cases for leveraging AI to
address fundamental needs of learning and provide recommendation on planning; (2) a
guidance on ethical principles: the project will develop ethical principles for the design,
deployment, and applications of AI in learning and education; and (3) AI competencies:
the project will develop a guiding framework on AI competencies needed by all learners
to live and learn with AI.
The Project will address both the technological and the human-oriented dimensions of
AI and the futures of learning across three sub-domains as identified in UNESCO’s AI and
Education: Guidance for policy-makers: learning with AI, learning about AI, and learning
to work and live with AI.
• Department/Division: Sector for Education
• Project Type/Output: Report/seminar/meeting:
o report on AI and the Future of Learning o Guidance for generative AI in

education and research
o AI competency framework for teachers and AI Competency framework for school
students

o The launch event of the AI and the Futures of Learning project was organized in
September 2021 to present the main activities and elicit comments, map out ongoing
initiatives, and build partnerships around the project.
o Since the launch of the project, UNESCO has carried out two surveys on government-
approved AI curricula targeting UNESCO Member States and non-governmental
organizations that provide AI curricula. The findings indicate that despite the wide
demand for AI-related competencies, only 11 countries have endorsed AI curricula to
180
date in their K-12 education systems, with a further 5 countries currently developing
AI curricula.
o A survey on the governmental use of AI in education was completed in early 2023,
UNESCO
covering assessments of Member States’ setting up of regulations on ethics of AI
and its use in education, strategies for AI in education, and national programmes
on developing AI competencies for teachers. The results of the survey informed the
AI Competency Framework for teachers. A series of extensive consultations
with experts, policy-makers, teachers, youth and other stakeholderswere organized
to inform the AI competency frameworks.
o In May 2023, UNESCO organized the first online ministerial roundtable on GenAI
gathering over 40 Ministers for an in-depth discussion on the implications and
potential of GenAI in ED.
o UNESCO has mobilized and coordinated groups of experts working on the
development of the two AI competency frameworks. The two drafts were presented
during the first edition of Digital Learning Week (Sept 2023). A 3-day working group
meeting took place at UNESCO HQ (Nov 2023) to validate the structure of the two
frameworks. The two competency frameworks are in the process of being finalized
and will be released in June 2024.
o The Guidance for Generative AI in Education and Research was released and launched
during Digital Learning Week, in September 2023. thThis is the first-ever global
guidance on generative AI (GenAI) in ED. The guidance is undergoing translation
into Arabic, Chinese, French, Portuguese, Romanian, Russian and Spanish.
o A call for contribution to the definition of Algorithm Literacy and Data Literacy was
launched in June 2023. The selected think-pieces will feed inputs for the development
of AI Competency Frameworks for students and teachers.
o UNESCO Digital Learning Week “Steering technology for education” took place in
September 2023: the topics addressed, on public digital learning platforms and GenAI,
have informed the work related to the three strands of the initiative. 17 ministers and
high-level representatives from 16 countries took part in the event
• Project Domain: Education

• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education
• Partnership(s)/Collaborator(s): Private Sector- This project is financially supported by TAL
Education Group and is open to a multi-stakeholder approach
o https://en.unesco.org/themes/ict-education/ai-futures-education; https://en.unesco
.org/artificial-intelligence/education
o https://en.unesco.org/artificial-intelligence/education
o https://w ww. unesco. org/e n/d igital- education/a i- future- learning/c ompetency
-frameworks
o https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research
• Contact information: Dr. Fengchun Miao (f.miao@unesco.org)
Project 11 Guiding the development of policies on AI and education

• Project Description: Based on UNESCO’s publication AI and education: guidance for
policy-makers, the project provides guidance and support with the development of
national policies on AI and education. The publication is made available in 7 languages
and offers guidance for policy-makers on how best to leverage the opportunities and
address the risks, presented by the growing connection between AI and education. It
starts with the essentials of AI, provides a detailed analysis of the emerging trends and
implications of AI for teaching and learning and introduces the challenges of harnessing
AI to achieve SDG 4 and offers concrete actionable recommendations for policy-makers
181
to plan policies and programmes for local contexts. The project also includes country
support and the organization of the awareness raising and knowledge sharing events,
such as launch events of the policy guidance or international forums on AI and education.
Moreover, policy examples, promising initiatives and best practices stemming from the
forum discussions are documented in the form of synthesis reports.
• Project Type/Output: Report/Policy Framework/Capacity building activities based on the
Guidance are under development.
• Project End Year: TBD
• Project updates: The Guidance for policy makers has been published in March 2021, it is
in 6 UN languages and Korean
• Online launch events, which included high-level ministry officials, were organized in the
LAC region and the Gulf States with the support of relevant Field Offices and UNESCO
Category 2 centres in the first of half of 2022. These events serve as the first step in a
series of activities to support the development of national policies on A and education in
concerned regions and beyond. The launch of the French version is being planned for
September 2022.
• Over the past four years, the following events were organized:
o International Conference on Artificial Intelligence and Education ( May 2019) resulting

in the adoption of the Beijing Consensus
o International Forum on AI and the Futures of Education (December 2020)
o International Forum on AI and Education: Ensuring AI as a Common Good to Transform
Education (December 2021)
o International Forum on AI and Education: Steering AI to Empower Teachers and
Transform Teaching (December 2022)

• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education
o Private Sector- Weidong Group (financial support)
• Contact information: Dr. Fengchun Miao (f.miao@unesco.org)
Project 12 Developing AI competencies for school students

• Project Description: Based on Teaching Artificial Intelligence at School project, the main
objectives are to develop a reference framework on AI competencies to raise awareness
and build capacities related to developing knowledge, skills and values needed to
live and work in the AI era. The project will further facilitate the planning of national or
institutional AI curricula for school education of fostering the competencies of both girls
and boys. A Report on Mapping of K-12 AI curricula based on the two surveys carried
out by UNESCO will inform the development of a guiding framework. Also, the project
involves capacity building activities and organization of workshops with countries based
on the tools and guidance developed by UNESCO. The project has been aligned with
the activities of the AI and the futures of learning project.
• Department/Division: Sector for Education
182

• Development of an AI skills framework for schools; Workshops to support the integration
of AI training into national or institutional school curriculum in a selected number of
UNESCO
countries.
• Project End Year: TBD
o In February 2022, UNESCO published a Report on Mapping of K-12 AI curricula based

on two surveys on government-approved AI curricula targeting UNESCO Member
States and non-governmental organizations that provide AI curricula. The findings
indicate that despite the wide demand for AI-related competencies, only 11 countries
have endorsed AI curricula to date in their K-12 education systems, with a further 4
countries currently developing AI curricula.
o In May 2022, based on the above mentioned report, UNESCO with the Ministry of
Education of Oman, RCEP, UNESCO Doha office and Ericsson organized an online
workshop to develop the capacities of more than 25 national curriculum developers
in integrating appropriate AI competencies (including knowledge, skills and values
relating to AI) into the national K-12 curriculum and/or institutional programmes.
o In collaboration with Lebanon's Ministry of Education and the UNESCO Beirut office,
UNESCO conducted a 3-day AI and coding training for 20 teachers and 12 Ministry
of Education representatives, from 10 to 12 May 2023, in Beirut.
o UNESCO has mobilized a group of experts working on the development of the AI
competency framework for school students. The framework is in the process of being
finalized and will be released in June 2024.

• Data Source: What the data contains, such as demographic data, satellite data
• Related Sustainable Development Goals (SDGs): SDG 4 Quality Education
• Partnership(s)/Collaborator(s): Private Sector: This project is being developed in
partnership with Ericsson and is open to a multi-stakeholder approach.
• Contact information : Dr. Fengchun Miao (f.miao@unesco.org)
Project 13: Flood forecasting for next 24 hours using AI

• Project description: Through the project “Enhancing Climate Services for Improved Water
Resources Management in Vulnerable Regions to Climate Change: Case Studies from
Africa and Latin America and the Caribbean” (CliMWaR) activities have been developed
on regional and national flood and drought monitoring platforms for the improvement
of relevant information on climate and water for decision-making, while making the
information actionable, locally relevant and timely.
The implementation of the platforms has been delivered by Princeton Climate Institute and
covers the continental area of Africa, and the national-level of Cameroon, Mozambique,
Zimbabwe, South Africa, Africa (at regional level) Namibia, Malawi. The platform is based
on a set of ground, satellite and modelled datasets, which are combined to provide a
consistent picture of hydrological conditions close to real-time, as well as forecasts out to
7-days for floods and out to 6 months for drought. The system is updated every day, about
183
1-2 days behind real-time and it runs a hydrological model with a 5 km resolution. The
system integrates an AI-driven flood forecast at 30m resolution, indicating the expected
area to be flooded in the next 24-36 hours.
• Department/Division: Natural Sciences Sector
• Project Type/Output: Other: Software tool/Application
• Project Domain: Disaster risk reduction; Artificial Intelligence
• Related Sustainable Development Goals SDGs: SDG 11 – sustainable cities and
communities (disaster risk reduction) SDG 13 – climate action
• Partnership(s)/Collaborator(s): Princeton Climate Institute, University of Montana
• Links and Multimedia: https://en.unesco.org/disaster-risk-reduction/ews-water
• Contact information: Koen Verbist (k.verbist@unesco.org)
Project 14: STEM education

• Project description: Development of the Robotics and Artificial intelligence program to
focus on designing codes, algorithms, linkage between hardware and software and 3D
printing to allow teachers, trainers, and students to create machines and AI applications
in MS especially in SIDS and Africa.
In Rwanda, about 120 students, teachers and UNESCO Trainers have convened in FAWE
Girls School/Gisozi for STEM Mentoring Boot Camp on Robotics, Artificial Intelligence,
3D Printing and Microscience. UNESCO in partnership with Rwanda National Commission
for UNESCO (CNRU), Ministry of Education, Ministry of ICT & Innovation, Rwanda Basic
Education Board, Rwanda TVET Board, FAWE Rwanda and the Creativity Lab have
organized the event.
Training on Artificial Intelligence Technology for University Teachers
In collaboration with Huawei Kenya, a series of training activities will be organized from
11-29th July on AI for teachers in Eastern Africa. The objective of this course is to provide
the most fundamental knowledge to the University / College teachers so that they can
understand what the A.I is. So, the teachers can introduce AI to their students during their
course of teaching.
Dominican Republic, The 4 days workshop took place from 15 to 18 March, 2022. 57
teachers from around the country participated in the training workshop. 15 of them have
participated in phase 1 workshop, dedicated to Artificial intelligence and robotics that was
held Virtually in July 2021. This was the first face-to-face workshops after the pandemic
period. The teachers were exposed to a number of open source softwares such as Cura,
Teachable machines etc.
• Project Type/Output: Other: training
• Related Sustainable Development Goals SDGs: SDG 4.- quality education
• Partnership(s)/Collaborator(s): Rwanda National Commission for UNESCO (CNRU),
Ministry of Education, Ministry of ICT & Innovation, Rwanda Basic Education Board,
Rwanda TVET Board, FAWE Rwanda and the Creativity Lab,
184

• Contact information: Imteyaz Khodabux (i.khodabux@unesco.org)
UNESCO
Project 15: Decision Making support tool using AI for school safety
• Project Description: Exposure of school infrastructure to multi-hazards poses significant
risk to vulnerable populations of students and their education process. In response to
this objective, UNESCO Chair in Disaster Risk Reduction and Resilience Engineering
developed a probabilistic resilience framework, combining two methodologies which
exploit Machine Learning technology, Agent-based (AB) and Bayesian Network (BN)
approaches, for system performance analysis. The framework aims to estimate the
disruption to education due to multiple hazards to quantify its resilience, considering
physical, functional, and social vulnerability aspects of the infrastructure. The resilience of
the system depends on the interdependencies of system variables from these qualitative
and quantitative aspects.
• Project Type/Output: Other: Software tool/Application
• Related Sustainable Development Goals SDGs: SDG 11 SDG 11 – sustainable cities and
communities (disaster risk reduction)
• Partnership(s)/Collaborator(s): University College London
• Contact information: Soichiro Yasukawa (s.yasukawa@unesco.org)
Project 16: Harnessing the power of AI to promote equal opportunities in the digital
world (McGovern I)
• Project Description: This project assists four countries (Chile, Morocco, Senegal, Brazil)
to translate the Recommendation into national institutional and regulatory frameworks
and to build a national consensus on the shared vision for AI. At the end of this project,
these countries will have national AI strategies developed from a bottom-up approach
and their capacities enhanced based on an ethical framework. A meta-strategy will be
developed for the global roll-out, based on the evidence and experience gained by the
four pilots.
• Project Domain: Environment, Education, Gender, Health, Human Rights, Justice,
Telecommunications, Artificial Intelligence will disrupt/impact almost all industries and
society. In particular, the Recommendation proposes concrete, impact-oriented policy
actions in 11 areas: Ethical Impact Assessment; Ethical governance and stewardship;
Data policy; Development and international cooperation; Environment and ecosystems;
Gender; Culture; Education and research; Communication and information; Economy
and labour; Health and social well-being
SDG 4 – Quality Education; SDG 5 – Gender Equality; SDG 8 – Decent Work and Economic
185
Growth; SDG 9 – Industry, Innovation and Infrastructure; SDG 10 – Reduced Inequality;

SDG 13 – Climate Action; SDG 16 – Peace, Justice, and Strong Institutions
• Partnership: Patrick J. McGovern Foundation, AIEB Network, key government entities,
local think tanks, academic experts, civil society
• Project Website (links): https://en.unesco.org/artificial-intelligence/ethics
• Contact Information: Irakli Khodeli (i.khodeli@unesco.org), ai-ethics@unesco.org
Project 17: Supporting Member States in lmplementing UNESCO's Recommendation

on the Ethics of Al through lnnovative Tools
UNESCO aims to advance the global implementation of the Recommendation on the Ethics of
AI by assisting countries in establishing robust national institutions for effective AI governance.
Over 70 UNESCO Member States have joined the Group of Friends, expressing interest in
technical assistance for building national institutions addressing AI technologies. The project
addresses these needs and will facilitate implementation of the Recommendation in the form
of support for developing and implementing national policy, legislation, regulation and other
governance mechanisms.
To build the capacities of Member States and support them in translating the Recommendation
on the Ethics of AI into policies and practices at the national level, UNESCO has developed the
following tools and mechanisms as mandated by the Recommendation:
> Readiness Assessment Methodology (RAM) to help governments determine how

prepared the country is for developing, adopting and using AI systems, by analysing
different dimensions of readiness, including legal, social, cultural, economic, scientific,
educational, technological and infrastructural.
> Ethical Impact Assessment (EIA) to enable countries to evaluate AI technologies based
on the standards of UNESCO Recommendation, underscoring the benefits and risks of a
specific system to individuals, society and the environment.
> The tools should identify impacts on human rights and fundamental freedoms, in
particular but not limited to the rights of vulnerable groups, poverty, digital divide, labour
rights, the environment and ecosystems, and ethical and social implications.
• Department/Division: Sector for Social and Human Sciences
• Project Type/Output: development of tools
• Project Status: ongoing
• Project Start Year: January 2022
• Project End Year: open-ended
• Project updates: The tools are being piloted and implemented in over 50 countries
around the world.
186

UNESCO
o Government: European Commission – DG INTPA, 70+ Member States of the Group of
Friends of the Recommendation (an op-ended alliance of countries spearheading the
implementation of the Recommendation). Particular governmental partners include
Japan, South Africa, Namibia, Mozambique, and Zimbabwe.
o Foundations: Patrick J. McGovern Foundation
• Links and Multimedia: https://en.unesco.org/artificial-intelligence/ethics

• Contact information: Irakli Khodeli (i.khodeli@unesco.org), ai-ethics@unesco.org
Project 18: Supervising AI by Competent Authorities
Project Description: UNESCO currently works with the Dutch Digital Infrastructure Authority on
a project funded by European Commission DG Reform on the supervision of AI by competent
authorities. The project will support the challenge confronting European competent authorities
in implementing the upcoming EU AI Act and other relevant legislations and international
standards. The project will provide technical support to enhance the capacities, competencies,
and knowledge of members of the Dutch and European Working Groups of Competent
Authorities on AI, to supervise AI in compliance with the forthcoming AI Act and other relevant
legislation and international standards, such as UNESCO’s Recommendation on the Ethics of AI,
developing case studies for good practices, and assessing different modalities for supervising
AI via a multistakeholder consultation.
The project has 3 phases:
1) comprehensive approach to understand the current state of AI supervision in the EU and

beyond presented in a comprehensive report;
2) capacity building efforts centring on specific domains and topics based on results of first
phase, in the form of a series of case studies to explore AI supervision. It will look at how
existing tools such as the EIA can contribute to the mission of the competent authorities;
3) a multistakeholder consultation to explore and develop options for operationalizing AI
supervision among Competent Authorities. This will culminate with recommendations on
the operational settings and requirements for AI supervision.
• Project Domain: Knowledge sharing, capacity building, AI governance, ethical AI, multi-
stakeholder participation, implementation of the Recommendation
• Partnership: European Union, Dutch Digital Infrastructure Authority
187
Project 19: Supporting Africa and SIDS to Benefit from AI Technologies while
Addressing Ethical Risks
• Project Description: This project aims to strengthen capacities and policies of Member
States in Africa and SIDS to benefit from Artificial Intelligence (AI) technologies and
to address the associated ethical risks, focusing especially on gender inequalities and
discrimination. This will be pursued through the development of innovative capacity-
building tools, including the Ethical Impact Assessment and Readiness Assessment
Methodology as outlined in the Recommendation on Ethics of AI, which upholds
protection, promotion and respect of human rights and fundamental freedoms, human
dignity and equality as its objective.
• Project Domain: A capacity building, implementation of the UNESCO Recommendation
on the Ethics of AI, Africa, SIDS
• Partnership: Government of Japan
• Project Website (links): https://en.unesco.org/artificial-intelligence/ethics
Project 20: AI Governance for Africa Toolkit - Regional and International Frameworks
• Project Description: This toolkit unpacks the context of AI governance, in Africa and
globally, and considers advocacy approaches for future governance. It does so in
the following ways: Part 1 examines existing AI governance instruments in Africa with
a particular focus on the East African Community. This chapter outlines continental
responses and details existing governing measures in Africa. Part 2 unpacks existing
international governance measures. In doing so it considers governance trends and
important considerations included in governance instruments. Part 3, which is a separate
document, explores a series of key questions for the design of advocacy strategies on
AI governance, particularly in African contexts. The aim of the toolkit is to empower
journalists and civil society organizations to inform public discourse, drive policy and
regulatory change and advocate for ethical and responsible AI deployment.
• Project End Year:
• Project Domain: [some keywords about the project]
• Partnership: Thompson Reuters Foundation, Patrick J. McGovern Foundation
• Project Website (links): https://www.trust.org/
188
Project 21: Business Council for Ethics of AI

• Project Description: The UNESCO AI Business Council is a strategic initiative designed
UNESCO
to collaborate with companies worldwide in promoting the ethical development and use
of artificial intelligence (AI). Drawing on the UNESCO Recommendation on the Ethics of
Artificial Intelligence, the council aims to implement ethical principles in AI applications,
foster the exchange of best practices, and enhance ethical impact assessments to
promote AI that respects human rights globally.
• Project End Year: Indefinite
• Project Domain: Ethical Artificial Intelligence
• Related Sustainable Development Goals (SDGs): The project contributes to multiple
SDGs, including but not limited to SDG 8 (Decent Work and Economic Growth), SDG 9
(Industry, Innovation, and Infrastructure), SDG 10 (Reduced Inequalities), SDG 13 (Climate
Action), SDG 16 (Peace, Justice, and Strong Institutions), and SDG 17 (Partnerships for the
Goals), by promoting ethical standards in the development and use of AI technologies.
• Partnership(s), Collaborator(s):
o Co-chairs: Microsoft, Telefónica

o Members: Bimbo Group, Salesforce, LG AI Research
o Observers: GSMA
• Links and multimedia: https://www.unesco.org/en/artificial-intelligence/business-council

• Contact Information: Melguizo, Angel a.melguizo@unesco.org ; Eleonora Lamm e.lamm@
unesco.org; Gonzalez, Natalia n.gonzalez@unesco.org
Project 22: AI Ethics Experts Without Borders (AIEB)
The AI Ethics Experts without Borders (AIEB) Network assists Member States in implementing
the UNESCO Recommendation on the Ethics of AI by providing on-demand support and
tailored policy advice and by contributing to UNESCO’s knowledge production to guide the
implementation of the Recommendation. The key tools used by the experts are the Readiness
Assessment Methodology and the Ethics Impact Assessment. The AIEB Network represents
an expert facility featuring international pool of recognized experts (independent consultants,
academics and researchers, governmental officials and civil servants, experts working in private
sector, etc.) that have a relevant experience in designing or implementing policies to advance
ethical governance of AI, fostering collaboration, building capacities and competences, and
sharing knowledge and experience globally.

• Project Type/Output: networks/event/research
• Project updates: Project was launched at the Globl Forum on the Ethics of AI in February
2024.
• Project Domain: Education, Gender, Health, Human Rights, Justice, Telecommunications,
Artificial Intelligence, Ethical governance and stewardship; Data policy; Development
and international cooperation; Environment and ecosystems; Culture; Education and
189
research; Communication and information; Economy and labour; Health and social well-
being
o Government: 70+ Member States of the Group of Friends of the Recommendation

(an op-ended alliance of countries spearheading the implementation of the
Recommendation).
o Academia: Alan Turing Institute, IE University, London School of Economics, MIT,
Technical University of Munich, Harvard University, Singapore Management University,
University of Ghana, Universidad del Rossario, United Nations University, Chinese
Academy of Science.
• Links and Multimedia: https://www.unesco.org/ethics-ai/en

Project 23: The Global AI Ethics and Governance Observatory
The Global AI Ethics and Governance Observatory is a collaborative and information-sharing

platform to increase understanding of the challenges and impact of AI and promote peer
learning. The Global AI Ethics and Governance Observatory will host the results of the analytical
work from all countries participating in the RAM exercise, together with the contributions from
experts from around the world. The aim of the Observatory is to provide a global resource for
policymakers, regulators, academics, the private sector and civil society to find solutions to the
most pressing challenges posed by Artificial Intelligence.

• Project Type/Output: research/database
• Project Start Year: December 2023
• Project End Year: Open-ended
o UN Partners: UN
(an open-ended alliance of countries spearheading the implementation of the
Recommendation).
190

UNESCO
Academy of Science.
• Links and Multimedia: https://www.unesco.org/ethics-ai/en

Project 24: Women for Ethical AI Network (Women4EthicalAI)
The Women for Ethical AI platform (Women4EthicalAI) leverages the knowledge, contribution
and networks of leading Artificial Intelligence (AI) experts to advance gender equality in the
AI agenda. This new multi-stakeholder collaborative platform supports governments and
companies’ efforts to ensure that women are represented equally in the design, use and
deployment of AI, and to promote trustworthy, gender-friendly and inclusive AI systems.
Women4Ethical AI’s experts further contribute to advance all the other policy chapters and to
implement the provisions contained in the UNESCO Recommendation on the Ethics of Artificial
Intelligence, from education and research, to economy and labour, and the environment and
ecosystems.

• Project Type/Output: networks/event/research
• Project Domain: Education, Gender, Health, Human Rights, Justice, Telecommunications,
Artificial Intelligence, Ethical governance and stewardship; Data policy; Development
and international cooperation; Environment and ecosystems; Culture; Education and
research; Communication and information; Economy and labour; Health and social well-
being

Recommendation).
Academy of Science.
• Links and Multimedia: https://www.unesco.org/en/artificial-intelligence/women4ethical

-ai
191
Project 25: Global Forum on the Ethics of AI
The Global Forum on the Ethics of AI will mark an important milestone in the building of a robust
international coalition towards ensuring the ethical development and use of AI worldwide. The
Forum will comprise three major components:
1. Ethical Development and Use of AI across the world reinforcing national regulatory
frameworks and institutions.
2. Global Collaboration on the Implementation of the Recommendation on the Ethics of
AI: This session will feature digital ministers from across the world, including Africa, Latin
America and the Caribbean, and Asia and the Pacific.
3. Special Topics: parallel sessions will be dedicated to the exploration of the state of AI
from different angles – Gender Equality, Environmental Protection, Readiness and Ethical
Impact Assessments, Transparency and Non-discrimination.
• The Forum will take stock of the national efforts to promote the ethical development and
use of AI and identify the best regulatory practices and institutional settings to ensure the
ethical development of these technologies.
• Project Type/Output: event
• Project Start Year: December 2022
• Project End Year: recurrent

Recommendation).
• Links and Multimedia: Global Forum on the Ethics of AI event website, Youtube Playlist
with all livestreams of the sessions
Project 26: Sensitizing and building capacity of the stakeholders to address ethical
issues of AI
Project Description: To build the capacities of stakeholders of AI technologies (developers and

beneficiaries), UNESCO has organized a series of roundtables on the selected topics of ethics
of AI inviting eminent experts around the world. A series of short educative and informative
video of each topic of the roundtable are produced and made available through UNESCO’s
192
YouTube channel for educational purposes. This project was financially supported by the
Japanese Ministry of Education, Culture, Sports, Science and Technology.
UNESCO
• Project Type/Output: Roundtables, short videos
• Project Status: completed
• Project Start Year: July 2019
• Project End Year: July 2022
• Project updates: Three roundtables on selected topics of Ethics of AI had been organized
at UNESCO HQ inviting eminent experts from different regions as follows:
o The 1st roundtable on Ethics of AI “Changing Relationship between Artificial

Intelligence and Humans”, December 2019.
o The 2nd roundtable on Ethics of AI “Shaping the Future of AI through Cultural Diversity”,
March 2021.
o The 3rd Roundtable on the ethics of AI “Challenges of AI Ethics and Governance”,
February 2022.
• The video recording of the roundtables are freely accessible through UNESCO website.
5 short educational videos on the topics addressed during the roundtables are freely
accessible through UNESCO’s youtube channel with several language subtitles.
• Project Domain: Gender, Health, Human Rights, Justice, Telecommunications, Artificial
Intelligence will disrupt/impact almost all industries and society.
SDG 4 – Quality Education; SDG 5 – Gender Equality; SDG 9 – Industry, Innovation
and Infrastructure; SDG 10 – Reduced Inequality; SDG 16 – Peace, Justice, and Strong
Institutions
• Partnership(s)/Collaborator(s): Japanese Ministry of Education, Culture, Sports, Science
and Technology
o Academia: Twente University (Netherlands), Harvard University (USA), Tokyo University

(JPN), Kyoto University (JPN), Cardiff University (UK), Glasgow University (UK), Robotics
Institute (Spain), Article 19, Concordia University (Canada), Rio de Janeiro State
University (Brazil), Australian National University (Australia), Stanford Digital Civil
Society Lab (USA), Pollicy organization (Uganda), Pretoria University (South Africa)

Roundtable video recordings:
o 1st roundtable: "Changing Relationship between Artificial Intelligence and Humans",

http://webcast.unesco.org/live/vod/2019/shs/20190312_shs_room-02/en/
o 2nd roundtable: “Shaping the future of AI through Cultural diversity”, https://www
.youtube.com/watch?v=Rdp6hQXVpqM
o 3rd roundtable: “AI Ethics and Governance: From Principles to Practices”, https://www
.youtube.com/watch?v=Innb8D9NLp8
Educational short video links:
o Do you know AI or AI knows you better? https://youtu.be/im0XTC91qMI

o Does AI make better decisions than humans? https://youtu.be/2E7l1hdjHsg
o Shaping AI through Cultural diversity https://www.youtube.com/watch?v=
AiK0iYZuNS0
o Evolving interactions between Humans and AI https://youtu.be/xDmQMpwiHdA
193
• Contact information: Orio Ikebe (o.ikebe@unesco.org), Irakli Khodeli (i.khodeli@unesco

.org), ai-ethics@unesco.org
Project 27: Landscape Study of AI Policies and Use in Southern Africa
Project Description: Guided by the UNESCO Recommendation, this project includes a policy
mapping of AI and relevant digital policies in 9 countries in Southern Africa (Angola, Botswana,
Lesotho, Mozambique, Namibia, South Africa, Swaziland, Zambia and Zimbabwe), as well as
a light-weight review of the extent of AI use across four key sectors in the 9 countries. These
analyses will adopt strong human rights and ethical standpoints, And will Highlight the areas of
potential ethical risks and threats on the long-term fulfillment of fundamental rights, protection
of social cohesion, and promotion of justice and equality will be essential for the future of
technological development.
UNESCO will use the study findings to inform the expert and policy discussions ensuring that
dialogues are evidence-based. In this framework, the specific objectives of this study are to:
1. Understand and raise public awareness on the actual utilization of the AI technologies in
the Southern Africa region, taking some sectors as examples and case studies.
2. Alert policy makers and technicians on the potential risks and threats of AI utilization on
human rights and fundamental freedom of individuals, on social justice and equality, on
governance of public affairs, and on the environment.
3. Guiding the process of ethical governance by investigating existing stewardship principles
from industries and government agencies.
4. It also seeks to promote human rights respecting governance frameworks by assessing
ratification to international, continental and regional policies/frameworks/strategies.
The study will include two segments: (i) a mapping review of policies that will focus on providing
critical analysis of the policy and normative environment from a human rights and ethical
standpoint, which can frame the AI utilization;1 and (ii) a sectoral landscape review of how the AI
technologies have been used concretely, without necessarily the public and the policy makers
being aware of all the human rights and ethical implications.
• Department/Division: UNESCO Harare office

• Project Type/Output: Research
• Project Start Year: May 2022
• Project End Year: August 2022
• Project Domain: Research/report
SDG 4 – Quality Education; SDG 5 – Gender Equality; SDG 9 – Industry, Innovation
and Infrastructure; SDG 10 – Reduced Inequality; SDG 16 – Peace, Justice, and Strong
Institutions
o Government: Angola, Botswana, Lesotho, Mozambique, Namibia, South Africa,

Swaziland, Zambia and Zimbabwe
o Academia: Research ICT Africa
• Links and Multimedia: Link to the final study: https://unesdoc.unesco.org/ark:/48223/

pf0000385563
194
• Contact information: Phinith Chanthalangsy p.chanthalangsy@unesco.org
UNESCO
Project 28: UNESCO-Southern Africa sub-Regional Forum on Artificial Intelligence
(AI)
• Project Description: The Forum aims to provide a platform for Member States, international
organizations, civil society, academia, the AI industry, and other stakeholders to discuss
issues and challenges related to AI development in Africa, including equity and ethics
and encourage intra-African cooperation in AI.
• Department/Division:
• Project Type/Output: Conference, synthesis report
• Project Domain: Artificial Intelligence; human rights; STEM education; data protection;
gender equality
• Related Sustainable Development Goals SDGs: SDG 17-Partnership for the Goals, SDG
5- Gender Equality, SDG 4- Quality education, SDG 10- Reduced Inequality
• Partnership(s)/Collaborator(s): Ministry of Higher Education, Technology and Innovation
of Namibia
• Links and Multimedia: www.sarfai2022.org
Project 29: UNESCO IPDC Handbook on Reporting on Artificial Intelligence for

Journalists
• Project Description: Against the backdrop of AI’s emerging ubiquity, AI coverage must
inform audiences about the implication of the technology itself, beyond journalism.
For instance, reporting on the power dynamics in the changing relationship between
companies, authorities, citizens and computer chips, and between data and algorithms.
While many AI deployments serve public interest, journalists also need insight and
expertise to alert about aspects like exclusions, unequal benefits, and violations of human
rights. As part of its journalism education series, UNESCO’s International Programme
for the Development of Communication (IPDC) has supported the World Journalism
Education Council in commissioning this handbook. The handbook aims to inspire and
empower journalism educators to help both journalism students and working journalists
do justice to one of the major issues of our times. Trainings for journalists based on the
handbook have already been piloted in Rwanda, in support of the government’s effort
to raise awareness about its National AI Policy among the population, and to strengthen
journalist's ability to report on AI in Africa in an accurate, nuanced and informed way.
• Department/Division: Communication and Information
• Project Type/Output: Handbook/ Capacity Building
• Project Domain: Gender, Human Rights, Telecommunications, Communication and
Information, Medi, Education
• Related Sustainable Development Goals (SDGs): SDG 4 – Quality Education; SDG
5 – Gender Equality; SDG 8 – Decent Work and Economic Growth; SDG 9 – Industry,
Innovation and Infrastructure; SDG 10 – Reduced Inequality; SDG 16 – Peace, Justice, and
Strong Institutions
195
o UN Partners: World Journalism Education Council (WJEC)

o Academia: Çukurova University, Utrecht University, Christ Nagar College, Trivandrum,
Saint Petersburg University
• Contact information : Misako Ito (m.ito@unesco.org)
Project 30: UNESCO Framework for Culture and Arts Education 2024
The UNESCO Conference on Culture and Arts Education took place from 13-15 February 2024
in Abu Dhabi, UAE and resulted in the unanimous adoption of the first Global Framework for
Culture and Arts Education, strengthening the synergies between education and culture in
formal, non-formal, informal, lifelong and life wide education and learning, including TVET,
across the broad spectrum of cultural resources from cultural heritage to living expressions
and the creative economy. The Framework includes two specific references to AI among its
policy guidance and recommendations:
• (II. Objectives, point ii) “Harness contemporary advances and seize the opportunities
opened up by technology, while identifying, preventing and mitigating risks, notably
in digital technologies and Artificial Intelligence (AI), in order to support and promote
reflection, creativity, initiatives, and ethical and responsible use in this domain, in particular
to the benefit of the educational, cultural and creative sectors.”
• (IV Strategic Goals, a) Access, inclusion and equity; point 14): “All learners, educators
and teachers should have equitable and inclusive access to infrastructure and resources
and learning opportunities to develop the skills and competencies to benefit from digital
technologies and AI. While digital technologies and AI have expanded new ways to
access and engage with culture and arts education, addressing the digital divide and
the imbalance in cultural diversity and expressions online have become equally critical
priorities to remove barriers to participation due to economic, geographical and social
disparities, and to equip learners, teachers and educators with the relevant knowledge
and skills they need, including media and information literacy.”
• Department/Division: UNESCO Division of Cultural Policies and Intercultural Dialogue
• Project Type/Output: Policy Guidance
• Project Status: Adopted
• Project updates: Framework adopted on 15 Feb 2024
• Project Domain: Culture and Education
Gender Equality; SDG 9 – Industry, Innovation and Infrastructure; SDG 10 – Reduced
Inequality; SDG 16 – Peace, Justice, and Strong Institutions
o Government: UAE
o Steps: https://www.unesco.org/en/frameworkcultureartseducation
o Adoption: https://www.unesco.org/en/articles/unesco-member-states-adopt-global
-framework-strengthen-culture-and-arts-education
196
• Contact information: Paola Leoncini Bartoli, Director of UNESCO CPD, p.leoncini

-bartoli@unesco.org
UNESCO
Project 31: UNESCO High-level discussion on AI in the audio-visual industry (2023)
• On19 October 2023, UNESCO organized a high-level discussion on AI in the audio-visual
industry entitled “The Film sector on the frontlines” at its HQ in Paris, which brought
together film stakeholders and AI experts from across the globe to voice a diversity
of perspectives on the challenges and opportunities presented by AI in the cultural
and creative sectors. Recommendations from the discussion highlighted the need for
UNESCO and other international organizations to develop inclusive and participatory
international guidelines and standards that are human-centered and guarantee:
o Consent: artists and cultural professionals should be informed of and consent to the
use of their work for training generative AI models;
o Fair remuneration: artists and cultural professionals whose work is being used by
generative AI tools should be remunerated and have their intellectual property rights
protected;
o Transparency: users of generative AI tools should be aware of the sources used to
create the content they are requesting, notably to avoid unintentional plagiarism.
Similarly, works that were generated through AI should be clearly indicated;
o Cultural diversity: generative AI models should be trained and corrected to avoid
perpetuating biases and stereotypes and ensure that diverse and representative
stories or images are produced and disseminated.
• Department/Division: UNESCO Diversity of Cultural Expressions Entity

• Project Type/Output: Awareness raising, policy recommendations
• Project updates: Event organized 19 October 2023
• Project Domain: Culture and Creative Industries/ Audio-visual sector (Film)
o UN Partners: consultations with ILO, WIPO

o Private Sector: CISAC
o https://www.unesco.org/en/articles/unesco-unites-diverse-perspectives-inform
-policies-ai-creative-sectors
• Contact information: Karalyn Monteil, Head of Programmes & Stakeholders Outreach,

UNESCO Diversity of Cultural Expressions Entity, k.monteil@unesco.org
Project 32: UNESCO Expert Reflection Group on the diversity of cultural expressions
in the digital environment (2024-2024)
• In the framework of the UNESCO 2005 Convention on the Protection and Promotion
of the Diversity of Cultural Expressions, UNESCO has established a reflection group on
the diversity of cultural expressions in the digital environment. The reflection group's
197
mandate for 2024-2025 will be to exchange knowledge and experiences regarding the
implementation of the 2005 Convention in the digital environment and to formulate
recommendations, which will be presented to the UNESCO Intergovernmental
Committee on the Protection and Promotion of the Diversity of Cultural Expressions at
its eighteenth session in February 2025 and then transmitted to the tenth session of the
Conference of Parties of the 2005 Convention in June 2025. The group is composed of
18 experts from 6 regions. They will participate in two meetings to be held in May and
September 2024.
• Department/Division: UNESCO Diversity of Cultural Expressions Entity
• Project Type/Output: Policy recommendations
• Project updates: Experts identified Jan 2024
• Project Domain: Culture and Creative Industries
o Government: Canada, France
o https://www.unesco.org/creativity/en/articles/closed-call-experts-reflection-group
-diversity-cultural-expressions-digital-environment
• Contact information: Laurence Mayer-robitaille, l.mayer-robitaille@unesco.org, UNESCO

Diversity of Cultural Expressions Entity
Project 33: Addressing Hydro-Climatic Vulnerability through Citizen and Open

Science
• The project aims to improve the ability to forecast hydro-climatic extremes, such as
droughts and floods through innovative technologies, such as AI, to build water resilience
against climate change and improve water management in pilot countries of Sub-Saharan
Africa and Latin America. A mapping of Machine Learning (ML) and Artificial Intelligence
(AI) applications for water resource management is currently underway, beginning from
November 2023. These exercises will establish the baseline and identify potential for ML/
AI applications in sustainable water management.
• Department/Division: UNESCO Natural Sciences Sector, Water Sciences Division
• Project Type/Output: research, databases
• Project Domain: water, climate change, disaster risk reduction
• Related Sustainable Development Goals (SDGs): SDG 6 Clean water and sanitation; SDG
13 Climate action
198
o Government: Government of Flanders, Belgium; UNESCO IHP National Committee
UNESCO
of the Republic of Korea
o Academia: Delaters
• Links and Multimedia: https://www.unesco.org/en/ihp/hydro-resilience

• Contact information: k.verbist@unesco.org Koen Verbist UNESCO Intergovernmental
Hydrological Programme

SDGs 3, 4, 5, 8, 9, 10, 11, 13, 16, and 17
3. Relevant links
https://www.unesco.org/en
Contact information
Clare Stark (c.stark@unesco.org), and Misako Ito (m.ito@unesco.org)
199
United Nations Framework Convention on Climate

Change
Project 1: UNFCCC Technology Mechanism Initiative on Artificial Intelligence for

Climate Action
• Project Description: The UNFCCC Technology Mechanism Initiative on Artificial
Intelligence for Climate Action (#AI4ClimateAction) aims to explore the role of AI
as a powerful technological tool for advancing and scaling up transformative climate
solutions for mitigation and adaptation action in developing countries with a focus
on least developed countries (LDCs) and small island developing States (SIDS), while
also addressing the challenges and risks posed by artificial intelligence. In 2023, the
Conference of the Parties (COP) of the UNFCCC requested the UNFCCC Technology
Executive Committee (TEC) and the Climate Technology Centre and Network (CTCN)
to implement the initiative with special attention to capacity needs and technology
priorities of developing countries and in support of the Technology Mechanism joint
work programme for 2023–2027;
• Division: Means of Implementation Division/Technology Sub-Division
• Project Type: Initiative
• Project Domain: Climate action
• Related Sustainable Development Goals (SDGs): All, in particular SDG 13 – Climate Action
O UNEP/CTCN
• Relevant Links and Multimedia: https://unfccc.int/ttclear/artificial_intelligence

• Contact information: Moritz Weigel (tec@unfccc.int)
Project 2: AI Innovation Grand Challenge

• Project Description: The UNFCCC Technology Executive Committee (TEC) launched
the AI Innovation Grand Challenge in partnership with Enterprise Neurosystem during a
High-level Event on Artificial Intelligence for Climate Action at the UN Climate Change
Conference in Dubai in December 2023. The AI Innovation Grand Challenge is a
climate innovation competition to identify new AI applications for climate mitigation and
adaptation action in developing countries and to make them available open-source as a
public good.
• Division: Means of Implementation Division/Technology Sub-Division
• Project Type: Innovation contest
200

UNFCCC
• Project Domain: Climate action
• Related Sustainable Development Goals (SDGs): All, in particular SDG 13 – Climate Action
O Enterprise Neurosystem
• Relevant Links and Multimedia: https://unfccc.int/ttclear/artificial_intelligence

• Contact information: Moritz Weigel (tec@unfccc.int)

All SDGs: SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17
3. Relevant Links
https://www.unfccc.int/
Contact information
Moritz Weigel (tec@unfccc.int)
201
United Nations Population Fund
Project 1: TABOO
• Project Description: Taboo is a service that offers users relevant information and to easily
obtain and visualize disaggregated and significant data about myths and misconceptions
of sexual and reproductive health contained in the genuine opinions of the Spanish-
speaking population on Twitter.
Through scraping techniques this data is meant to be a complement to the official
information gathered by the entities in charge of guaranteeing sexual and reproductive
rights in Colombia and other Spanish-speaking countries. In this way, it is possible to
obtain a more robust picture of what is happening around this issue, based on this
compilation of information to make better decisions when we approach the citizens with
new actions and strategies that seek behavioral changes.
Implementing trusted and tailored messages, in the right channels will bust myths and
unveil misinformation and help women, young women and adolescents, to exercise their
sexual and reproductive health and rights and prevent unwanted pregnancies.
The project is centered in eliminating barriers encountered in the demand of family
planning related to taboos, lack of understanding of RH and limited decision making.
Since we intend to unveil the myths and misconceptions in family planning, our innovation
will contribute to preventing unplanned/unwanted pregnancies.
• Department/Division: Innovation Lab UNFPA Colombia
• Project Type/Output: Dataset; Software tool
• Project updates: Main results: Through Taboo, it was possible to carry out a classification
of myths regarding sexual and reproductive health and to collect 212.257 Tweets.
After this first exercise we obtained three categories of myths: 1. Sexually transmitted
diseases, 2. Modern contraceptive methods and 3. Unscientific methods of contraception.
From the first category, five major misconceptions were obtained: 1. You cannot get an
STI by practicing oral sex, 2. Condoms do not protect people from STIs, 3. Gay people are
more likely to get STI’s, 4. There is a cure for HIV/AIDS, and 5. Only promiscuous people
get STI’s, of which 177.522 related tweets were found.
From the second category, 11 great myths were obtained: 1. More babies die from
abortion in two days than all deaths from COVID- 19, 2. Abortion is never medically
necessary, 3. The use of the pill can cause varicose veins, 8. You should only use the IUD
if you have children, 9. Occasionally you should stop taking contraceptives to "clear" or
"take a break" from hormones, 10. The morning-after pill causes abortions, and 11. The
202
morning-after pill can be used as a regular contraceptive, of which 20345 related tweets
were found.
Finally, from the third category, 6 myths were obtained: 1. The first relationship does
UNFPA
not make you pregnant, 2. There are safe days for a woman not to become pregnant, 3.
During menstruation a woman cannot become pregnant, 4. Parsley serves as an abortion,
of which 14.390 related tweets were found.
Additionally we have arranged the complete database in a dashboard so that it can be
explored by users.
This information will allow the UNFPA Colombia office, other spanish-speaking COs and
LACRO to better understand the beliefs of the population with respect to sexual and
reproductive health, so that from the lack of clear knowledge these beliefs can be broken
by means of more accurate messages.
Activities achieved:
o A clear methodology for the collection of information by looking out some fake news.
o An improvement was made to the scraping algorithm to obtain the expansion
information on each myth (Retweets, likes, followers).
o The brand identity of the project (brandbook) was designed: identity guidelines, logo,
logo marks and color palette.
o A functional microsite was designed and developed, which contains: 1. Technical
information that counters the myths in the form of storytelling, 2. Myths Explorer
Dashboard and 3. Tweets Explorer Dashboard.
o By updating the scraping algorithm improvements, 3 categories containing a total of
21 myths were obtained.
Three main assumptions tested and proven:
1. People talk and share their beliefs and myths about family planning and
contraception on Twitter and Google search:
Our first scrapping process allowed us to get more than 800.000 tweets. It was
possible to carry out a classification of myths regarding sexual and reproductive
health and to collect 212.257 Tweets.
After this first exercise we obtained three categories of myths: 1. Sexually
transmitted diseases, 2. Modern contraceptive methods and 3. Unscientific
methods of contraception.
2. There are areas of the country that have different discussions about contraception
and sexuality.
Our main finding here is that not just in the same country we can find different
myths, also throughout Latin America we discovered diverse myths and discussion
about contraception and sexuality. The gathered information enables UNFPA
Colombia office, other Spanish-speaking COs and LACRO to better understand
the beliefs of the population with respect to sexual and reproductive health. We
identified that the borders on the internet are the languages and not the country
borders, so we found posts from all the Spanish speaking countries.
3. We can gather the information, including new insights, needed (by user) to build
an effective campaign for myth busting through scraping.
The main source of information explored was the Twitter posts (near to 300.000)
from July 5, 2007 to December 3, 2020, which contain words that we have
identified as relevant and recurrent in conversations about contraceptive myths
(our own taxonomy). However, some publications with these same terms try to
disprove these misconceptions, not to spread them. The nature of social networks
means that we cannot ensure that all trills refer to these myths or want to spread
them, which is why in the data exploration tools we have referred to the issues
associated with contraceptive myths and not the myths themselves.
203
We found the next insights related to myths and misconceptions around contraception:
1. You cannot get an STI by practicing oral sex

2. Condoms do not protect people from STIs
3. Gay people are more likely to get STI’s
4. There is a cure for HIV/AIDS
5. Only promiscuous people get STI’s
6. More babies die from abortion in two days than all deaths from COVID-19
7. Abortion is never medically necessary
8. The use of the pill can cause varicose veins
9. You should only use the IUD if you have children
10. Occasionally you should stop taking contraceptives to "clear" or "take a break"
from hormones
11. The morning-after pill causes abortions, and 11. The morning-after pill can be
used as a regular contraceptive
12. The first relationship does not make you pregnant
13. There are safe days for a woman not to become pregnant
14. During menstruation a woman cannot become pregnant
15. Parsley serves as an abortion
• Project Domain: Education; Health; Human Rights

• Data Source: Taboo was aimed at searching for information on social networks,
specifically on Twitter. More than 330 thousand tweets were collected between July 5,
2007 and December 3, 2020 that were related to sexual and reproductive health and
were classified into 3 major areas (Sexually transmitted infections, Non-scientific methods
of contraception and Modern methods of contraception).
• Link to data: https://datasketch.github.io/unfpa/
• Technology/Platform: NLP (Word vectorization), Web Scraping - Python, D3.js, Google
Cloud
SDG 4 – Quality Education; SDG 17 – Partnerships for the Goals
o UN Partners: UNFPA Innovation Secretariat

o Government:
o Ministry of Health
o Major´s office of Bogotá
o National Planning Department
o ICT Ministry
o Secretariat for Youth in Medellín
o Private Sector: DataSketch
o Academia: Universidad del Norte (Barranquilla, Colombia)
• Lessons Learned: Make sure you have the staff and tools you need: One of the great
lessons we learned during the project was related to human capabilities and technological
tools required at the right time. Knowing the resources needed for the project from the
beginning allows us to investigate the processes required to obtain them or discover
the elements that the Fund already has that can be used. Taboo as a new project in
the organization, had administrative processes unknown to the team, which generated
reprocessing and delays in service delivery.
204
Build on what is built: As a great learning process we understood that innovation does
not work when it is done individually, in the environment there are multiple actors who
have worked on scraping social networks and discovering myths around sexual and
UNFPA
reproductive health, actors who can help others not to make the same mistakes and
move forward in an optimal way in the development of the project. Before you invent
something "new" look around, explore and be inspired by others.
Involve the users in the whole process: Including the users in the process of creating the
project is an essential step that cannot become a barrier to the continuity of the process,
it is important to establish the different objectives that you want to achieve with the users
in each session to make the most of them and to collect all the valuable information, it is
also important to adjust the expectations of the users' participation in each stage of the
process, this will guarantee a better appropriation of the service.
o https://datasketch.github.io/unfpa/
o https://drive.google.com/drive/folders/1qiRe4NkXgEHYb43QCcQB2shMRoLqaWy6
?usp=sharing
• Contact information: Jaime Aguirre (jaaguirre@unfpa.org)
Project 2: ECHO: Amplifying citizen’s voices for the SDGs”

• Project Description: ECHO is a unique tool that uses Automatic Speech Recognition,
Cognitive Computing, and Data Analytics to improve the efficiency in processing large
amounts of information in real-time. ECHO collects information from individuals of all
backgrounds, including minorities and vulnerable populations
ECHO is a tool powered by artificial intelligence that promotes citizens’ participatory
planning and awareness about the SDGs through real-time guided public discussion.
ECHO is seeking to link conversational and informal citizen’s language to SDGs language
using a classification model, developed by UNFPA Colombia. After the first phase of the
implementation of the interviews in Medellin through ECHO, we obtained, among other
things:
o More than 4,800 guided discussions were carried out, whose results in 56.22% were
performed in women and 43.8% in men. Of the total number of respondents, 44.8%
were young and 18.3% were older adults.
o A pact for the SDGs was signed by more than 10 public institutions in Antioquia.
These entities include Antioquia Governorate, National Police, EPM, Medellin Metro,
Metropolitan Area of the Aburrá Valley, Inder, Medellín City Council, Teleantioquia
and TeleMedellín, which makes Medellín the first city in Colombia to use AI to make
better public policies and make deep commitments around the 2030 Agenda.
o A draft of a Dashboard of the SDGs related to the main concerns of the people in
these areas. It also contains a call to action and "What to do" related to the SDGs that
resulted from the analysis.
o Data collection campaigns have been conducted in many new cities such as Cartagena,
Villavicencio, and the Venezuelan immigrant population in Medellin. Thus obtaining
more than 3,000, 15,000 and 1,200 new testimonies in each city respectively.
The process will cover groups of populations of Medellin, Bogotá and Cartagena, three
different zones and two of the largest cities of Colombia.
• Project Domain: Participatory planning, Freedom of Speech
• AI approach: Automatic Speech Recognition, Cognitive Computing, Natural Language
Processing (NLP)
• Technology: GraphQL, Deep Learning IA, live speech to text
205
• Datasets: ECHO stores all recorded and processed voice information (with NLP
technology) in a structured manner. This then involves a large amount of information from
recorded voice testimonials converted to text, stored as documents in noSQL databases.
• Related Sustainable Development Goals (SDGs): All SDGs, especially SDG 17 –
• Partnership(s)/Collaborator(s): UNFPA Innovation Secretariat, Antioquia Governorate,
National Police, EPM, Medellin Metro, Metropolitan Area of the Aburrá Valley, Inder,
Medellín City Council, Teleantioquia, TeleMedellín
• Links and Multimedia: www.echo -vis -2020.herokuapp.com, Echo Interviewer System:
www.echo.carinalab.co. If you are interested and want to get access please write an email
to nieto@unfpa.org, https://echo.carinalab.co/#/mMedellin1, Social Media Data Scraper
(Beta): http://165.227.124.98/tweetsunfpa/
• Challenges: a) The urgency to achieve our organizational priority areas and leverage
the power of AI toward that end. b) The prevalence of myths and misconceptions about
contraceptives. c) The prevalence data and visualization. d) Humanitarian crisis: internal
and external migration
• Opportunities: a) To accelerate our mandate through AI and cutting-edge technologies.
b) The improve our impact including BC and C4D campaigns through among others the
internet and social network messages, radio, public tv channels, public spaces, etc.
Project 3: Big Data for Family Planning Inequalities Analysis (BiDaFPInAs)

• Project Description: The Big Data for Family Planning Inequalities Analysis (BiDaFPInAs)
project aims to capture social media conversations of Filipinos about family planning and
develop a machine learning (ML) algorithm that converts those conversations to insights.
These insights will help provide real-time data on awareness and perception of Filipinos
about family planning to government planners. Using appropriate keywords and filters,
the system is programmed through a series of iterations to identify levels of awareness
and perception of Filipinos on family planning.
• Department/Division: UNFPA Philippines
• Project Type/Output: Dataset and Analysis, Data Visualisation and Dashboard
Through BiDaFPInAs, a social media dashboard was developed to capture and broaden
the scope of data on Family Planning among key population groups. Social media data
was harnessed to provide insights that improve the planning, targeting and execution of
Family Planning programs in the Philippines.
The initial phase of analysis explored the wider dataset that included publicly available
tweets, posts, and comments; however, this approach led to poor data quality that made
the management of data and the generation of useful insights very challenging. By
working with an implementing partner and an academic expert, and with a focus on the
data from Reproductive Health Care Info, a Facebook community dedicated to family
planning (FP), the analysis of FP-related social media posts was completed at the end of
2021. Another challenge was the difficulty in accessing and using social media data due
to data privacy concerns.
The evidence generated from the analysis validated social media as a means to connect
Filipinos into several networks, which then facilitated knowledge generation and sharing
as to the benefits and disadvantages of FP. More importantly, it suggests the usability
of the Big Data Sentiment Analysis platform and lifewise recommends a more targeted
206
approach of pre- identifying family planning related social media accounts in order to
more effectively capture and analyse the needed data.
• Reported as part of 2022 Compendium on AI Activities? Yes
UNFPA
• Datasets: Social Media (Twitter and Facebook)
• Partnerships and Collaborators: Philippine Legislators Council for Population and
Development (PLCPD)
• Related Sustainable Development Goals: SDG 3 (Good Health and Well Being), SDG 5
(Achieve Gender Equality and Empower All Women and Girls)
Project 4: AI based Early Warning System (EWS) on Youth and SRHR

• Project Description: The early warning system is aimed to provide UNFPA access to
real time information on dialogues, social media behaviour, reported and/documented
programme interventions by development partners with regards to adolescents and
young people. This data is expected to be analysed by an AI algorithm to predict rapid
and critical response to address the root causes and/or prevent further escalation of
the early signs of opposition/backlash against ASRH deduced from the data analysis.
Additionally, we seek to bridge the gap between knowledge and (social media)
behaviours by shifting attitudes among the target population group and subsequently
adopting positive behaviour choices.
Young people, who are connected to the internet, use social media to share their thoughts,
feelings and impressions about various SRHR issues. Some of the conversations include
sexual behaviour, contraceptive use, intergenerational relationships, peer violence.
Moreover, social media is often used by opposition groups against SRHR and CSE, this
tool will facilitate timely action in response to growing opposition before it escalates and
allow relevant UNFPA’s offices to counteract misconceptions and opposition to SRHR.
UNFPA have identified three middle income countries (current suggestions include:
Zambia, to largely track opposition opinions to SRHR; Namibia and Rwanda, due to their
access to technology and smaller population) implementing the Safeguard Young People
Programme with high internet usage especially among young people under 30 years to
form a part of the pilot implementation of the early warning system.
The developed AI based system will be applied to Twitter and/or Facebook during the
pilot phase to generate data to train the AI algorithm to familiarise with the conversations.
Initially, UNFPA will facilitate conversations through targeted questions to the young
people in the target group to test the response of the developed algorithm. In addition
to this, we will feed the AI system UNFPA program documents, reports, intervention
plans etc. to align the discussions to program response. It is expected that the system
will generate reports with recommendations of urgent action in the form of graphs,
infographics, word clouds etc. for each use of the information.
A front facing visualisation platform will be developed to aggregate and present the
reports in a user friendly manner. This visualisation platform will be connected to existing
UNFPA tools including social media platforms to make the generated report available to
all stakeholders.
• Department/Division: Youth and Innovation Units UNFPA ESARO
• Project Type/Output: Data visualisation, AI Algorithm
• Datasets: Social Media (twitter and facebook), publicly available data
• Project updates: AI Platform currently being developed.
207
• Project Domain: Currently being developed

• Related Sustainable Development Goals (SDGs): SDG 3, 4, 5, 9
o Safeguard Young People Project - Funding the project

o Private Sector AI developer
• Links and Multimedia: NA

• Contact information: Sydney Hushie- hushie@unfpa.org , Renata Tallarico- tallarico@unfpa
.org, Isabelle Jost - jost@unfpa.org
Project 5: Hayati Chatbot

• Project Description: ‘Hayati’, meaning ‘my life’ is a mobile application being developed by
Men4Women in collaboration with UNFPA South Sudan. The chatbot is a way forward for
South Sudan given the increase in technology and phone user uptake across the country.
The chatbot is critical for the need to provide young people with a platform where they
can access SRH and GBV information confidentially from wherever they are.
In future, the chatbot also plans to incorporate live information such as reporting of sexual
harassment in public spaces etcetera. But currently it will only look at providing free,
timely and confidential information on SRH and GBV to the people.
Key features of the App:
The chatbot is in a form of an Android based Mobile Application with integrated services/
features which includes:
1. Pinning location from the app with corresponding reason(s)/ information the person
seeks to receive.
2. Integrated Interactive Voice Response (IVR) through the National Helplines 623, 662
and 885.
3. Canned/ automatic responses linked to the FAQs. Saves time when answering
common questions.
4. Knowledge base or frequently-ask-questions integration. This allows you to incorporate
topics of discussion into a support section that can be used for instant answers and to
troubleshoot problem areas of the website/app.
5. Ticket creation and management. This allows visitors to submit a request after business
hours. Ideally, it includes auto follow-ups to remind visitors you are waiting for their
reply.
6. Support and agent ratings to identify areas needing improvement.
7. Visitor tracking. By seeing the pages, the visitor is on or has visited, requests can better
be understood.
8. Analytics integration so you can see which services were assisted by chat. It should
also provide periodic reports
9. Chat tags and keywords to help you find common issues and questions.
10. Mobile dashboard or app, to monitor chats while on-the-go.
11. Visitor banning, to minimize “trolls” from taking up your time.
12. Translation features- It is linked to Google Translate, which is free and can be helpful
for the Arabic speakers.
13. Transcript of emails can be done
14. Long-term logging to extend visitors’ visiting journey.
15. Chat log and data exporting feature.
16. Chat transfers among agents. This eliminates the need for the visitor to repeat her or
himself.
208
17. Emoji support.

18. File upload support to help visitors describe what they need.
UNFPA
• Department/Division: UNFPA South Sudan and Innovation Unit UNFPA ESARO
• Project Type/Output: Data visualisation, AI Chatbot
• Project updates: AI Chatbot currently being developed.
• Project Domain: Currently being developed
o Men4Women - Youth tech startup supported by UNFPA South Sudan

• Contact information: Shruti upadhyay - shupadhyay@unfpa.org, Sydney Hushie- hushie@
unfpa.org
Project 6: SophiBot
• Project Description: Sophie Bot is an AI application that directly answers SRHR questions
in a private and confidential way. The information is pre-fed in the platform and in
instances where no response is provided, the user is linked to a direct chat where they
are able to communicate with someone to receive the correct responses and also join
the discussion forum. Sophie Bot can also be accessed through Messenger, Telegram
or Twitter. The solution seeks to address poor SRH outcomes among young people
attributed to inadequate access to comprehensive and correct information on SRH. In
Kenya, it is estimated that one in every five teenage girls between the ages of 15-19 years
is either pregnant or has had their first child while 51% of new HIV infections occur among
young people between the ages of 15-24.
• Department/Division: UNFPA Kenya
• Project Type/Output: AI Chatbot
• Project End Year: NA
• Project updates: Major results since the launch of SophieBot in December 2016:
1. SophieBot has been downloaded 1700 times;

2. 38,549 questions have been asked across all platforms (most of them through the
app and Facebook);
3. $40,000 raised through the KCB Lions Den Show;
4. Enrolled in the Merck Accelerator Program (3 months acceleration cycle);
5. Awarded the Microsoft Insiders4Good East Africa Fellowship, which grants support
from Microsoft and the broader Windows Insider community including hardware,
software, tailored mentorship and access to Microsoft’s global network;
6. Enrolled in the SRHR Africa Trust leadership programme for building capacity for
advocacy in sexual and reproductive health rights.
• Project Domain: Youth, Reproductive Health

209
1. Nailab
2. National AIDS Council - Kenya,
3. Discover JKUAT
o http://misssophiebot.com/
o https://www.theguardian.com/careers/2018/feb/21/sex-education-at-the-push-of-a
-button-the-apps-changing-lives-worldwide
• Contact information: Shruti upadhyay - shupadhyay@unfpa.org, Sydney Hushie- hushie@

unfpa.org
Project 7: #TECH4YOUTH AND #NLP SOLUTIONS TO AMPLIFY SRHR ACCESS AND

ACCELERATE THE ACHIEVEMENT OF #SDG3
• Project Description: The number of languages spoken in Africa varies between 1,000
to 2,500, depending on different estimates and definitions. Almost half (48 per cent)
of Sub-Saharan African countries have an African language that is spoken by over 50
per cent of the population as a mother tongue. With the additional secondary speakers
sometimes at mother-tongue proficiency level, the proportion increases to more than
two-thirds (67 per cent). Sixteen of Africa’s shared cross-border languages have more
than 150 million speakers. Outside the education sector, at least 56 African languages
are used in administration and at least 63 African languages are used in the judicial
system (26 sub-Saharan nations allow African languages in legislation). In written business
communication, at least 66 African languages are used, and at least 242 African languages
are used in the mass media. In short, the existence of so many languages within a single
country and their right not only to survival but also to development represent a matter of
importance that has to be considered over and above the categories into which they fall.
This diversity is in itself perceived as an inherent challenge in matters of communication,
governance and education. Such a multiplicity is perceived as a communication barrier
and viewed as synonymous with conflicts and tension but also a great opportunity to
spread messages and information till the last mile.
• Five UNFPA country offices, BENIN, TOGO, GHANA, BURKINA FASO, and NIGERIA, have
decided to work collaboratively to use technology-based solutions in Adolescent and
Youth Sexual
• and Reproductive Health (AYSRH) to improve knowledge and access to information. In
order to reach all youth, especially those left behind, the innovation and technology team
set out to
• integrate natural language processing (NLP) in four local languages (Fon-Gbe, Haussa,
Fulfulde, and Yoruba) into their existing digital solutions.
• Department/Division: UNFPA Benin, Burkina Faso, Ghana, Nigeria and Togo
• Project Type/Output: Natural language processing, AI Algorithm
• Datasets: 25 hours of professional Hausa audio for text to speech, french and english with
African accent licences
• Project Domain: SRHR education
• UN partner: ITU
210
• Partnership(s)/Collaborator(s)/: Tsinghua University, Dorabot, W.B, European Institute of

technology, Voxygen, Etrilabs, Africa Design School
• Project updates : Due to the lack of datas on african languages and to accelerate the
UNFPA
development of this NLP solution, we are establishing a research & development
partnership with European Institute of technology on BIG Data and AI issues so that,
through its specialized Data/AI Masters, students can undertake research specifically on
automatic translations into natural language support from UNFPA Benin.
• Links and Multimedia: Tech4Youth
• Contact information: Wilfried ROUAMBA- rouamba@unfpa.org , Djawad RAMANOU
-djawad@unfpa.org, Wilfried GOSSAN- gossan@unfpa.org
Project 8: SAS & UNFPA Partnership

The overall goal of this partnership is to: (1) increase awareness of GBV, in particular
targeting the business community in the region; and (2) develop solutions to analyze GBV
using data and artificial intelligence to tailor policies on GBV prevention.
As part of the second goal, SAS and UNFPA will jointly explore how to leverage and
analyse large administrative data sets to understand the risk and protective factors using
AI, and explore a model that can be used by the public sector and civil society to prevent
the incidence of violence against women, and improve GBV response programmes.
A proof of concept will be developed and showcased for local authorities and public
opinion using available data sources such as for example survivors’ interviews and police
records, and the analytical model (AI) will be applied to identify risk factors that will inform
prevention plans at local and national level.
SAS will convene its internal data scientists, as well as data scientists from SAS customer
companies, to explore data sets and analytical techniques to assess what can be learned
from data to prevent GBV, improve services for survivors and optimise GBV programs.
SAS will also bring its business client base to the advocacy component of the partnership.
UNFPA will liaise with its country office staff and partners throughout the region to identify
countries and public sector bodies who would partner in this project and share data from
its own GBV programs. UNFPA will also provide technical expertise and support accurate
and robust messaging around the topic of GBV.
Both organizations will also collaborate on and support each other messaging or
communications campaigns to elevate public awareness of GBV.
Background of AI technology and SAS projects:
o The first phase of SAS technology was based on improving the existing system to
ensure that every woman is given the correct risk factor and receives the correct
support and protection. The data included information from the victims, complaints
made and the different risk assessments, both at the time of registration of the case
and during its evolution.
o The studies compiled a range of indicators regarding the psychological profile of
the perpetrators and the vulnerability of the victims. This includes information such
as suicide attempts, addictions, and the perpetrator's family history, giving a total of
more than 50 indicators. This provided a very large data pool to be used for modeling
purposes.
o After initial data processing, a two-stage modeling strategy was chosen to manage
the different levels of information received.
o This includes drawing on the experience and professionalism of police officers in
their reports. Using only these reports, a predictive model was created that assigns
the probability of recidivism. In the second stage, an analytical model was developed
using the probability generated in the first stage and offender-related indicators were
added to assign the eventual recidivism probability.
211
o Overall, a machine learning model was found to help reduce the risk of recidivism,
and AI technology was found to be actionable by authorities. It is believed that the
effective use of predictive analytics and machine learning can help prevent a large
number of cases of gender-based violence.
• Department/Division: UNFPA Eastern Europe & Central Asia Regional Office (EECARO)
• Project Type/Output: Partnership on Development of Data Analysis Methods and AI
Technology to Adapt Policies for GBV Prevention
o Three webinars were organized to introduce the project to its potential stakeholders;
o The project is in its initial stage of development and SAS is developing a white paper
to agree on specific outputs
• Project Domain: Implementing AI Technology and Advanced Data Analysis Methods to

Prevent Gender-Based Violence
• Related Sustainable Development Goals (SDGs): SDG 5 – Gender Equality
1. Private Sector – SAS

SAS is a private US-based global company, operating for more than 40 years, offering
the most advanced analytics and AI solutions to its clients.
o https://eeca.unfpa.org/en/news/unfpa-and-sas-partner-work-preventing-gender
-based-violence-power-ai
• Contact information: Gabriela Alvarez-Minte, UNFPA EECARO, alvarezminte@unfpa.org
Project 9: AMMA App - Period & Pregnancy Tracker

• Project Description: ''Amma'' is an international mobile application that allows women
and their families to have an informed, safe, and healthy pregnancy. The application,
which continues to work in cooperation with UNFPA Eastern Europe and Central Asia
Regional Office (EECARO), provides information sharing in a wide scope. Within the
application, detailed information can be obtained about the development of the baby,
the changes that the mother may experience during pregnancy, and nutrition & moods.
In addition, with the AI technology created, the weight gain and abdominal growth of
the expectant mother are controlled, and pregnant women can count the contractions
they experience and send their data to their doctors. The application also contains high-
quality ultrasound images and explanations. In addition, there is a personal calendar for
each day of pregnancy with reminders for medications and mood changes.
With the support of UNFPA EECARO, additional content in mobile applications is created
to convey more accurate information on safe pregnancy, prenatal care, and sexual and
reproductive health. With the AI technology to be developed, it is aimed to conduct joint
research to understand user information, behavior, and information needs on maternal
health, sexual reproductive health, and gender issues.
• Department/Division: UNFPA Eastern Europe & Central Asia Regional Office (EECARO)
• Project Type/Output: Mobile Application
212

UNFPA
o Amma currently has 1.5 million active users. The partnership with UNFPA is quite new
(officially launched in 2022), however UNFPA has been able to feed quality SRHR-
related content to the application, which has potentially reached the users of the app.
UNFPA plans to expand the partnership and tap into the potential the app provides
to reach users with accurate SRHR information and data.
o The Amma app is downloaded by 10 mln pregnant women every year (or 7% of all
pregnant women in the world). As content is provided in 13 languages, there are users
in almost all countries of the world.
o With UNFPA support, 3 articles have been published and 12 more topics will be
covered before the end of the year. So far, the published articles have had close to
5,000 active engagement with Amma users.
• Project Domain: Education; Sexual and Reproductive Health

• Related Sustainable Development Goals (SDGs): SDG 3, 4, 5
o Private Sector (Financial Support) - PERIOD TRACKER & PREGNANCY AND BABY
CALENDAR LIMITED
o https://pregnancytracker.app/
o https://amma.family/#landing
• Contact information:
o Tamar Khomasuridze - khomasuridze@unfpa.org, UNFPA EECARO
Project 10: Just Ask! Khulke Poocho! AI Chatbot

• Project Description: UNFPA India's Just Ask!: The 'Just Ask!' Chatbot, launched by
UNFPA in India, is an AI-driven digital platform tailored for adolescents, young adults
and couples, aimed at enhancing their understanding of sexual and reproductive health
and rights. This initiative offers a safe, confidential, personalized, and non-judgmental
space for individuals aged 15-29 to access reliable information, dispel misconceptions,
and seek services related to their sexual and reproductive well-being. The chatbot is part
of the project “My Rights, My Choices” aimed to increase awareness and access to SRHR
services and is co-financed by Bayer.
The digital engagement platform is designed on the principles of behavioral science, user-
centered design, data security and Artificial Intelligence aligned to the UN 2.0 Quintet
of Change. The Chatbot has been developed by UNFPA experts and was launched in
collaboration with National Health Mission (NHM), State Government of Madhya Pradesh
in July 2023.
The chatbot is designed to encourage informed decision-making and foster interest
in often misunderstood or stigmatized topics such as puberty, menstruation, sexual
and gender identity, safe sex, and contraception. The platform will also introduce more
topics such as safe cyber use (highlight areas around TF-GBV) and mental health in
2024. Adhering to UNFPA's guidelines for the safe and ethical use of technology, the
chatbot prioritizes user data privacy in its design, ensuring a secure and responsible
digital experience for its users. This project reflects UNFPA's commitment to using digital
innovations as an accelerator to promote sexual and reproductive health among young
213
people, particularly in settings where access to accurate and comprehensive health

information is limited.
• Location/Country Office: UNFPA India
• Department/Division: Innovation Unit in collaboration with Programme, State and
Resource Mobilization & Partnership Units
• Project Type/Output: My Rights, My Choice - Increasing awareness and access to SRHR
services
• Datasets:
o Content - The content for the AI chatbot was designed in accordance with the
Ministry of Health & Family Welfare (MoHFW), Government of India guidelines and
International Sexuality Guidelines. The dataset was validated by SRHR experts and
agencies specialized in this domain. Currently, the chatbot has accumulated over 2.8
million user queries and continues to enhance its dataset regularly.
o Technology - The chatbot is built on an open source AI-tech stack and operates within
the predefined content boundaries. It has been deployed on WhatsApp platform
considering its wide user base in India.
• Project Status: On going

• Project Domain: SRHR and Innovations
• UN partner: none
• Partnership(s)/Collaborator(s)/: Bayer AG, Amazon Web Services (AWS), State
Government of Madhya Pradesh and Rajasthan (India)
o Co-designed and Piloted in 2023 with 500 adolescent and young adults.
o Currently scale up in 2 States with aim of reaching 1mn users by 2025 across India.
• Links and Multimedia: https://www.unfpa.org/updates/india-unfpa-launches-just-ask

-chatbot-sexual-and-reproductive-health-and-rights
• Contact information: Sahil Kapoor, kapoor@unfpa.org, Swati R Dutt, dutt@unfpa
.org
Project 11: NeMa Smartbot in India

• Project Description: The Maternity Foundation, in partnership with Finnish AI startup
Neuvo Inc. Global and facilitated by the United Nations Population Fund (UNFPA), has
introduced an innovative AI tool, the NeMa smartbot, to enhance midwifery skills and
knowledge in low-resource settings. This tool is integrated into the Maternity Foundation's
Safe Delivery App, a resource already reaching over 375,000 healthcare workers globally,
to provide real-time, evidence-based answers on pregnancy and childbirth. Designed
with an ethical approach to AI and data, the smartbot functions offline and uses clinically
verified material, ensuring reliable access in areas with limited connectivity. The initial
pilot, involving healthcare professionals in Punjab, India, and financed by Organon, marks
a significant advancement in using technology to improve maternal and newborn health,
especially in underprivileged regions. UNFPA's facilitation of this project underscores
its commitment to reducing global maternal and newborn mortality and highlights
the potential for AI to transform healthcare education and delivery in challenging
environments.
• Location/Country Office: India
• Department/Division:
214
• Project Type/Output: AI enabled capacitation for midwives

• Datasets:
UNFPA
• Project Domain: SRHR
• UN partner: UNFPA
• Partnership(s)/Collaborator(s)/: Maternity Foundation, Neuvo Inc.
o Links and Multimedia: https://www.unfpa.org/updates/unfpa-and-organon-support

-maternity-foundation-using-ai-strengthen-midwifery-skills
o https://www.maternity.dk/maternity-foundation-and-neuvo-inc-global-launch-new-ai
-tool-to-ensure-safer-childbirths-in-low-resource-settings/
o Maternity Foundation: Head of Communications, Helle Degn: helle@maternity.dk /

www.maternity.dk
o Neuvo Inc. Global: Data Ethics and Communications Specialist, Ada Kangas: ada.
kangas@neuvo.ai / https://neuvo.ai/
o UNFPA: Chief, Innovation Unit, Nigina Muntean: muntean@unfpa.org / https://www
.unfpa.org/innovation
Project 12: BabyChecker

• Project Description: The UNFPA Honduras Country Office is utilizing advanced AI
technology to address maternal health challenges, especially in remote areas with
limited healthcare services. The AI-powered tool, "BabyChecker", enables individuals
without specialized knowledge in prenatal care to effectively conduct ultrasounds on
pregnant women. This technology can estimate the gestational age of the fetus, detect
whether there is more than one fetus, determine potential delivery risks due to fetal
positioning, and identify early warning signs that require specialist referral. This initiative
aims to significantly enhance maternal and prenatal care in resource-limited settings,
demonstrating the potential of AI to bridge gaps in healthcare accessibility and quality.
• Location/Country Office: Tegucigalpa, Honduras
• Department/Division: Sexual Reproductive Health component
• Project Type/Output: Pilot Project for the use of AI in medical services to Reduce Maternal
Mortality
• Datasets:
• Project Start Year: Initiated in 2023 with a formal agreement between the Ministry of
Health of Honduras (SESAL), Delft Imaging and UNFPA to initiate the use of BabyChecker,
an AI application to collect data and help utilize the use of the equipment to detect early
signs of potential risks for the mother and the fetus.
• Project Domain: Country level in Honduras
• UN partner: UNFPA Honduras country office
• Partnership(s)/Collaborator(s)/: Delft Imaging https://www.delft.care
215

o Jose Manuel Pérez - joperez@unfpa.org

o Silvia Padilla - spadilla@unfpa.org
Project 13: UNFPA ESARO's iSAY

• Project Description: UNFPA's East and Southern Africa Regional Office's AI-based
platform, iSAY - Intelligent Sentiment Analysis on Youth and SRHR Platform, employs
advanced machine learning and natural language processing to monitor and analyze
public dialogues and social media trends related to adolescent sexual and reproductive
health and rights (ASRHR). This innovative platform filters relevant content, classifies
opinions and information into categories, and extracts key insights, enabling UNFPA
to make real-time program predictions, follow-ups, and recommendations based on
credible data from diverse sources. iSAY's capabilities include temporal analysis of
conversations, intent detection, sentiment analysis, gender gap analysis, age filtering, and
geographical analysis, making it a comprehensive tool for understanding and addressing
ASRHR issues. Additionally, its expandable nature allows for potential application in other
areas, reflecting its utility across UNFPA's various mandate areas.
While the dashboard developed in 2022 is only accessible by selected UNFPA ESARO
staff, in December 2023 a semi-public platform will be accessible for all UNFPA staff in
the ESARO region.
• Location/Country Office: UNFPA ESARO - covering all 23 countries in the region
• Department/Division: ESARO Youth and Innovation unit
• Project Type/Output: Data Visualization
• Datasets:
o Data is collected daily from online conversations in publicly available sources, including
Twitter, Facebook public pages, online forums, news comments, and blogs.
o More specifically, from each data source we collect the following information:
• Twitter: All the comments in the country of interest that mention the defined keywords.
• Facebook: We select a list of Facebook Public Pages of interest and analyse the comments
on those pages.
• Web content: Comments on websites, news, blogs and forums that mention the defined
keywords.
• Project End Year: Current licence ending December 2024
• Reported as part of 2022 Compendium on UN AI Activities? no
• Project Domain: https://dashboard.citibeats.com/beta/organization/citibeats-projects/
project/sexual-and-reproductive-hea-ji/transparency/(not publicly available)
• UN partner: none
• Partnership(s)/Collaborator(s)/: Citibeats (service provider)
• Project updates: NA
• Contact information: jost@unfpa.org and hushie@unfpa.org

All SDGs: SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17
216
3. Relevant Links
https://www.unfpa.org/
UNFPA
Contact information
Jaime Aguirre (jaaguirre@unfpa.org)
217
United Nations High Commissioner for Refugees
Project 1: Azure Open AI – Generative AI-powered chatbot

• Project Description: UNHCR’s Microsoft Azure Open AI solutions aim to provide
personalized and individual chatbots for different use cases with an agile deployment
approach. The chatbots are developed as internal “sandboxes” for operations: they are
individual website instances that serve as “playgrounds” for teams to safely test generative
AI capabilities, such as talking to internal pools of documents and producing new content
based on the input (user prompts).
The pilot phase for these sandboxes encompasses more than fifty use cases, representing
a diverse range of operational scenarios, including country operations, regional teams,
and inter-agency work at our headquarters (e.g., HR Policy chatbot, ERP Job Aid chatbot).
For the 50+ use cases, we deployed more than 20 chatbots in 2023 with an interface
where colleagues can chat with their own data to generate content, find information,
create summaries, discover patterns, and more. All case studies aim to leverage their
text-based data such as:
o Documents/Text most representative to HQ teams: 1) Legal documentation, 2) Policies,

rules, and regulations, 3) Reports: evaluation, financial audit, universal periodic review,
oversight, submissions proposals, 4) Talking Points, etc
o Text from Regional teams: 1) Query situational reports to find incidents or trends, 2)
Summarize data (e.g., refugee case interviews), to mention some
o Text from Country Operations: 1) Security notes, 2) Mission reports, 3) Protection
monitoring documents, cluster reports, 4) Translations of original language reports,
4) Participatory data.
UNHCR also adopted Copilot in Bing in the course of 2023, enhancing the capabilities
of the chatbot instances with additional information and content from the web. Copilot
in Bing is rolled out as a secure way to use Generative AI across UNHCR.
• Department/Division: Division of Information Systems and Telecommunications (DIST),
Innovation Service, UNHCR
• Project Type/Output: Web-based GAI chatbot for individual business units
• Project Domain: Generative AI (GAI), RAG (Retrieval-augmented generation), Foced
displacement and policy
• Data Source: In the interest of agility, UNHCR’s business units’ non-sensitive data.
Exceptions are made for use cases with sensitive/confidential data.
• Data publicly available: No
• Technology/Platform: Generative AI, GPT 3.5 turbo and 4.0 from Azure OpenAI
(Microsoft), Web site instances are hosted in a dedicated Microsoft Azure subscription
connected to UNHCR corporate Microsoft 365 tenant.
218

– Peace, Justice, and Strong Institutions; SDG 17 – Partnership for the Goals
UNHCR
• Partnership(s)/Collaborator(s): Microsoft, OpenAI/GPT
o OCHA publication (UNHCR case study portrayed): https://reliefweb.int/report/world/

generative-ai-humanitarians-september-2023
• Lesson Learned:
o Reskilling personnel: on prompt completion, limitations of GAI (e.g., quantification of

text-based inputs), and ethical/human rights considerations on Automated Decision-
Making Systems (ADMs)
o Governance: organization-wide solutions on GAI have high costs associated with it
- Directcosts: IT infrastructure, indexing costs plus staff time (semi-automated and

sometimes manual dataset curation process), prompt completion (currently pay-
as-you-go), website hosting, semantic searches costs.
- Indirect: Staff time for training and implementation. This project has a
multidisciplinary team (mostly working part-time): a) Project management
dedicated personnel, b) business analyst (for curating/selecting case studies,
solution engineering/prompt engineer, c) data scientist, d) AI/data engineer, e)
ML DevOps engineer (cognitive search), f) Web/Full stack developmentand g)
policy researcher.
o Data protection: Microsoft's framework agreement with the UN includes privileges

and immunities for data hosting.
o Capacity building: the user feedback from walk-through sessions confirms the need
for enhanced capabilities for prompt engineering to get the best results out of Gen
AI.
o Tailored prompt engineering and follow-up/testing support facilitates successful
adoption by business partners.
o The approach for us needs to be agile and follow innovation methodologies (e.g.,
scrum methodologies, agile DevOps)
• Contact information: Nicole Henderson hendersn@unhcr.org , Carolynn Oleniuk oleniuk@

unhcr.org, Marc Schittek schittek@unhcr.org
Project 2: Monitoring hate speech with data-driven alert system

• Project Description: UNHCR’s HateFree project aims to provide an internal digital platform
for humanitarians, allowing them to easily monitor, analyze, and archive hate speech
targeting refugees, stateless, and forcibly displaced people on social media. The platform
offers humanitarians three main tools:
o A dashboard, where harmful activities online can be easily researched and analyzed
with various interactive data visualizations and filters based on sentiment, date,
language, and origin. Research parameters can be saved or results downloaded for
more in-depth analysis.
o A situation page acts as a centralized space to help UNHCR colleagues coordinate their
understanding, response, and tracking of specific cases of hate speech. Colleagues
can access background information, updated if-asked lines, focal points information
and directly upload evidence of hate on the platform.
o An alert system that serves as a complement to the situation page, where colleagues
can subscribe to situation of interest to them and receive alerts and notifications when
219
there is a shift in tone of conversation, sudden spikes, or influential mentions made

within that conversation.
The digital platform sits within UNHCR’s domain, and is built on AWS, to facilitate the
scalability and flexibility of the project, with the intention of enabling a deeper and wider
analysis of harmful online conversations by allowing easy expansion of the platform with
new tasks, models and languages. The platform currently hosts a bespoke scalable AI
model allowing the detection of hate speech targeting refugees in English. As of now,
three additional AI models are being developed or have been developed to compliment
the initial model, including a Spanish AI hate speech model, a thematic analysis model
and a sentiment analysis model.
• Department/Division: UNHCR Innovation Service, Data Innovation Fund 2022-2023,
UNHCR Division of External Relations (DER)
• Project Type/Output: In-house digital platform / web application. Two AI models for
detection of hate speech targeting refugees (English, Spanish), one Spanish hate speech
against refugees dataset, one AI model for sentiment analysis, one AI model for thematic
analysis.
• Project Domain: Forced displacement, Human rights,
• Data Source: Meltwater
• Technology/Platform: Python, Transformer models (BERT), Large Language Models
(Llama2), React/Next, Django, AWS, GCP, Huggingface. The platform consists of a
frontend, backend and a data pipeline and is built on a serverless architecture, using
AWS Lambda, AWS S3, AWS SQS, AWS SNS, and AWS RDS Postgres, AWS Fargate
and AWS Amplify. The machine learning models are sourced in 3 ways: Using existing
open-source models, retraining existing models with open-source data as well as data
created internally or in collaboration with other teams and academia. The models are
deployed using AWS SageMaker, and the predictions are made using AWS Lambda. The
predictions are then stored in a database, and the results are displayed in the frontend.
• Partnership(s)/Collaborator(s): MDH ECHO Project, DPO
o Project site: not public (hatefree.unhcr.org)
• Lesson Learned:
o Scale of project is hard to keep: Despite starting off as a project with a very limited scale
(i.e.: supporting the Analytics unit monitoring hate speech as per the SG’s Strategy and
Action plan), it became evident from early consultations with additional teams across
the organisation that the platform could be a great asset for other UNHCR colleagues,
thus widening the scope of HateFree and the workload necessary to make it come to
fruition.
o Expertise and staffing is difficult to gain / retain: Project relied heavily on expertise
not available or not part of the team’s core work. Acquiring NLP experts and full stack
developers for such a timebound project was very difficult, as the team was looking
for senior expertise, putting the team in a competitive hiring environment with private
sector companies and conditions. Overall, this seems to be a problem across the UN
system, with many teams building similar tools having shared the same challenges.
220
o This project is still on-going and will be developing for the year to come, so additional
lessons will inevitably be learnt in the coming months. The ones shared above seemed
to be the most important challenges to mention.
UNHCR
• Contact information: Patricia Fabi fabi@unhcr.org, David Howes howes@unhcr.org,
Kwabena Adwabour adwabour@unhcr.org
Project 3: Machine learning to boost digital education uptake and impact

• Project Description: Learning Equality and UNHCR have been working to enable
streamlined curriculum digitization and alignment processes, helping edtech platforms
and implementers to source and organize relevant materials, educational content
providers to contextualize their work to serve new audiences, and curricular bodies to
understand gaps and similarities across curricular standards. We’ve developed end-to-
end processes with human-in-loop workflows. Using generative AI, we’ve automated the
process of digitizing curricular standards from raw source documents into a machine-
readable format that is usable within an educational platform. Next, we co-hosted a Kaggle
machine learning competition to develop initial curricular recommendation machine
learning models that match relevant content items from a large library to particular
learning objectives in a curriculum. These workflows are currently being prototyped.
• Department/Division: UNHCR Innovation Service, Data Innovation Fund 2022-2023,
UNHCR Education Unit
• Project Type/Output: A content recommender tool that can make recommended
matches for digital content according to curricular standards.
• Project Domain: Generative AI (GAI), Forced displacement
• Data Source: Metadata (titles. descriptions, topic trees, and text) from aligned content in
the Kolibri Library
• Data publicly available: https://www.kaggle.com/competitions/learning-equality
-curriculum-recommendations/data
• Technology/Platform: Open Source, Python, OpenAI’s GPT-4 API
• Partnership(s)/Collaborator(s): Learning Equality (project developer)
o Kaggle competition: https://www.kaggle.com/competitions/learning-equality

-curriculum-recommendations
o GitHub documentation: https://github.com/learningequality
• Lesson Learned:
o Value of competition: Kaggle convenes an active community of machine learning

specialists who are primed for engagement in competitions to solve different problems
using these technologies. The Kaggle community helped to propel this work forward
in a way that would have otherwise taken longer or have been less effective.
o Advancements in generative AI: A few years ago we attempted to digitize curricular
documents through OCR, computer vision, and statistical methods, but it was time
intensive and not generalizable.
• Contact information: Kent Awiti Ojok awiti@unhcr.org,Sofia Kyriazi KYRIAZIS@unhcr.org
221
Project 4: Brazil Predictive Analytics for UNHCR Improved Contingency Planning

• Project Description: This project seeks to forecast the number of people who cross the
border into Brazil and the shelter capacity in Boa Vista and Pacaraima, the Brazil border
with Venezuela. This project utilises interlinked tools that consists of three computer-
based solutions using different techniques: a queueing modelling tool for simulating
future border crossing scenarios under different conditions, nowcasting model efforts for
estimating the current urban population and potentially identifying interest in population
movements to Brazil using big data sources, and the development of predictive models
for forecasting future arrivals and population movements. This is done using different
tools (e.g., web engines search, surveys, and social media analysis) as well as different
computer and data science techniques (e.g., artificial intelligence, specifically supervised
machine learning for time series analysis).
These tools have been developed in collaboration with operational teams over the past
year and a half, and currently are deployed in the Brazil operation. The desired effect of
the project is increased evidence-informed decision-making in contingency planning,
using scenarios triggered by estimated nowcasting and forecasting figures related to
population movements. These tools have been developed in collaboration with UNHCR
Brazil operational teams over the past year and a half, and currently are deployed in the
Brazil operation. UNHCR Brazil has already used the preliminary findings to advocate and
prepare for arrivals and to have medical services and shelter available according to their
needs. In the past. For example, UNHCR utilised these tools to expanding isolation areas
to treat COVID-19 cases, and UNHCR and its partners have advocated for the delivery of
vaccines according to the level of demand from new arrivals.
• Department/Division: UNHCR Innovation Service
• Project Type/Output: Report, Software tool, Medium Blog Post
• Project Domain: Forced Displacement
• Data Source: Population data: population historical data , Other variables: COVID-19,
market prices, transportation, exchange rates
• Data publicly available: No, partial datasets only. Publicly available: https://data.humdata
.org/group/ven
• Technology/Platform: Open Source - Python application
o UN Partners: UN Global Pulse, UNHCR Brazil
o Project site: not public (https://predictive-analytics-brazil.unhcr.org)

o Academic papers: https://arxiv.org/abs/2303.15614
o Blog post: (blog post 1)
• Lesson Learned: For predictive analytics of mixed-migration and forcibly displaced

populations, it is important to take into consideration existing policies - such as the UNHCR
contingency planning and emergency-related policies- to strengthen ongoing work
and avoid duplication. It is also important to keep in mind that the sporadic nature of
population movements across a number of informal crossings undermines many field
operations’ ability to collect data crucial to their work, which makes the application of
predictive analytics a useful tool that can be one of datasets used to inform decision
making. Additionally, a key lesson learned is the need to assess the risks related to
222
data management – particularly data protection-related. There are risks associated with
predictive/forecasting analytics data/information/variables (inputs), as well as the results/
predictions (forecast outputs). These may be direct or unintended consequences (e.g.,
UNHCR
lack of access to territory/asylum) derived from analytics results. It is important to consider
and adequately mitigate these risks through the implementation of safeguards when
embarking on any predictive analytics project.
• Contact Information: Hovig Etyemezian etyemezi@unhcr.org, Oscar Sanchez Oscar
sanchezo@unhcr.org
Project 5: Project Jetson (Predictive Analytics of Forced Displacement, Somalia)

• Project Description: Project Jetson is UNHCR’s first AI-based predictive analytics project to
predict population flow, specifically the movement of internally displaced persons within
Somalia. Project Jetson also aimed to understand the factors that cause or exacerbated
that forced displacement. This project focused on the development of predictive
models for forecasting future arrivals and population movements in each region. This
is done using a variety of datasets (e.g., conflict, weather/climate anomalies, market
commodity prices, and historical population movement) and utilizes different computer
science and data science techniques (e.g., artificial intelligence, specifically supervised
machine learning for time series analysis). The project follows UNHCR guidance for
data responsibility, data protection, ethics and human rights due diligence, as well
as OCHA peer-review framework for predictive analytics projects. The Innovation Service
intends for Project Jetson to lead to evidence-informed decision-making in contingency
planning, improved humanitarian action through preparedness and risk reduction, and
strengthened protection for those who are forcibly displaced.
• Project Type/Output: Academic paper, Software tool (application)
• Project Status: Project Jetson ran from 2017-2019, but the UNHCR Innovation Service
deploys the project by the request of operations in ned and with similar operational
contexts
• Project Domain: Human Rights, Weather, Conflict, and Forced displacement
• Data Source: PRMN-UNHCR Somalia, ACLED, FAO SWALIM, FAO FSNAU
• Link to data:
o https://unhcr.github.io/dataviz-somalia-prmn/index.html
o https://data.unhcr.org/en/geoservices/
o https://data.humdata.org/group/som
o https://acleddata.com/#/dashboard
o https://dashboard.fsnau.org/
o https://www.faoswalim.org/article/swalim-online-systems-virtual-launch

• Technology/Platform: Open Source, Python Jupyter Notebooks and R Markdown
notebook, R several packages, including R-shiny https://unhcrinnovation.shinyapps.io/
Somalia/
• Project Update: UNHCR undertook a formal procurement process to acquire DataRobot
software www.datarobot.com to test an alternative to open source models and the
models produced great results. While Project Jetson has not had any specific technical
updates, the teams are currently exploring ad hoc support for UNHCR Somalia Operation
in times of crises. The project was featured at a Somalia inter-agency IDP working group
223
meeting in May 2023. Conversations are ongoing both with different members of the
UNHCR in Somalia to identify new trends and assess current needs and capacity, as well
as with researchers at Essex University exploring the link between conflict and climate in
Somalia. Additionally, UNHCR Project Jetson was featured highlighting relevant lessons
for decision makers in the IOM Big Data for Migration Handbook (BD4M), Chapter 5
(2023).
Peace, Justice, and Strong Institutions, and SDG 17 – Partnership for the Goals
o UN Partners: UN Global Pulse

o Civil Society: Uptake Foundation (Data Fellows Program) (Capacity Building), Omdena
Foundation (Challenge Team)
o Academia: University of Essex (Human Rights, Big Data & Technology Project HRBDT)
o The project page for Project Jetson on UNHCR’s website http://jetson.unhcr.org

o The book chapter that explains the journey on Predictive Analytics for UNHCR
o Some academic papers highlighting Jetson (paper 1, paper 2, paper 3 on ethics)
o Some blogs about project jetson (blog 1, blog 2, blog 3, essay)
o The code repositories (repo 1)
o Some media articles highlighting Jetson (here, here, here, here and here in German)
• Lesson Learned: Because this was UNHCR’s first application of predictive analytics
experiment to try to predict movement and understand some of the unique drivers
and indicators of displacement within Somalia and along the Somali-Ethiopian border,
it prompted significant learnings. The most crucial lessons learned highlighted the
importance of engaging the populations you serve; Project Jetson received crucial inputs
from refugees and IDPs in Somalia who described the act of selling their goats before
fleeing. This information allowed the Innovation Service to identify the market prices for
goats as a predictor for potential displacement behavior. Additionally, although Project
Jetson represented the first time predictive analytics had been leveraged by UNHCR
to understand displacement, it was never operationalized and used by field or country
operations to inform their decision making prior to arrivals. It just served as proof of
concept that an AI-based system could be used to anticipate displacement. Linkages
with emergency-related policies, ethics, human rights-based approach to AI and other
issues such as closure of borders due to predictions, need to be taken into account prior
triggering decision-making based on AI products.
• Contact Information: Hovig Etyemezian (etyemezi@unhcr.org), Rebeca Moreno Jimenez
morenoji@unhcr.org
Project 6: Detecting drought with computer vision (AI) and satellite imagery
• Project Description: This project came out of engagement with Human Rights, Big Data
and Technology Initiative, University of Essex during a previous innovation project, Project
Jetson. This project saw the utilisation of computer vision (AI sub-domain) and satellite
imagery of Somalia to detect drought in the different regions and correlate with conflict
and displacement patterns. This project represents a nascent exploration for UNHCR to
incorporate the detection of extreme weather events due to climate change as part of
regular monitoring to perhaps one day better anticipate the flow of internally displaced
people due to drought. This project utilised Landsat 8 satellite imagery to detect indicators
of drought on the terrain, which was cross-validated with weather datasets from the same
regions.
224
• Project Type/Output: Code repository, Software tool (application), Blog article

UNHCR
• Project Domain: Forced Displacement, Computer Vision
• Data Source:
o Conflict Data: https://acleddata.com/, Satellite imagery: Landsat8

o Geography: Somalia, admin level 2 (subnational, per region)
• Link to data: https://data.unhcr.org/en/geoservices/

• Data publicly available: Yes
• Technology/Platform: Docker. Open Source, Python with the following libraries: espa-
bulk-downloader, rasterio, scikit-image matplotlib, earthpy and geopandas. GitHub for
the project available here: https://github.com/unhcr/Jetson/wiki/3c.-Experiment-3
• Project Update: built a project docker for computer vision/AI researchers and other GIS-
experts to download the software package in their local computer in Project Jetson site.
The researchers need their own credentials for accessing satellite imagery data (e.g.
Landsat)
Peace, Justice, and Strong Institutions, and SDG 17 – Partnership for the Goals
o Academia: University of Essex - Human Rights, Big Data and Technology (HRBDT)
Initiative
o Project Site: http://jetson.unhcr.org

o Github repository: (link)
o Software tool: (application link) and in http://jetson.unhcr.org
• Lesson Learned: For a successful project implementation, there was a requirement for
extensive computer vision (AI) and geographical information systems (GIS) knowledge,
a need that was satisfied through the ongoing partnership with the HRBDT team from
University of Essex. The project also needs significant server storage space for the satellite
imagery tasking and processing. Additional data protection considerations need to be
taken in case the imagery is high definition (HD) and therefore capturing settlements or
other areas of concern.
• Contact Information: Hovig Etyemezian etyemezi@unhcr.org, Sofia Kyriazi kyraizis@unhcr
.org
Project 7: ARiN
• Project Description: ARiN is a software solution that uses machine learning techniques
for the screening of applications submitted to UNHCR’s human resources talent pools,
according to a set criteria. It assists the pre-screening phase, where the recruiters make
the first parsing of the thousands of talent pool applications. It is a web application
developed by UNHCR’s Innovation Service for the Affiliate Partnerships and Recruitment
Section (APRS) within DHR. The application is machine-learning based and supports them
with the screening process for external candidates coming from the UNHCR external
talent pool applications. The talent pools are the most sought-after functional profiles
within UNHCR, and they are dedicated to help respond urgently to forced displacement
crises. There are approximately 29 talent pools (Evergreen job vacancies) that receive
on average 8000, mostly text-based, applications per month. Contrary to other off the
225
shelf tools, ARiN was customized in order to comply with the internal policies and rules
for talent acquisition within UNHCR, which includes transparency of process, gender and
diversity considerations.
• Department/Division: Division of Human Resources (DHR), UNHCR Innovation Service
• Project End Year: 2022, officially hand over to UNICC
• Project Domain: Forced Displacement, Human resources allocation
• Project Updates: ARiN was developed as a proof of concept by UNHCR Innovation. The
software has been officially handed over to UNICC for maintenance with user requirements
coming from the human resources team. ARiN is currently being analysed for potential
interoperability/integration to Workday (new human resources system, Workday).
Growth, SDG 9 –Industry; Innovation and Infrastructure; SDG 16 – Peace, Justice, and
Strong Institutions; SDG – 17 Partnership for the Goals
o UNICCC
o Project Site: https://arin-hr.unhcr.org/account/login?returnUrl=%2Ftalent-pools

o Project description: https://www.unhcr.org/innovation/division-of-human-resources
-dhr-arin/
o Presentation: ARIN public presentation
o Blogs: (Blog 1, Blog 2, Blog 3, Blog 4, Blog 5)
• Contact Information: Hovig Etyemezian etyemezi@unhcr.org, Sofia Kyriazi kyriazis@unhcr

.org

SDG 3, 8, 9, 10, 13, 16, and 17
3. Relevant Links
https://www.unhcr.org
https://www.unhcr.org/innovation/data-innovation/
Contact information: Hovig Etyemezian (etyemezi@unhcr.org), Rebeca Moreno Jimenez

(morenoji@unhcr.org), and Sofia Kyriazi kyriazis@unhcr.org
226
United Nations International Computing Centre
UNICC
Project 1: An AI Approach to Flag Sexist Text Content on Social Media Channels in

Latin American Countries
Project Description: Rapidly growing access and use of information and communication
technologies (ICT), accelerated by the COVID-19 pandemic, has had multiple impacts on
gender equality and women’s rights, including the further exacerbation of existing forms of
sexism, abuse and violence against women (VAW). At the same time social media platforms,
especially X (Previously Twitter), have become powerful conduits for communication in Latin
American countries. The rise of online sexism and the perpetuation of harmful narratives
necessitate proactive measures.
To address this concern, UNICC and UN Women developed an artificial intelligence (AI) model
designed to automatically detect and flag sexist text content on X (Previously Twitter) across
Spanish-speaking countries in Latin America.
The AI model employs advanced supervised algorithms, carefully chosen and trained on diverse
labeled datasets, to identify sexist and abusive language that may contribute to perpetuating
harmful stereotypes. Recognizing the linguistic and cultural diversity across Latin American
countries, the model incorporates a multilingual approach, considering regional nuances to
ensure accurate detection.
Despite linguistic and cultural challenges, the model demonstrates promising results in
detecting and categorizing sexist content. Continuous refinement and ethical considerations
remain integral aspects of the approach we used, emphasizing the importance of flagging the
content with freedom of expression.
Ultimately, the AI model serves as a valuable tool for addressing online sexism in Latin
American countries. Its insights and trends gleaned from the analysis will play a pivotal role in
informing the development of a systematic approach to prevention and response to sexism
and technology-facilitated Violence Against Women (TF VAW). This includes evidence-based
prevention interventions aimed at transforming harmful social norms, enhancing understanding
of the issue in the region, and fostering the creation of safe online spaces for women and girls.
Moreover, the model will contribute to the development of effective enforcement mechanisms
and consistent standards for content moderation, specifically tailored for UN Women initiatives.
• Project Type/Output: Policy Framework, Software tool

227

• Reported as part of 2022 Compendium on AI Activities? No
• Project Domain: Gender, Human Rights, Justice
• Data Source:
o Publicly available Social Media Data (X- Formerly Twitter)

o Filtered data specific to Gudalajara, Mexico.
• Technology/Platform: Python, Jupyter notebook.

• Related SDGs: SDG 3 – Good Health and Well-Being, SDG 5 – Gender Equality, SDG 16
– Peace, Justice, and Strong Institutions, SDG 17 – Partnerships for the Goals
• Partners:
o UN Partners: UN Women
o Private Sector: Twitter
o Academia: UPV Valencia
o https://www.unicc.org/news/2022/07/20/unicc-academic-partnerships-help-students
-take-learning-beyond-the-classroom/
Effectiveness of OpenAI/GPT in Data Translation: The application of OpenAI/GPT in data

translation has proven highly effective, maintaining exceptional standards of quality and
performance.
Successes and Limitations of Keyword Approach: Leveraging keywords, we successfully

identified sexist tweets with a detection rate of 10-12%. However, this approach fell short in
fully distinguishing between different types of sexism.
Insights into the Relationship Between Sexism, Emotions, and Hate Speech: Our research has
provided valuable insights into the complex relationship between various forms of sexism,
emotions, and hate speech, enriching our understanding of these interconnected phenomena.
Competitiveness of Our Model: Despite the limitations, our model's performance is comparable
to other Spanish language models, highlighting its competitiveness within the field.
Areas for Improvement and Expansion: Moving forward, we have identified areas for
improvement and expansion. Initiatives such as implementing a dashboard and conducting
hyperparameter search are underway to enhance efficacy and efficiency. Additionally, we aim
to broaden the project's scope by adapting the methodology to encompass Spanish-speaking
and other LATAM countries, facilitating code reutilization and scalability.
Potential Applications: Our research findings pave the way for several potential applications.
These include the development of a real-time app to alert users about sexist content, as well as
the creation of an API for seamless integration of our models into existing systems, benefiting
stakeholders such as UN agencies and LATAM countries.
Overall Impact and Contribution: Through these endeavors, we aim not only to advance the
field of AI-driven data translation but also to contribute to societal well-being by combating
harmful online behavior and fostering a safer online environment.
228
• Contact information: Anusha Dandapani (dandapani@unicc.org), Lizzette Soria, Andrea

Chazaro (andrea.chazaro@unwomen.org)
UNICC
Project 2: An AI Approach to Incident Management
Project Description: The United Nations International Computing Centre (UNICC) is the leading
digital service provider within the UN system. UNICC’s AI Approach to Incident Management
integrates Information Technology Infrastructure Library (ITIL) processes, providing a structured
framework for efficient incident management and problem resolution. This innovative approach
aims to establish a robust decision support mechanism, refining the monitors responsible for
initiating incidents. By enhancing the accuracy and effectiveness of incident detection, our
approach contributes to more streamlined infrastructure management. Through the application
of AI technologies, we strive to optimize the monitoring process, ultimately minimizing
downtime and improving overall operational efficiency.
Background: ITIL is a renowned framework of IT best practices, aimed at aligning IT services

with both customer and business needs. Within this framework, incidents are defined as
unplanned interruptions or reductions in the quality of IT services, while problems represent
the root or potential causes of such incidents.
ITIL categorizes changes as additions, modifications, or removals of elements that could impact
IT services, including hardware, software, and application patching. Workarounds, considered
temporary fixes, aim to restore service provisionally without addressing the underlying
problem's root cause. Meanwhile, UNICC employs a catalog system to provide customers with
self-service capabilities, encompassing categories, catalog items, variables for customization,
and the generation of requests and requested items.
Incident management within UNICC focuses on promptly resolving service-related issues

to minimize negative impacts. All incidents are meticulously recorded in the Service Now
(SNOW) tracking and analysis. Our problem management approach aims to reduce the
incidence and impact of recurring issues by identifying root causes and implementing effective
solutions. Solutions may range from component-level adjustments to organizational process
improvements.
Incident resolution follows a structured process within UNICC, starting with the creation of
incident records triggered by monitoring tools or user reports. Service Desk agents then
identify the affected service and associated Configuration Item (CI), subsequently assigning the
issue to the relevant team for resolution. Technical teams troubleshoot the problem, implement
workarounds or permanent fixes, and close the incident once resolved.
In summary, UNICC's utilization of ITIL processes ensures efficient incident and problem
management, fostering continuous service improvement and enhancing the delivery of IT
services to UN system organizations.
Solution: We harnessed the power of AI to enhance decision-making processes at UNICC.

Facing the challenge of effectively managing infrastructure status, we devised a groundbreaking
solution leveraging AI technology. UNICC's Operation (OPX) team implements sophisticated
monitors to continuously assess various parameters against predefined thresholds. Upon
detecting breaches in these thresholds, incident tickets are automatically triggered, each
assigned with designated priorities. This system is dynamic, with parameters, thresholds, and
229
priorities subject to regular review and ad-hoc adjustments by the team. Our objective is to
establish a robust decision support mechanism aimed at refining the monitors responsible for
initiating incidents, thereby elevating the overall management of our infrastructure. Importantly,
input for this mechanism includes a list of incidents, ensuring data privacy through masking
techniques.
• Project Type/Output: Academic paper, Software tool

• Project Status: Ongoing (deployed)
• Project Domain: Operational improvement & Process efficiencies
• Data Source: Incident data from Internal Tools such as Service Now /SNOW and ITIL
• Technology/Platform: Python Notebook.
• Related SDGs: SDG 7 – Affordable and Clean Energy, SDG 9 – Industry, Innovation, and
Infrastructure, SDG 11 – Sustainable Cities and Communities, SDG 17 – Partnerships for
the Goals
• Partners:
o UN Partners: UNICC OPX Operations

o Academia: University of Salento
Text Cleaning and Data Anonymity: Our journey began with the critical step of text cleaning,
focusing on enhancing data anonymity and confidentiality. Named Entity Recognition (NER)
techniques were instrumental in removing proper names, while word embeddings tailored
to the IT domain effectively captured domain-specific semantics. Additionally, refinement of
stopwords and exploration of automatic techniques like NER for identifying organization names
were vital in ensuring data quality.
Prioritization Strategies: We underwent an iterative process of experimenting with prioritization

strategies. While some approaches initially seemed promising, we encountered roadblocks that
hindered the classification process. Eventually, we settled on a method that involved reducing
the number of classes from five to three and modeling the problem as binary classification. This
decision allowed us to distinguish critical from non-critical incidents effectively.
Custom Vocabulary Development: A key insight gained was the importance of curating a
custom vocabulary comprising words specific to high-priority incidents. This focused approach
facilitated the development of a more accurate and impactful model.
Optimizing Data Analysis Workflows: Through the implementation of comprehensive

processing and prioritization methods, our aim was to optimize data analysis workflows. By
deriving actionable insights and driving informed decision-making, we sought to enhance
organizational efficiency in incident management.
Culmination of Industry Standards and Iterative Experimentation: Our approach represents

a culmination of industry-standard practices and iterative experimentation. By leveraging
230
established methodologies and continuously refining our techniques through experimentation,

we have developed a robust and effective methodology for data analysis in incident
management.
UNICC
By sharing these lessons learned and insights gained, we hope to assist others in navigating
similar challenges and achieving success in their endeavors.
• Contact information: Anusha Dandapani (dandapani@unicc.org)
Project 3: Detection of Misinformation in Social Media
Project Description: The Detection of Misinformation in Social Media project was a collaborative
effort conducted between UNICC and NYU Capstone participants from 2021 to 2022. The
objective of this joint endeavor was to leverage AI technology to develop a precise and
efficient classification system for identifying "fake news" (i.e. misinformation and disinformation)
encompassing misinformation and disinformation. The overarching goal was to empower users
to assess the reliability of information encountered on social media platforms.
The term fake news encompasses both misinformation and disinformation, disseminated
primarily through dedicated fake-news outlets, which purposefully fabricate and spread false
information (Janice & The Verified Initiative of the United Nations, 2021). Misinformation
constitutes unintentional errors, encompassing inaccuracies in statistical data, images, or
comments that are mistakenly perceived as accurate. On the other hand, disinformation
involves the deliberate creation or manipulation of audio or visual content, along with the
propagation of intentionally crafted conspiracy theories or rumors, all aimed at gaining some
form of advantage. Furthermore, disinformation can serve to suppress alternative viewpoints
or divert attention elsewhere.
To address this pervasive issue, UNICC collaborated with NYU Stern (students &faculty) to
develop a comprehensive data labeling approach aimed at categorizing information based
on its veracity. Together, we established the following classification criteria:
"True": Claims that have undergone rigorous fact-checking and are confirmed to be entirely
accurate by credible sources.
"Misleading": Claims containing varying degrees of falsehood, as determined by fact-checkers,

including partial falsehoods, questionable assertions, or misleading information.
"False": Claims that have been thoroughly debunked and proven to be entirely untrue by
reliable fact-checking sources.
"Unproven": Claims lacking sufficient evidence or scientific support to determine their veracity,
categorized as unproven, unsupported, or unfounded.
This collaborative effort resulted in a structured approach to classifying content, enabling more
accurate identification and assessment of information accuracy and reliability.
This AI based systematic classification was aimed at enabling users to discern the reliability of
information they encounter, empowering them to make more informed decisions and combat
the spread of misinformation and disinformation. By fostering transparency and accountability in
the dissemination of information, this project’s approach can play a crucial role in safeguarding
231
the integrity of public discourse and promoting a more discerning and critical approach to
media consumption.
• Project Type/Output: Report, Academic Paper, Dataset, Software tool

• Project Domain: Education, Human Rights, Justice
• Data Source:
o The data labeling leveraged a manual approach of annotating for 31K out of 350K
tweets from May 2021 to Jan 2022 using keywords “COVID-19 Vaccine” and its
variation in English.
o 31K tweets are labeled based on the whitelist and the fact-check website Politifact.
com. The label categories are simplified based on the fact-check website Politifact.
com.
o Publicly available Social Media Data (X- Formerly Twitter).
o Filtered data specific to Gudalajara, Mexico.
o Whitelist - BBC News, CNN, Al Jazeera, Deutsche Welle, The Hindu, AllAfrica, the big
four global news wire services: UPI, AP, Reuters, and Agence France Presse, and the
World Health Organization.
o Fact-checking website – Politifact.com will be used for identifying the tweets outside
of the whitelist. We further simplified the Truth-O-Meter into four labels. True, False,
Misleading, and Unproven, a newly created label for those data that need to be
verified in the future.
• Technology/Platform: Python notebooks & Final presentation uploaded to UNICC

GitHub, Reusable Python classes and modules.
• Related SDGs: SDG 4 – Quality Education, SDG 5 – Gender Equality, SDG 16 – Peace,
Justice, and Strong Institutions, SDG 17 – Partnerships for the Goals
• Partners:
o Academia: NYU Stern Business Analytics
Enhancements in Dataset: Significant improvements were noted in our dataset, which now
comprises a broader array of attributes for comprehensive analysis. With over 350,000 tweets
collected for this project, each accompanied by relevant attributes, we have gained a rich
repository of data to draw insights from. Despite fields like location and country remaining
largely empty, their absence does not impact our analysis, given our focus on English-speaking
tweets.
Impact of COVID-19 Pandemic on Twitter Activity: A noteworthy observation was the sharp
increase in Twitter account subscriptions coinciding with the onset of the COVID-19 pandemic.
This phenomenon suggests a mix of established institutions creating accounts during Twitter's
early years and opportunistic users capitalizing on the pandemic.
232
Effective Use of CRISP Methodology: The CRISP methodology for data mining proved to be
effective in guiding our analysis efforts. This structured approach enabled us to efficiently
navigate the complexities of the dataset, extracting meaningful patterns and insights.
UNICC
Commitment to Robust Methodologies: As we continue to explore the data and its implications,
our project remains committed to leveraging robust methodologies and innovative approaches
to uncover actionable insights. We are continuously seeking new methodologies and techniques
to enhance our analysis and drive informed decision-making.
Future Plans for Expansion: Moving forward, we are exploring the possibility of trying other AI
methodologies and techniques to further enhance our analysis. Additionally, we are actively
creating new annotated datasets related to social media to expand our research further and
unlock additional insights in this domain. This commitment to continuous improvement and
expansion reflects our dedication to delivering tangible benefits through our project.
• Contact information: Anusha Dandapani (dandapani@unicc.org), Lizzette Soria, Andrea

Chazaro (andrea.chazaro@unwomen.org)

SDGs 3, 4, 5, 7, 9, 11, 16, 17
3. Relevant Links
https://www.unicc.org/news/2022/07/20/unicc-academic-partnerships-help-students-take
-learning-beyond-the-classroom/
Contact information: Anusha Dandapani (dandapani@unicc.org)
233
United Nations Children’s Fund
Project 1: End-Year Summary Narrative Quality Assurance Tool

• Project Description: End-year reports are important instruments for accountability and
monitoring of UNICEF's ongoing performance for all offices. The End-year Summary
Narrative (EYSN) component is posted on UNICEF’s external website, making it accessible
to the general public – including through UNICEF’s Transparency Portal (https://open
.unicef.org/). It needs to be written in a style and tone that is appropriate, credible,
relevant, and understandable for external audiences. It is important to ensure the quality
of the report, as well as guarding against the inclusion of statements that may give rise to
reputational risk to UNICEF and partners in the public domain.
• Collaborating with UNESCWA, UNICEF DAPM developed an AI-driven quality assurance
tool that can quickly scan the draft EYSN reports and assess the degree of adherence to
reporting guidelines, flag any potential reputation risk language, and perform certain
editorial check functions. This tool was offered to COs during 2021 end year reporting
period and helped achieve efficiency gains and improve the overall quality of final reports.
• Department/Division: Division of Data, Analytics, Planning and Monitoring (DAPM)
• Domain: Improved quality of end-year summary narrative reporting that’s made available
to the public, as well as guarding against the inclusion of statements that may give rise to
reputational risk to UNICEF and partners in the public domain.
• Data Source: Executive analyses of key aspects of annual results and management
performance for both external and internal audiences.
• Link to data: Past reports - Country, Regional and Divisional Annual Reports 2020 | UNICEF
• Technology/Platform: Net core with ML running on Python. Frontend build with Angular
Infrastructure; SDG 17 – Partnerships for the Goals
• Partners: UN Partners: UN ESCWA
• Contact information: Sanjana Gaddam, Timothy Takona (sgaddam@unicef.org, ttakona@
unicef.org)
Project 2: AI4D Research Bank

• Project Description: Data gaps hinder the development sector from implementing
evidence-based programs for children and other vulnerable populations. Access
234
to updated and high-resolution data would help unlock timely insights for decision
making such as vulnerability, economic activity, and environmental conditions. The
AI4D Research Bank will leverage various geospatial analytic solutions to help fill in and
UNICEF
augment data gaps, scaling poverty estimation efforts and air quality exploratory research,
while promoting open science by giving other data scientists access to datasets and
machine learning model training references. Access to updated and high-resolution data
would help unlock timely insights for decision making such as vulnerability, economic
activity, and environmental conditions. The resulting web app is designed for both data
scientists and program staff, and will offer a Data Catalogue, a Model Catalogue, and API
Functionality to ensure replicability. Funding for this initiative is made possible through
the UNICEF Venture Fund, Office of Innovation.
• Department/Division (if applicable): Frontier Data Hub, Regional Office for East Asia and
the Pacific islands (EAPRO), Office of Innovation
• Domain: Environment and Poverty
• Data Source: Our poverty estimation model features a low-cost and efficient
implementation that will use publicly available data sources:
o Google Earth Engine satellite images

o Facebook High Resolution Settlement Layer
o OpenStreetMap
o Ookla Internet Dataset
• The exploratory air quality research side of the platform will use data from:
o Satellite Imagery MODIS, VIIRS (NASA Earth Data)

o Mapillary Street-level Sequences Dataset image dataset for Bangkok
o Meteorological data from the World Air Quality Index

• Technology/Platform: The web app itself will be open sourced and data analysis made
available via an API. The repository will be maintained on GitHub.
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 3 – Good
Health and Well-Being; SDG 13 – Climate Action; SDG 16 – Peace, Justice, and Strong
Institutions
• Partnerships:
o UN Partners: UNICEF
o Private Sector: Thinking Machines
• Links: Partner’s website: https://thinkingmachin.es/

• Lessons Learned: The bulk of the development work for the AI4D Research Bank will
happen in 2022. During 2021, UNICEF and the technology partner co-designed the
project through a constant communication of respective priorities and interests. This
alignment is paramount on such projects, where the capacity and expertise of partners
is crucial to guarantee the rigor and speed of execution. We will create an impact
assessment report to evaluate the use of the AI4D Research Bank, poverty estimation
models, and air quality estimation research. We will gather information on how these
models and tools have helped other users’ work and inputs to future iterations.
• Contact information: Karim Ghalaini (kghaliani@unicef.org)
235
Project 3: Safer Chatbots

• Project Description: Chatbots, including those run by UNICEF, often do not detect,
or respond to users’ attempts to seek urgent help with potentially life threatening
experiences. The Safer Chatbots project aims to standardise safeguarding mechanisms
within any chatbot reaching vulnerable communities, especially those engaging girls
and women. Safer Chatbots includes 3 options for chatbot implementers who wish to
improve the ability of their chatbot to detect and respond automatically to safeguarding
disclosures, for example, messages from users indicating they have experienced or are at
risk of serious harm including Gender Based Violence or suicidal ideation. Implementers
can take inspiration from tried and tested mechanisms which have been accessibly
documented via Implementation Guidelines and piloted by 3 UNICEF country offices.
Chatbot developers do not have to be running an AI powered chatbot, but one of the
options is powered by an NLP model developed and tested by external partners with
our support. Mechanisms and guidelines can be implemented on any chatbot-building
software but the guidelines have been tailored for those using platforms such as RapidPro,
Turn.io and Bothub. By implementing a Safer Chatbot mechanism, chatbot developers
can improve the safety of their chatbots’ users and ensure they get the support they need
from appropriate referral services.
• Department/Division: Gender (EAPRO)
• Project Type/Output: The main output of this project is implementation guidelines
including RapidPro flows. Through a partnership with external organisations we have
also produced a first draft of a safeguarding NLP model.
• Domain: Gender, Health
• Data Source: Data used for the development of the NLP model option was collated
anonymously based on training data from external partners. No other data is involved in
this project until the point of implementation, at which point the data collected depends
on that collected by the implementer, but may include anonymised safeguarding
disclosures, and demographic data including age and gender.
• Technology/Platform: Templates have been created and made available using RapidPro
(.json files available as part of project documentation). One of the options is available via
the Turn.io platform as a template, and the other via Weni/Bothub as well as via Github.
• Related Sustainable Development Goals (SDGs): SDG 3 – Good Health and Well-Being;
SDG 5 – Gender Equality; SDG 16 – Peace, Justice, and Strong Institutions
• Partnerships:
o Private Sector: Weni, technical partner

o NGO: Girl Effect, technical partner
• Lessons Learned: One of the biggest challenges was understanding what other initiatives
may already be in motion to address this important problem within UNICEF, and ensuring
we were not doubling up our efforts. Secondly, although local teams running or planning
chatbots recognised the importance of this issue, it was hard for them to prioritise
addressing it amongst the many other pressing issues involved in running/developing a
chatbot. For us this emphasised the importance of making this mechanism available as
part of the ‘standard’ templates for teams wishing to roll out a new chatbot, and a pre-
condition for its release to vulnerable users.
• Contact information: Gerda Binder (Gbinder@unicef.org)
236
Project 4: Spotlight
• Project Description: Spotlight is a project to predict country-level changes in the ACLED
UNICEF
data set (events and fatalities) using the news-report event data set GDELT. It will be used
to inform the UNICEF Horizon Scan, a monthly process that seeks to identify a short list of
countries facing imminent increases in humanitarian need, to support enhanced country
office preparedness. It aims to solve the problem of an absence of statistically proven
quantitative forecasting inputs related to changes in conflict intensity.
• Department/Division (if applicable): EMOPS
• Domain: Risk analysis and emergency preparedness, Risk mitigation
• Data Source: News and media events (GDELT), conflict events (ACLED)
• Link to data:
o https://www.gdeltproject.org/
o https://acleddata.com/data-export-tool/

• Technology/Platform: Microsoft Azure Databricks (Pyspark) Notebooks, Microsoft Azure
ML (Pyspark/TensorFlow) Jupyter Notebooks, Microsoft Azure Data Factory, Microsoft
SQL Server, Power BI
• Related Sustainable Development Goals (SDGs): SDG 16 – Peace, Justice, and Strong
Institutions
• Links: Tool is only available on UNICEF internal network
• Lessons Learned: The computing costs associated with calculations on a large dataset are
a constraint on exploration. Internal bureaucracy is an impediment to experimentation. A
lack of internal expertise is a limit on development.
• Kevin Wyjad (kwyjad@unicef.org)
Project 5: UNICEF Venture Fund

• Project Description: Launched in 2016, the UNICEF Venture Fund makes $50k - $100k
early-stage investments in technologies for children developed by UNICEF country offices
or companies in UNICEF program countries. By providing flexible funding, UNICEF has
the ability to quickly assess, fund and grow open-source technology solutions that show
potential to positively impact the lives of vulnerable children.
The Venture Fund has made specific calls for solutions that are in the data science and
AI space. Specific focus is placed on using satellite imagery or social media to map
infrastructure, using natural language processing, generating insights from big data,
and curating data from various sources.
In 2021, the fund launched a call for AI & ML solutions to improve access to and delivery
of digital services and systems, which resulted in investments in nine startups developing
AI solutions in the health and education space. In 2022, the fund also announced 2 new
investments into Child Online Safety solutions, both utilizing AI to strengthen children’s
skills to stay safe in the digital world. In 2023 the Fund announced investments in several
AI-powered startups as part of its call for climate change solutions.
The Fund has made investments of over $5.2MM to date in AI-related projects in
countries including Argentina, Brazil, Chile, Guatemala, India, Kenya, Mexico, Philippines,
South Africa, and Tanzania. For example, UNICEF invested in Afinidata, which offers a
237
personalized, virtual assistant that guides parents with early childhood development tools
for children under the age of six. The company has delivered over 2 million educational
sessions. Using machine learning, Afinidata has built a content personalization algorithm
which increases the use of suggested content, benefiting over 80k families in Latin
America.
Startup Investments
Portfolio Country Programmatic/ SDGs One-Line Summary

Company Name Operational Area
Afinidata Guatemala early child develop- 4 Artificial intelligence
ment, education powered personalized
educational develop-
ment centered (ECD)
advice for parents
AfriLearn Nigeria education 4 A personalized learn-

ing platform for West
African curricula that
engages students with
the help of gamification,
recommender systems,
and generative AI
Angaza Elimu Kenya education 4 AI powered eLearning

platform that deliv-
ers personalized and
mastery based learning
experiences to students
and amplifies teacher
student engagement
AQAI (Formative India environment, health 3, 11, 13 Multi-modal machine

Resilience) learning models to
assess exposure to
climate-related pollut-
ants, to identify at-risk
demographic groups
including children
Avyantra India newborn health, 3 Using machine learning

child survival for early diagnosis of
neonatal sepsis
Bioverse Labs Brazil environment, 11, 13, 15 Machine learning on

climate remote sensing imag-
ery to identify key tree
species that can be
sustainably managed
and harvested by indig-
enous communities
238
(continued)
UNICEF
Bookbot Indonesia education 4 Improving pronuncia-
tion, comprehension
and reading among
primary school children
with the help of real-
time, on-device speech
recognition technology
in Bahasa Indonesian
Cboard (Cireha) Argentina, support for children 4 Helping those with

ECARO with disabilities, speech and language
education, social impairments to commu-
inclusion nicate by symbols and
text-to-speech
Project AEDES Philippines health 3 Predict dengue cases

(Cirrolytix) from climate and digi-
tal data and pinpoint
possible hotspots from
satellite data
OTTAA Chile education, support 4 AI enabled commu-

(Comunicacion for children with nication assistant for
Aumentativa) disabilities, social children with speech
inclusion impairments using AI
and pictograms
Daktilograf Serbia education 4 A speech-to-text and

(Om3ga) text-to-speech solution
that makes it possible
for children with disabili-
ties to communicate
Datawheel Chile Chile social and economic 1, 9, 10, Data visualization

policy evaluation 11 engine for integrating
data from multiple
sources
Dymaxion Labs Argentina emergency 11 Using satellite imagery

response, social and and machine learning
economic policy to deliver geospatial
evaluation analytics insights
Equinoct India environment, 11, 13 AI and machine

climate, emergency learning with commu-
response nity-sourced data
to provide real-time
monitoring and early
warnings of climate
change hazards
239
(continued)
Eyebou UAE health 3 A smart phone applica-
tion that can detect eye
abnormalities such as
strabismus
Geospoc India connectivity, educa- 3, 4, 8, 9, Platform using artifi-

GeoSpatial tion 11 cial intelligence and
machine learning with
satellite data to map
schools and hospitals in
India
Giraffe South Africa youth engagement, 4, 8, 10 Job matching portal

education, social connecting employers
inclusion, upskilling and jobseekers, while
young people’s also generating insights
digital skills for the around critical skills in
future demand.
Weni (Ilhasoft) Brazil education, health, 3, 4 Open platform for train-

emergency ing and sharing Natural
response, child Language Processing
protection datasets in multiple
languages
Elsa (Inspired Tanzania health 3 Health Assistant tool,

Ideas) powered by artificial
intelligence
Intelligent Albania environment, 6, 11, 13, Integrating machine

Network climate, water and 15 learning with a
Solutions sanitation geographic informa-
tion system platform
to detect illegal waste
landfills while monitor-
ing changes in legal
landfills
ISTEM India education, support 3, 4, 10 Dashboard for students

for children with with disabilities
disabilities, social providing automated
inclusion workflows, accessible
content conversion, and
mentorship matching
services
Jellow (Ninaad India education, support 3, 4, 10 An open-source app for

Digital for children with children with speech
Technology) disabilities, social impairments to commu-
inclusion nicate using symbols
240
(continued)
UNICEF
Jobzi Brazil education, capacity 4, 8, 10 Forecasting the
building, upskilling influence of school
young people’s connectivity on employ-
digital skills for the ment and educational
future outcomes
Kimetrica Kenya nutrition, health, 2, 3 AI-based tech to detect

child survival severe malnutrition from
an image
Map Action Mali environment, 11, 13 A dynamic platform that

climate empowers citizens and
organizations to report
on environmental inci-
dents
Map&Rank Cameroon environment, 11, 13 Machine learning

climate platform to deliver
climate risk information
and facilitate disaster
preparedness in vulner-
able communities
Neural Labs Kenya health 3 Deep learning and

Africa computer vision to
screen medical images
in real time to identify
diseases such as pneu-
monia and tuberculosis
Towi (Pixframe Mexico education 4 Game-based learning

Studios) tools customized using
data science
Portal Brazil health 3 A telehealth and

Telemedicina population health
management platform
for streamlining informa-
tion from fragmented
and disparate data
sources for faster, more
reliable, and lower-cost
diagnostics
qAIRA Peru environment, 3, 11, 13 AI to process sensor

climate, health data and create air
pollution maps that visu-
alize air quality
241
(continued)
SolarRec Uruguay environment, 11, 13 An AI-powered platform
(Renovus Clean climate connecting solar gener-
Energy) ators to create carbon
offsets and finance
additional solar panel
installations
Similie Timor Leste environment, 11, 13 An Early Warning

climate, emergency System using machine
response learning and low-cost
sensors to assess risk
levels and mitigate the
effects of natural disas-
ters
Simple Maps Chile environment, 11, 13 Detects, classifies, and

(Rentadrone) climate organizes the errors
and damaged modules
in solar plants and
assesses crop disease
using thermal imagery
Somleng Cambodia emergency 11, 13 An interactive voice

(Chatterbox) response response platform that
was used to create an
early warning system
that sends cell phone
warning messages
to people at risk of
seasonal flooding
Talk2U Brazil education, child 4, 5 An interactive learning

protection experience to raise
awareness about the
impact of hate speech
on young people with
the goal of strength-
ening emotional
intelligence and mental
wellbeing in youth
Thinking Philippines emergency 10, 11 Mapping hard-to-reach

Machines response, social areas and connect-
protection ing communities to
resources using arti-
ficial intelligence and
geospatial analysis
Tilli Sri Lanka education, child 4, 5 An AI-powered play-

protection based social-emotional
learning (SEL) tool for
children
242
(continued)
UNICEF
Vrapeutic Egypt education, social 4 Virtual reality content
inclusion to promote skill-based
learning in children with
ADHD and autism, with
the help of AI to enrich
customizable educa-
tional scenarios
WonderTree Pakistan education, social 4 Augmented Reality to

inclusion create games for chil-
dren with disabilities
that are based on phys-
iotherapy and cognitive
exercises
Cryptolab Argentina social protection 1, 10 A financial services

(Xcapit) platform powered by
blockchain and AI that
enables financial plan-
ning, gamified savings
and wealth manage-
ment
Yuddee 2 (Beijing China education, support 4 Personalized learn-

Daokoudai) for children with ing for children with
disabilities, social autism to help develop
inclusion complex sentence struc-
tures
Country Office Investments
Country Project Description

Colombia Big Data for AI for emergency preparedness in Colombia
Emergencies and
Resilience
Iraq Monitoring child Poverty estimation using big data and AI

poverty through
new methods and
data
Brazil Big Data for Big data to predict and prevent tropical disease trans-
Epidemics mission in Brazil
India Safetipin Using Machine Learning to analyze night-time images

taken by taxi cabs to provide insights on safety of public
spaces to women and girls
Mexico MatematIA AI-powered game based learning tool to enhance

(Pixframe Studios) mathematics skills in upper secondary school children.
Developed by Innovation Fund alumni Pixframe.
243
(continued)
Country Project Description
India UniLearn Learning management system that will use AI / ML to
provide personalized learning for children
EAPRO, Thinking Machines Data Science & ML to use new and different data
Indonesia sources for better poverty estimations and insights
on pollution. Implemented jointly with fund alumni
Thinking Machines.
Mongolia, Oky A period tracking app that promotes privacy and

Indonesia provides positive and accurate information about repro-
ductive health
Greece, FunDoo A digital life skills development tool for adolescents and
Eastern young adults that is integrated into U-Report
Caribbean
• Department/Division: Office of Innovation

• Project Type/Output: Software tool, Investments into early stage open source frontier
technologies (including AI, ML, data science) developed by startups in UNICEF’s
programme countries or UNICEF’s Country Offices.
• Domain: Environment, Education, Gender, Health, Poverty, and Risk mitigation
• Related Sustainable Development Goals (SDGs) SDG 1 – No Poverty; SDG 3 – Good
Health and Well-Being; SDG 4 – Quality Education; SDG 5 – Gender Equality; SDG 8 –
Decent Work and Economic Growth, SDG 9 – Industry, Innovation, and Infrastructure,
SDG 10 – Reduced Inequalities, SDG 13 – Climate Action; SDG 15 - Life on Land; SDG
16 – Peace, Justice, and Strong Institutions
• Partnerships:
o Government: Country governments that are implementing partners to the Country

Office solutions
o Private Sector: Startup companies
• Links:
o https://www.unicefinnovationfund.org
o https://www.unicef.org/innovation/venturefund
o https://www.facebook.com/unicef.innovation
o https://twitter.com/UNICEFinnovate
o https://www.instagram.com/unicefinnovate/
o https://www.linkedin.com/showcase/unicef-innovation/
o https://www.unicef.org/innovation/research-and-reports/download-impact-brief
-2023
• Lessons Learned: As AI becomes increasingly pervasive across all programmatic areas

in which UNICEF works, the demand and the opportunities for investing in AI solutions
for children keep growing. It is important to discern where AI can deliver the greatest
impact in advancing UNICEF’s priorities, providing added value to the programmes, and
realizing the transformative potential of AI to scale solutions for children.
244
• Contact information: Sunita Grote (sgrote@unicef.org)
Project 6: UNICEF Policy Guidance on AI for Children
UNICEF
• Project Description: The Office of Global Insight and Policy led a two-year project to
better understand how Artificial Intelligence (AI) systems can protect, provide for, and
empower children. Key to this project was the development of a guide for creating
and implementing AI policies and systems that protect children’s rights and brings the
attention of the public and private sectors to how AI systems impact on children. To
develop the guidance over 200 experts were consulted in 5 regions, and almost 250
children were consulted on AI issues.
• Department/Division: Office of Global Insight and Policy
• Project Type/Output: Report, Policy Framework, and Conference (multiple meetings on
a theme)
• Project updates: UNICEF released a report of the consultation with 245 children as well as
the consultation workshop methodology. The draft policy guidance was put out for public
consultation to solicit feedback. UNICEF also worked with eight governments, companies,
and non-profit organizations to “pilot” the guidance. The results were released as case
studies and a version 2 of the Policy Guidance on AI for Children was published during
the first-ever Global Forum on AI for Children. The guidance has been translated into
French, Spanish and Arabic.
• Domain: Education, Human Rights, Child rights, and Digital technologies
Growth
• Partnerships:
o Government: Government of Finland

o Civil Society: IEEE Standards Association, World Economic Forum, the 5Rights
Foundation
o Academia: Berkman Klein Centre for Internet & Society, Harvard University
• Links: https://www.unicef.org/globalinsight/featured-projects/ai-children #ai4children

• Lessons Learned: There is a real need for greater guidance on how to develop AI policies
and systems that uphold child rights, that include children in the design process. Many
of the valuable lessons learned by the organizations piloting the guidance are captured
in the case studies. The Government of Scotland formally adopted the policy guidance
in its National AI Strategy in 2021, a key success for the project.
• Contact information: Steven Vosloo (svosloo@unicef.org)
Project 7: UNICEF’s Good Governance of Children’s Data

• Project Description: Since data is a key input into AI systems, the governance of children’s
data to ensure child-centred AI is critical. UNICEF developed a Manifesto that sets
aspirational benchmarks to guide governments, the private sector and international
organizations in developing data governance that take full account of children’s issues
and rights. The Manifesto proposes the world we want and aims to address ambiguous
or sensitive areas where there are no straightforward answers. To develop this Manifesto,
a working group of 17 global experts from the private sector, academia, think tanks
and others provided analysis, insights, guidance and comments. They wrote short
245
commentaries examining data governance approaches, evidence, gaps and grey or

conflicting areas. A wider group of experts was engaged through convenings, webinars
and consultations throughout the year.
• Department/Division: Office of Global Insight and Policy
• Project updates: A Manifesto that articulates a vision for a better approach to children’s
data has been published.
• Domain: Human Rights, Child rights, Digital technologies, Data
• Links: https://www.unicef.org/globalinsight/good-governance-childrens-data
• Related Sustainable Development Goals (SDGs): SDGs 1, 3, 4, 5, 8, 9, 10, 13, 16, 17
• Contact information: Jasmina Byrne (jbyrne@unicef.org)
Project 8: More Water More Life - Groundwater Mapping Project Using Satellite Data
• Project Description: UNICEF’s More Water More Life initiative was designed to increase
the drilling success rate for groundwater by mapping deep aquifers—bodies of permeable
rock that can contain groundwater—with satellites and science. Since 2017, UNICEF has
collaborated with the European Union Joint Research Centre (EU-JRC) to develop this
new way of mapping groundwater using geospatial imagery and data. Together, EU-
JRC and UNICEF found a way to use satellite images in conjunction with conventional
exploration techniques to reduce the time and resources required for groundwater
feasibility studies. More Water More Life’s data-driven approach helps find groundwater
before drilling, doubling the drilling success rate. It has improved water access for 1.2
million people (including 740,000 children) in Ethiopia and Madagascar.
• Project Type/Output: Map
• Project status: Ongoing
• Project end year: ongoing
• Project domain: Water and Sanitation
• Technology/platform: Remote sensing, geospatial AI
• Related SDGs: SDG 6 – Clean water and sanitation, SDG 11 – Sustainable cities and
communities
• Partnerships: European Union Joint Research Centre (EU-JRC)
• Relevant links: https://www.unicef.org/innovation/stories/finding-water-driest-places
• Lessons learned: Geospatial technology allowed UNICEF to access far-flung areas that
would otherwise be impossible to map using conventional survey methods. Remote
sensing data allowed UNICEF to cover wide areas, including those traditionally known
to be too remote and too fragile due to conflict. In spite of difficult terrain and conflict,
More Water More Life proved effective. This provides a strong case for expanding More
Water More Life in East and Southern Africa and beyond.
246
Project 9: Climate-Based Ensemble Machine Learning Model to Forecast Dengue

Epidemics in Peru and Brazil
UNICEF
• Project Description: Researchers from UNICEF’s Office of Innovation worked alongside
colleagues from the Latin American and Caribbean Regional Office (LACRO) – plus data
scientists from the European Space Agency and researchers across a wide range of
universities and institutions – to develop a climate-based ensemble machine learning
model to forecast dengue epidemics more accurately. Combining 20 years of data about
dengue cases from Brazil and then Peru, satellite-based meteorological products and
socioeconomic data, they trained the ensemble machine learning model to learn from
the past in order to forecast the near future. The model then autonomously correlated
dengue cases with meteorological conditions like temperature, rainfall and altitude, as
well as the socioeconomic conditions of the area under analysis, to enable it to forecast
caseloads under the different conditions. The forecasts proved to be notably more
accurate than estimates using previous approaches, including those that had been based
on single machine learning architectures. This indicates that the new ensemble-based
approach could provide local health authorities with the timely information they need to
prepare for a likely outbreak and ensure intervention to contain the spread of disease and
save lives. First piloted in Brazil, the model appeared to be as accurate when tested in a
second country – Peru – giving the team confidence that it could be reproduced in other
countries.
• Project Type/Output: Epidemiological model
• Project end year: Ongoing
• Project domain: Health, infectious disease, climate impacts
• Technology/platform: Geospatial modeling, remote sensing, ensemble machine learning
• Related SDGs: SDG 3 – Good Health and Well-Being, SDG 13 – Climate action
• Partnerships: European Space Agency, University of California San Diego, Newlight
Technologies
• Relevant links:
o https://www.unicef-irc.org/publications/pdf/Best-of-unicef-research-2022.pdf
o https://www.unicefusa.org/stories/using-artificial-intelligence-fight-dengue
o https://philab.esa.int/%CF%86-lab-and-unicef-joint-dengue-fever-research-receives
-further-award/
o https://www.climatechange.ai/papers/icml2021/10
• Lessons learned: The single most important finding was that the ensemble machine
learning model provided considerably more accurate predictions of dengue outbreaks
than any previous approach. The study also looked at the factors that appeared to have
the greatest influence on dengue incidence. While it was no surprise that the case rate for
the previous month came out on top, followed by temperature, the next most influential
factors were wind speed and direction. The analysis was conducted first in Brazil, which
had a longer time-series of dengue cases freely available compared with Peru. The
ensemble machine learning model was first trained and validated using the Brazilian
dataset, then transferred to the Peruvian context. This proved to be a highly effective
solution for countries that lack sufficient dengue data to train machine learning models.
• Contact information: Do-Hyung Kim (dokim@unicef.org), Hanoch Barlevi (hbarlevi@unicef
.org)
247
Project 10: Anomaly Detection in Yemen Cash Transfers

• Project Description: In Yemen, the OOI is piloting AI-based anomaly detection for
UNICEF’s cash transfer program. The project hopes to build on existing anomaly detection
algorithms and move beyond detecting “predictable anomalies” with the use of AI. The
goal is to help find suspicious transactions and avoid the loss of much-needed funds,
tackling fraud in a cash transfer program that has already reached 1.4 million beneficiaries
and impacted 9 million people, almost a third of the country’s population.
• Project Type/Output: Anomaly detection algorithm for risk mitigation
• Project domain: Social protection
• Data source: Yemen Service Center Management Information System
• Data is publicly available: No
• Technology/platform: Anomaly detection capability integrated into an existing
Management Information System
• Related SDGs: SDG 1 – No poverty, SDG 2 – Zero hunger
Project 11: Machine Translation for Emergencies

• Project Description: The Office of Innovation and EMOPS are prototyping AI-based
machine translation tools in several countries, including the Democratic Republic of the
Congo. The solution makes humanitarian knowledge and training materials available
to local emergency responders and volunteers in their own language (including low-
frequency languages), improving first-responders’ capacity to achieve full operational
readiness for emergencies.
• Department/Division: Office of Innovation, EMOPS
• Project Type/Output: Translation tool and translated texts
• Project domain: Emergency response
• Technology/platform: open-source machine translation models, large language models
• Related SDGs: 13 – Climate action, 16 – Peace, justice and strong institutions
Project 12: Integrating Speech Technology into the U Youth Application in Burkina
Faso
• Project Description: UNESCO data from 2021 show a literacy rate of 65% among
young people in Burkina Faso, placing it among the 10 countries with the worst literacy
rates. These young people who are challenged with low literacy, many of whom only
speak local languages, are in urgent need of equal access to information. Similarly, this
population is currently cut out of key feedback and accountability mechanisms that allow
them to participate in and report on programmes implemented by UNICEF and other
actors. However, many of these youth have access to mobile phones and potentially
smartphones. Therefore, Burkina Faso is upgrading the UYouth application to integrate
248
speech-to-text and other AI-powered language tools, including in low-resource languages

such as Mooré. The solution will make the platform more user-friendly, interactive, and
inclusive, increasing engagement and expanding its capacity to amplify the voices of
UNICEF
young people.
• Project Type/Output: AI-based speech-to-text functionality in a mobile application
• Project domain: Youth engagement and empowerment, Social inclusion
• Technology/platform: Speech-to-text, text-to-speech, machine translation, and NLP
• Related SDGs: 8 – Decent work and economic growth, 10 – Reduced inequalities
Project 13: Accessible Textbooks with the help of Generative AI

• Project Description: In most countries, learning materials rely primarily on the printed
word, which is not accessible to everyone. Millions of children have little or no access to
books in a language they can understand, limiting their basic human right to literacy and
education. Children with disabilities remain one of the most marginalized and excluded
groups and, for them, gaining access to accessible learning materials is a key to quality
education. However, to date, producing these materials has proven costly. UNICEF
has worked with governments on producing Accessible Digital Textbooks (ADTs) that
enhance learning for all children. However, it currently takes 6 – 9 months to produce a
single accessible textbook. Generative AI can significantly reduce the cost of production of
accessible digital formats. Following Universal Design for Learning principles, accessible
textbooks can be made available and affordable in days, not months, for students of all
abilities. This project aims to explore the potential of generative AI to create interactive
digital textbooks for a wide audience of learners (with an initial focus on Latin America and
Eastern and Southern Africa, where ADTs are already in use), and to generate evidence
on the use of AI for inclusive education.
• Department/Division: Office of Innovation, ICTD, Innocenti
• Project Type/Output: Textbooks
• Project domain: Education
• Technology/platform: Generative AI
• Related SDGs: 3 – Good health and well-being, 4 – Quality education, 10 – Reduced
inequalities
• Relevant links: https://www.accessibletextbooksforall.org/

SDG 1 – No Poverty; SDG 3 – Good Health and Well-Being; SDG 4 – Quality Education; SDG
5 – Gender Equality; SDG 8 – Decent Work and Economic Growth, SDG 9 – Industry, Innovation,
and Infrastructure, SDG 10 – Reduced Inequalities, SDG 13 – Climate Action; SDG 15 - Life on
Land; SDG 16 – Peace, Justice, and Strong Institutions
249
3. Relevant Links
• https://www.unicefinnovationfund.org
• https://www.unicef.org/innovation/venturefund
250
United Nations Interregional Crime and

Justice Research Institute
UNICRI
Project 1: Operation of the Centre for Artificial Intelligence and Robotics

• Project Description: The Centre for Artificial Intelligence and Robotics is a specialized
centre of the United Nations Interregional Crime and Justice Research Institute (UNICRI).
It was established in 2017 and is located in The Hague, the Netherlands.
The focus of the Centre is to advance understanding of the challenges and opportunities
brought by artificial intelligence (AI) and other new and emerging technologies from the
perspective of justice, crime and other security threats. It seeks to support Member States
both through practically-oriented policy guidance, research and reporting, as well as
capacity building activities to leverage the potential of these technologies in a responsible
manner.
• Department/Division: Centre for Artificial Intelligence and Robotics
• Project Type/Output: Through its specialized Centre for AI and Robotics in The Hague,
UNICRI advances understanding on the risks and benefits of AI, robotics and related
technologies vis-à-vis crime, terrorism and other threats to security and seeks to support
Member States to leverage the potential of these technologies in a responsible manner.
• Project Domain: Crime Prevention, Justice
o Relying on its convening power as a UN entity, UNICRI organized several events,

including information-sharing meetings, technical workshops, training courses, multi-
stakeholder policy discussions, and high-level awareness-raising and visibility events.
o UNICRI has laid the groundwork to continue advancing the Centre’s mission, exploring
new opportunities and emerging research areas related to the challenges and
opportunities brought by AI and other new and emerging technologies from the
perspective of justice, crime and other security threats.
• Related Sustainable Development Goals (SDGs): SDG 16 –Peace, Justice, and Strong
Institutions
• Relevant Links and Multimedia: https://unicri.it/topics/Toolkit-Responsible-AI-for-Law
-Enforcement-INTERPOL-UNICRI
• https://unicri.it/index.php/topics/ai_roboticsttps://unicri.it/News-AI-for-Safer-Children
-Regional-Training-Singapore
251
• Partnership(s)/Collaborator(s): Multiple stakeholders in the fields of artificial intelligence

and criminal justice, including Member States, UN Partners such as ITU, UNESCO, UNPOL,
UNODC, UNICEF and OICT; and other International Organizations such as INTERPOL and
the European Union.
• Lesson Learned:
Challenges:
1. Lack of a global governance framework on AI and a diverse and rapidly changing

regulatory landscape regarding AI and law enforcement’s use of AI.
2. Constant developments within the field of AI and digital transformation in law
enforcement and beyond.
3. Distinct levels of engagement from different stakeholders, with varied levels of interest
and availability to invest time and resources in the topic, resulting in difficulty with
ensuring a balanced representation across stakeholder groups, demographics and
regions across the globe.
4. Stakeholders in the field have various backgrounds and often different priorities,
objectives and motivations, making it challenging to balance their diverse perspectives
and foster collaborations.
Lessons learned:
1. Engagement and strong collaboration with diverse stakeholders (law enforcement,

criminal justice practitioners, technology providers, human rights experts, etc.) are
essential for successfully pursuing the Centre’s mission.
2. The project implementation requires flexibility, constant updating of knowledge and
review of practices to keep pace with the ongoing and rapid changes in the realm of
emerging technologies.
3. Translating technical concepts into language accessible to non-technical stakeholders
such as policy makers and the general public is crucial for effective communication
and knowledge-building.
Future work:
1. Furthering the Centre’s mission, further advancing the understanding of the challenges
and opportunities brought by AI and related technologies from the perspective of
justice, crime and other security threats.
2. Building an understanding of citizens’ perceptions of law enforcement’s use of AI and
how it relates to trust in law enforcement.
3. Conducting research on current practices and proposing strategies for enhancing
transparency around law enforcement's use of AI.
4. Developing recommendations for communication strategies to foster public trust in
law enforcement’s responsible use of AI.
5. Exploring new areas related to the core mission of the Centre, focusing on the
development of educational resources on AI to empower children, parents, and
caregivers, as well as on building knowledge on AI and law enforcement of the
broader criminal justice community, targeting practitioners such as prosecutors.
• Contact Information: Irakli Beridze (irakli.beridze@un.org); Odhran McCarthy (odhran.

mccarthy@un.org)
Project 2: Responsible Artificial Intelligence Innovation in Law Enforcement

• Project Description: Fostering responsible AI innovation within the law enforcement
community is one of the priority areas of the UNICRI’s Centre for Artificial Intelligence and
Robotics. In this context, UNICRI has established a partnership with INTERPOL to create
a unique global forum to discuss the advancements and impact of using AI for law
252
enforcement. The purpose of this project is to demystify the world of AI, not only
for law enforcement officers but also for policymakers, practitioners, industry
partners, academic researchers, civil society and the general public.
UNICRI
Within this project, UNICRI and INTERPOL organize annual global meetings on AI
for law enforcement, conducting in-depth interviews with several law enforcement
experts and other partners and hosting virtual meetings with a multi-disciplinary group
of experts to identify AI use cases for law enforcement and discuss issues related to
the dichotomy between the potential opportunities and challenges presented by
AI, coupled with the lack of guidance tailored to the relative novel use of AI in law
enforcement.
During this process, UNICRI and INTERPOL have identified the need for developing
operationally oriented support and guidance for law enforcement in the
design, development and deployment of AI in a trustworthy, lawful and responsible
manner and, in this regard, have undertaken the development of a “Toolkit for Responsible
Artificial Intelligence Innovation in Law Enforcement”. The Toolkit, the development
of which is funded by the European Commission DG HOME, is intended to provide
guidance and support for law enforcement agencies on the development, procurement
and deployment of AI in their work in a human rights compliant and ethical manner.
Through the Toolkit, law enforcement agencies globally will have access to the knowledge
and resources needed to enable them to tap into the positive potential of AI. At the same
time, many different stakeholders play an important role in promoting responsible AI in
law enforcement. As a result, the resources developed as part of the Toolkit will also cater
for and be approachable by other stakeholders, including industry, academia, civil society
and the general public.
• Project Type/Output: Multi-stakeholder platform bringing together law
enforcement, industry and academia from all over the world to explore the positive
potential of AI in law enforcement in a trustworthy, lawful and responsible manner, as
well as an actionable Toolkit consisting of resources that include principles, guidance
documents and actionable recommendations to support law enforcement to make more
informed decisions and prevent possible risks related to the implementation of AI in their
work.
• Project Domain: Justice
o The Core Group of Experts, which was established in November 2020, got together
for an additional Virtual Discussion Room organized. This Virtual Discussion Room
focused on the presentation of the near final version of the AI Toolkit ahead of its
release to the public at the INTERPOL-convened Police Science Congress in June
2023. The central objective was to gather final feedback and specialist perspectives
from the Core Group, particularly regarding the style, thoroughness, and content of
the resources which constitute the AI Toolkit.
o A validation session with law enforcement representatives worldwide also took place.
The goal of this session was to collect feedback from the law enforcement community
about the draft AI Toolkit ahead of its finalization. This validation was considered the
final checkpoint before public release at the INTERPOL-convened Police Science
Congress in June 2023.
o In June 2023, UNICRI and INTERPOL organized a special event on responsible AI
innovation as part of the Police Science Congress in Singapore on 8 June, gathering
representatives from the global law enforcement community. This event served to
253
formally and publicly launch the AI Toolkit, which was announced by Secretary-General
of INTERPOL, Dr Jürgen Stock as part of his keynote address
o The testing phase of the project took place from March until November 2023,
with a view to having the AI Toolkit tested in a practical setting by volunteering law
enforcement agencies. This exercise was a crucial step in the development process
of the AI Toolkit. A total of 15 law enforcement agencies participated in the testing
and provided valuable insights that helped improve its functionality, utility and
completeness. The results were used to refine the AI Toolkit for the release of a revised
version in 2024.
o The survey on public opinions on the use of AI by law enforcement, launched in
November 2022, ran until June 2023. The survey was shared with UNICRI’s network
and publicized on different online platforms including the UNICRI website, and its
accounts on Facebook, LinkedIn , Twitter and YouTube. 759 people responded to
the survey, and their contributions are currently being analysed.
o Three specialized training sessions on responsible AI innovation in law enforcement
were conducted, building on the AI Toolkit’s content:
- On 8 September 2023, UNICRI together with INTERPOL and CARICOM IMPACS,

hosted a webinar to present the AI Toolkit and introduce the concept of responsible
AI innovation in law enforcement. The webinar was attended by mid to senior-level
management professionals within the law enforcement community in the region
of the Caribbean Community.
- On 4 and 5 October 2023, UNICRI and Naif Arab University for Security Studies
(NAUSS) co-hosted the 1st Artificial Intelligence Forum for Law Enforcement for
Arab Countries. On the margins of the Forum, UNICRI and INTERPOL convened
a specialized workshop for participants focused on responsible AI Innovation. The
workshop aimed to introduce participants to the AI Toolkit, explain what responsible
AI is and why it is important and, finally, to help participants to understand how the
contents of the AI Toolkit can be applied through an interactive scenario-based
exercise.
- On 13 November 2023, UNICRI and INTERPOL together with CEPOL hosted
a webinar with the objective of enhancing awareness regarding responsible
and human-compliant engagement with AI in law enforcement agencies across
the European Union (EU). The webinar convened 247 experts from EU law
enforcement officials, including police officers, border guards, customs, and
others. The webinar's main goal was to provide insights into the principles,
organizational elements, and procedures essential for guiding agencies in the
ethical and human rights-compliant development, procurement, and use of AI
systems.
Institutions
• Relevant Links and Multimedia: https://unicri.it/topics/Toolkit-Responsible-AI-for-Law
-Enforcement-INTERPOL-UNICRI
o UN Partners: UNPOL, OICT

o Other International Organizations: INTERPOL, European Commission (European
Union)
• Lesson Learned:
Challenges:
1. Ensuring human rights compliant AI and building public trust in AI for law enforcement.
2. Lack of a global governance framework and the prevalence of policies and regulatory
frameworks centred in the global north.
254
3. Differences in rules and regulations on AI and the law enforcement use of AI.
4. Evolving development of AI and digital transformation in law enforcement.
5. Heterogenous level of technical capabilities and the level of engagement with AI by
UNICRI
law enforcement agencies around the world.
Lessons learned:
o It is essential to work with law enforcement directly to identify their needs and help
them fill their gaps in a practical and operationally oriented manner.
o The Toolkit should be a living document rather than a complete and definitive
statement on the topic, to keep pace with ongoing developments in the area.
o The Toolkit should build upon the already established legal and ethical discussions
and guidelines surrounding the use of AI, rather than redefining them.
o The Toolkit must promote trust in the law enforcement use of AI, by fostering a sense
of openness and transparency and engaging all relevant stakeholders from the outset.
Future work:
o A revised version of the AI Toolkit will be released in early 2024, including an interactive
website.
o Further pilot training sessions will be organized on the basis of the AI Toolkit.
o The 4th Global Meeting on AI for Law Enforcement will take place in early 2024,
disseminating the improved AI Toolkit for the law enforcement community.
o Additional areas of research, awareness-raising and capacity building will be explored
to ensure the continuous advancement of the concept of responsible AI innovation
in law enforcement across the globe.
• Contact Information: Irakli Beridze (irakli.beridze@un.org); Odhran McCarthy (odhran.

mccarthy@un.org)
Project 3: Building Capacities of Law Enforcement Worldwide to Leverage the Positive

Potential of Artificial Intelligence to Combat Child Sexual Abuse and Exploitation
• Project Description: In 2020, UNICRI, through its Centre for AI and Robotics, together
with the Ministry of Interior of the United Arab Emirates launched the AI for Safer
Children initiative. This project aims to explore the positive potential of AI to support law
enforcement agencies to prevent, detect and prosecute child sexual exploitation and
abuse.
The issue of child sexual exploitation and abuse has been growing exponentially over the
last decades and has experienced a further increase throughout the COVID-19 pandemic.
Developing technologies that allow for increasingly vast amounts of files to be stored and
shared, as well as AI image generation which allows more child sexual abuse material to
be created, both exacerbate the issue further. Law enforcement agencies face significant
challenges in combating this crime, including a high turnover of investigators - largely due
to the psychological burden of reviewing the abuse material, the difficulties in identifying
both perpetrators and victims in huge amounts of data, subsequent high backlogs, the
challenges posed by the global nature of the Internet and new technologies such as
encryption.
While many existing AI tools can and are already contributing to solving these challenges,
law enforcement agencies are not yet widely using them. Many are not aware of the
potential of the AI tools, or are sceptical about the technology’s benefits. Additionally,
selecting and implementing AI tools is a complex process for which law enforcement
agencies may lack the necessary knowledge and resources.
The project aims to tackle this issue through the continued development of its AI for
Safer Children Global Hub – a unique, online platform where law enforcement officers
can access a repository of information on 80 AI tools that are available to combat child
255
sexual exploitation and abuse. The Global Hub also allows law enforcement officers to
communicate with each other and learn how to implement and use AI tools responsibly
within their workflow.
To further distribute the Global Hub’s information and resources, the AI for Safer Children
initiative is also offering free online and in-person training to advance the capacities of law
enforcement agencies across the globe to combat child sexual exploitation and abuse.
These training sessions are tailored to the specific needs and contexts of participating
law enforcement agencies and cover a broad range of AI tools and techniques across an
investigative workflow.
• Project Type/Output: Online platform, trainings
• Data Source: AI tools that can be leveraged by law enforcement agencies in preventing,
detecting, and prosecuting child sexual exploitation and abuse.
• Technology/Platform: Microsoft SharePoint
Since the launch of the AI for Safer Children Initiative in 2020, UNICRI has:
o Developed and rolled out the Global Hub platform dedicated to supporting law
enforcement in leveraging AI to combat child sexual exploitation and abuse, including
a database of over 80 existing AI-based tools and guidance on how to implement
them responsibly, as well as a communication section. As of the end of 2023, the
Global Hub has over 570 registered law enforcement users from 107 countries.
o Continually updated and managed the AI for Safer Children Global Hub on SharePoint
by collecting, adding and adjusting information to display on the Global Hub (including
creating content such as learning videos for the law enforcement audience, collecting
information about AI tools from technology providers, and adding relevant news and
events as they arise), translating the Global Hub into all official UN languages and
onboarding law enforcement agents worldwide to the Global Hub.
o Designed and launched specialized AI for Safer Children trainings which adapt the
information on the Global Hub to local or regional contexts. In total, nine trainings have
already been conducted for 24 countries – including 11 Caribbean states – and 1,446
law enforcement participants, including 1,000 from Zimbabwe Republic Police alone.
These trainings take place both in person and online, and aim at teaching investigators
how they can responsibly implement a broad range of AI tools to facilitate their child
sexual abuse investigation workflows at each step of the process.
o o Established an Advisory Board composed of representatives of law enforcement
agencies, technology providers, academia, international organizations and civil society
organizations, to support the initiative and ensure its relevance to the needs and
requirements of combating child sexual exploitation and abuse. o Held many
meetings, including annual Advisory Board meetings in December 2023, February
2023, March 2022 and 2021; annual Technology Provider meetings in November
2023, September and February 2022; two Stakeholder Meetings for law enforcement
agencies in June 2022 and 2021; and an Ethics Expert Meeting on December 2021.
o Carried out extensive research and analysis on the ethical considerations associated
with the AI for Safer Children initiative, including those associated with the ethical
development and deployment of AI in law enforcement efforts to combat child sexual
exploitation and abuse. This resulted in the development of the core principles of
256
the initiative, the Terms of Engagement for Law Enforcement Users, the Terms of
Engagement for Technology Providers, and the Data Protection Policy. A Responsible
AI Questionnaire was also developed to collect information from the technology
UNICRI
providers on ethical and legal aspects related to the development of their tools.
o Presented and attended a variety of events and conferences including but not limited
to the United Nations General Assembly, INTERPOL Conference, the AI for Good
Summit and the United Nations Chiefs of Police Summit to promote the initiative and
further build the AI for Safer Children community.
• Related Sustainable Development Goals (SDGs): SDG 16 – Peace, Justice, And Strong
Institutions
o UN Partners: UNICEF, OICT, UNPOL, UNODC

o Government: Ministry of Interior of the UAE
o Private Sector: Griffeye, SafeToNet, Kaseware, Semantics21, Web-IQ, Cyacomb
Examinor, Qumodo, ZiuZ Visual Intelligence, Paliscope, CameraForensics Ltd,
FriendMTS, Trilateral Research, SUMURI, 1st1 Technologies, Netspark, GetData
Forensics, Nuix, Dark River Systems, Magnetic Forensics, Breakpoint Forensics, Wiretap,
Securium, Epieos, SAS Software, Amped Software, iCOP, Global Emancipation
Network, Palantir, OSForensics and PassMark Software, Privately SA, BlueBear.
o Civil Society: Red Papaz, ECPAT, International Justice Mission, National Center for
Missing & Exploited Children, Thorn, Childhood Foundation, Bracket Foundation,
RATI Foundation, Prerana, Canadian Center for Child Protection, wePROTECT
Alliance, Gucci Foundation, Project Vic, Child Rescue Coalition.
o Academia: University of Massachusetts Amherst.
o Other International Organizations: INTERPOL, European Commission (European
Union), Europol, Virtual Global Taskforce, Fund to End Violence Against Children.
o https://unicri.it/topics/AI-for-Safer-Children
o https://unicri.it/index.php/topic/AI-for-Safer-Children-Global-Hub
o https://unicri.it/News-AI-for-Safer-Children-Regional-Training-Singapore
o https://www.forbes.com/sites/markminevich/2023/12/26/revolutionizing-child
-protection-the-un-and-uaes-groundbreaking-ai-for-safer-children-collaboration/?sh
=5f912b871cdf
o Twitter: @AISaferChildren
o https://youtu.be/aYJJ2m2Y29g
o https://www.youtube.com/watch?v=sJShhYiTjoM&t=20s
o Challenges:
- Heterogenous level of technical capabilities, accessibility, and engagement with

AI by law enforcement agencies around the world.
- Reaching a truly global and diverse range of law enforcement agencies, both for
members of the Global Hub and to host trainings.
- Differences in rules and regulations in different countries and regions not only on
AI and the law enforcement use of AI, but also on online child sexual exploitation
and abuse.
- Lack of a global governance framework and the prevalence of policies and
regulatory frameworks centred in the global north.
257
- Ensuring ethical and human rights compliant AI and building public trust in the
use of AI by law enforcement.
o Lessons learned:
- The project should target law enforcement agencies in all Member States,
especially low- and middle-income countries and law enforcement agencies in
the global south.
- The project should seek to build a community around the responsible use of AI
by law enforcement, with a strong emphasis on a human rights-centred approach.
- The project should take a bottom-up approach, working with law enforcement
agencies directly to identify their needs and help them fill their gaps.
- The project should include appropriate measures and processes to ensure that it
is developed and implemented in accordance with ethical principles to warrant
that any possible issues are addressed early on.
- The project should seek to pursue specialized training for law enforcement to
increase capacity building and engagement of AI to combat child sexual abuse
and exploitation around the world.
- The project should seek to increase accessibility of AI tools around the world.
o Future work:
- Maintain, optimize, and grow the Global Hub, including desk-based research
and analysis on the continually developing field of AI tools and techniques for
the purposes of inclusion in the Global Hub and coordination with technology
developers; creation and curation of content for the Learning Centre; networking
with law enforcement agencies for onboarding purposes; promotion and visibility
activities of the Global Hub in general.
- Continue to provide and adapt trainings to build capacities and implement
knowledge imparted through the Global Hub by promoting such training
programmes and delivering the training courses to law enforcement agencies
from all over the world.
- Build on this connection with law enforcement users worldwide through trainings to
gather more information about their priorities and limitations, as well as feedback
about AI for Safer Children services and impact, through regular meetings,
interviews, and surveys both before and after trainings. This feedback will guide
research and development of the Global Hub and training materials in future.
- Leverage the Global Hub to foster further innovation, including organizing
hackathon challenges and sharing selected tools with the AI for Safer Children
community.
- Development of tailored resources for other non-law enforcement stakeholders
and conducting research and analysis on the evolving phenomenon of child
exploitation and abuse as well as on responsible AI.
• Contact Information: Irakli Beridze (irakli.beridze@un.org), Odhran McCarthy (odhran.

mccarthy@un.org)
Project 4: Building Knowledge on Counter-Terrorism in the Age of Artificial

Intelligence / Human Rights Aspects Of The Use Of Artificial Intelligence In Counter-
Terrorism
• Project Description: In 2020, UNICRI, through its Centre for Artificial Intelligence and
Robotics commenced a research initiative aimed at exploring the dual nature potential
of artificial intelligence (AI) from the perspective of counter-terrorism.
Together with the United Nations Counter-Terrorism Centre (UNCCT) of the United
Nations Office of Counter-Terrorism (UNOCT), UNICRI explored the dual nature potential
258
of AI, analyzing the concerning aspects of the advent of AI, such as the possibility of its
use with malicious intent by terrorist groups and individuals, as well as how AI might
be leveraged to support counter-terrorism efforts, in particular in terms of combatting
UNICRI
terrorist use of the Internet and social media. This resulted in the release of two reports
“Algorithms and Terrorism: The Malicious Use of Artificial Intelligence for Terrorist
Purposes” and “Countering terrorism online with artificial intelligence – An Overview for
Law Enforcement and Counter-Terrorism Agencies in South and South-East Asia”.
In partnership with UNOCT/UNCCT and the Office of the High Commissioner for Human
Rights (OHCHR), UNICRI further examined how artificial intelligence technology can be
used to counter terrorism in a manner that complies with human rights. The ongoing
research will result in the release of report on the human rights aspects of the use of AI in
counter-terrorism .
UNICRI, UNOCT/UNCCT and OHCHR organized a high-level briefing entitled “Counter-
Terrorism in the Age of Artificial Intelligence: Risks, Opportunities and Safeguarding
Human Rights” on 29 June 2021 as part of the Second Counter-Terrorism Week at
the United Nations. During this briefing the findings of their collective research were
presented.
o UNICRI and UNOCT organized an expert group meeting to explore the specific
application of AI-enabled social network analysis in counter-terrorism in March 2022
to develop knowledge and understanding of this application and develop a concept
for further activities in this domain.
o Department/Division: Centre for Artificial Intelligence and Robotics

• Project Status: Finished
Institutions
o UN Partners: UNOCT/UNCCT and OHCHR

o Government: Kingdom of Saudi Arabia, Government of Japan
o https://www.un.org/counterterrorism/sites/www.un.org.counterterrorism/files/
malicious-use-of-ai-uncct-unicri-report-hd.pdf
countering-terrorism-online-with-ai-uncct-unicri-report-web.pdf
• Lesson Learned:
o Challenges:
- Ensuring human rights compliant AI and building public trust in the use of AI by
law enforcement and counter-terrorism agencies, especially given the complex
interaction between human rights and counter-terrorism.
- Absence of consensus on the definitions of both terrorism and AI.
259
- Insufficient public information related to the level of technological readiness and

the current use of AI tools in the regions of South Asia and South-East Asia.
- Lack of understanding of how malicious actors could use AI.
o Lessons learned:
1. While the use of AI for terrorist purposes is currently not a developed threat, it is
important to not underestimate it.
2. The capacity of all stakeholders to identify and respond to the threat of the
malicious use and abuse of AI for terrorist purposes should be improved.
3. Efforts need to be made to raise awareness of governments and industry partners
about the role of AI in counter-terrorism.
4. It is essential to ensure that law enforcement and counter-terrorism agencies
appreciate the potential human rights impacts of AI, as well as the limitations and
fallibility of AI.
o Future work:
- Supporting the design of human rights compliant models for the use of AI for
counter-terrorism together with UNOCT/UNCCT and OHCHR.
- Building good practices for the use of AI for social network analysis for counter-
terrorism together with UNOCT/UNCCT.
- Further research and monitoring of the willingness and future ability of terrorists
to use or abuse AI.
- Further collaboration with UNOCT/UNCCT in the topic of responsible AI innovation
for counter-terrorism.
• Contact Information: Irakli Beridze (irakli.beridze@un.org), Odhran McCarthy (odhran.

mccarthy@un.org)
Project 5: Responsible Limits on Facial Recognition Technology

• Project Description: The development of facial recognition technology (FRT) presents
considerable opportunities for socially beneficial uses, mostly through enhanced
authentication and identification processes, but it also creates unique challenges. To
fully grasp these challenges and the trade-offs they may entail and to build appropriate
governance processes, it is necessary to approach FRT deployment through specific use
cases.
To this end, the World Economic Forum, in partnership with UNICRI, INTERPOL and the
Netherlands Police, has spearheaded a global and multistakeholder policy initiative to
design a robust governance framework for the responsible use of FRT for law enforcement
investigations
This workstream was started in November 2020 focused on the law enforcement use case
– identifying a person by comparing a probe image to one or multiple reference databases
to advance a police investigation. The ambition of this work is to support law- and
policy-makers across the globe to design an actionable governance framework that
addresses key policy considerations in terms of the prevention of untargeted surveillance,
the necessity of a specific purpose, the performance assessment of authorized solutions,
the procurement processes for law enforcement agencies, the training of professional
forensic examiners, and the maintenance of the chain of command for emergency
situations. This framework is structured around two critical components: a set of
principles that defines what constitutes the responsible use of FRT for law enforcement
investigations, and a self-assessment questionnaire that details the requirements that law
enforcement agencies must respect to ensure compliance with the proposed principles.
To validate this governance framework a pilot was run with police agencies in France,
Sweden, New Zealand, Netherlands and Brazil. All participating agencies provided very
260
positive feedback which served to improve the overall quality of the framework and to
ensure that what is presented is actionable, relevant and useable in an operational law
enforcement context. Besides these six, other law enforcement agencies have already
UNICRI
demonstrated interest in integrating this policy framework into their current and future
practices, including law enforcement agencies in Canada and Ireland.
o Law enforcement agencies in France, Brazil, New Zealand, Sweden and the
Netherlands participated in the pilot, which consisted of testing the 8 established
principles together with the self-assessment questionnaires and discussing the results
and suggestions for improvement during three pilot workshops.
o After the pilot, the final version of the Policy Framework was released in November
2022.
o Press releases were published by World Economic Forum, UNICRI, INTERPOL and
the Netherlands Police in several media channels.
o Promotion of the project results in relevant FRT forums such as the IDEMIA Users
Conference.
o Preliminary discussions with other law enforcement agencies that are interested in
implementing the Policy Framework at the national level.

• Project Status: Finalized
Institutions
• Partnership(s)/Collaborator(s): World Economic Forum; INTERPOL; the Netherlands
police; Brazilian Federal Police; New Zealand Police Agency; Swedish Police Authority;
Gendermerie Nationale, France; the Central Directorate of Judicial Police, France.
o https://unicri.it/A- Policy-Framework%20-for-Responsible-Limits-on-Facial-Recognition
o https://www.weforum.org/publications/a- policy-framework-for-responsible-limits-on
-facial-recognition-use-case-law-enforcement-investigations-revised-2022/
• Lesson Learned:
o Challenges:
1. The use of FRT by law enforcement is one of the most controversial uses of AI with
negative records of abusive practices, discriminatory decisions, wrongful arrests,
moratoriums and bans by several public and private entities.
2. There are several initiatives to regulate AI worldwide but most of them are general
frameworks that do not give technical and procedural specific recommendations
on how to implement FRT in a responsible way.
3. Because of the misuse of the technology, there is a lack of public trust on the use
of FRT by law enforcement.
4. Law enforcement agencies are not used to publicly share their use of technology
which further contributes to general distrust.
o Lessons learned:
261
- It is essential to work with law enforcement directly to identify their needs and help
them fill their gaps in a practical and operationally oriented manner.
- Among the pilot members, there were very different procedures on the use of
FRT which shows a lack of guidance and standardization on the use of FRT by law
enforcement.
- None of the pilot agencies implements real-time FRT which is the most controversial
use of FRT that has been banned by several countries.
- Pilot members were cognizant that transparency is the field where they could
develop the most to increase their trust by the public and improve their relation
with the communities.
o Future work:
- Global study on uses of FRT: mapping of the different uses of this technology by
law enforcement worldwide.

SDG 16
3. Relevant Links
http://www.unicri.it/
Contact Information
Irakli Beridze (irakli.beridze@un.org), Odhran Mc Carthy (odhran.mccarthy@un.org)
262
United Nations Institute for Disarmament Research
UNIDIR
Project: Security and Technology – AI and Autonomy Workstream

• Project Description: The UNIDIR Security and Technology Programme’s (SecTec) AI and
Autonomy workstream conducts original research and convenes international events to
promote a fact-based, technologically sound dialogue between policymakers, the tech
community, the private sector and other stakeholders working on AI technology and
its implications for peace and security. This project directly supports the Convention
on Certain Conventional Weapons Group of Government Experts on Emerging
Technologies in the Area of Lethal Autonomous Weapons Systems (GGE on LAWS)
in its efforts to advance multilateral debate on concepts such as human control and
responsibility, the human-machine interface, and the predictability and reliability of AI-
enabled conventional weapon systems (among other considerations). This project also
seeks to address considerations related to broader applications of AI in military systems—
particularly in decision-making support tools, cyber operations, and command and
control— which themselves raise novel concerns about understandability, reliability and
predictability; the potential for unintended interactions or outcomes; and susceptibility of
these systems to manipulation. The rate of technological progress in this space requires,
as the Secretary-General has described it, a “broader consideration of the impacts of
introducing autonomy and artificial intelligence into other military systems, and how
effective governance and risk mitigation can be achieved”. The implications of AI for
digital, physical and even political security require a fundamental reassessment and, in
some instances, re-equipping of the multilateral arms control toolbox.
UNIDIR’s AI and autonomy workstream seeks to a) support understanding of the
implications of military uses of AI in and beyond weapon systems and b) explore the
options available for AI governance and arms control. In addition to continuing its
work on autonomy in weapon and military systems, in 2022, UNIDIR adopted a new
research agenda on AI and Autonomy that focuses on building an understanding of
different types of AI, their different purposes and military uses, the broader international
security implications of advancements in AI, convergence between AI and other new and
emerging technologies and issues of AI governance.
• Project Type/Output: Framework/Strategy/Policy, Research (Fully fledged Development)
• Project Domain: Peace and Security, Lethal Autonomous Weapons Systems, Military
applications of AI
• Data Source: Research, Events
• Project Updates: In previous submissions, UNIDIR outlined that its future work will
encompass research studies and events at the regional and international level on the
science, significance, and solutions related to artificial intelligence and the weaponization
of increasingly autonomous technologies. This research agenda and program of
263
convenings form a fundamental part of UNIDIR’s efforts in support of the Group of

Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous
Weapon Systems, as well as other stakeholder communities. As part of this research
agenda, UNIDIR completed the following activity in 2023: the Programme updated a
resource paper (first published in 2022), which provides a comprehensive and comparative
analysis of the Proposals submitted to the GGE on LAWS. Furthermore, SecTec published
two research reports, which were launched during the General Assembly, both exploring
different aspects related to AI in the military domain and in the context of international
security . One study offers an analysis of uses of AI beyond weapon systems and another
provides a taxonomy of risks of AI (the latter is part of the UNIDIR project on confidence-
building measures for AI). The Programme’s ongoing research projects focus on the
following key themes: governance of AI and confidence-building measures, the role of
AI in disinformation, the AI-cybersecurity nexus. . In addition, the Programme continues
to develop AI Policy Portal (AIPP) which aims at gathering available information at the
national, regional and international level on policies, processes and structures that are
relevant to development and use of AI for military or security purposes.
Institutions
• Partnership(s)/Collaborator(s): Technology and academic research community, arms
control practitioners and other experts in the area of technology governance and
regulation.
• Links and Multimedia
o https://unidir.org/programmes/security-and-technology
• Lesson Learned:
o Challenges: Complexity and constant evolution of the research Project Domain,

uncertainties regarding possible future applications of military AI, low technological
literacy of many policy makers and a reluctance to adopt multistakeholder approaches
(particularly in cooperation with the private sector and technical community) to
international security challenges, exploring the governance of applications of AI that
do not fit within the mandate of existing multilateral arms control processes.
o Opportunities: demand is high for UNIDIR primers, briefings and multistakeholder
convenings such as the Innovations Dialogue (in 2023, the theme of the Innovations
Dialogue was The Impact of AI on Future Battlefields Through these outputs we seek
to create spaces to build knowledge, raise awareness among policy makers and
convene multi-stakeholder discussions on new technology issues, as mandated by
the Secretary-General in his Agenda for Disarmament.
o Lessons learned: – Need and demand for focused research on specific topics,
clarification of the scope and exact meaning of broadly used terms or concepts, as
well as description of process pertaining to the use of AI in the framework of military
operations or the weaponization of AI. – The multi-stakeholder approach continues
to be valuable for finding common ground and building constructive approaches
among stakeholders holding divergent or competing points of view. – Neutral expert
analysis is very much welcomed by stakeholders and policy makers.
• Contact Information: Dr Giacomo Persi Paoli (giacomo.persipaoli@un.org)"

SDG 16
3. Relevant Links
https://unidir.org/
264
Contact Information
Ioana Puscas (ioana.puscas@un.org), Sarah Grand Clement (sarah.grandclement@un.org) and
UNIDIR
Giacomo Persi Paoli (giacomo.persipaoli@un.org)
265
United Nations Industrial Development Organization
Project 1: Fostering Slovenian – Cuban innovation cluster for biopharma, medical

and nanotechnologies sectors 2020 – 2023
• Project Description:The project intends to develop a Slovenian – Cuban innovation
cluster model for biopharma, medical and nanotechnologies sectors along the period
2020 – 2021. The project can be considered a pilot model for interregional cooperation,
facilitating knowledge sharing and technology transfer from and to participant countries,
and could be replicated or extended to other countries in Latin American and Caribbean
region, and in Southeastern Europe/The Balkans region as well. More specifically the
project seeks:
o to support competent authorities in the upgrading of legal / regulatory framework for

development of the bio-pharma cluster;
o improving the conditions for larger business interventions (technology transfer, joint
design and production);
o increase the competitiveness in the field of bio-pharma industry of participating
countries;
o promote the development of innovative products, technologies and processes
contributing to improve the human health and quality of life.
UNIDO launched the e-training course “Application of Advanced Technologies and

New Business Models in the Biopharma and Medical Sectors” in May 2023. This training
course is intended to provide a solid foundation on the biopharma and medical industries
within the context of the Fourth Industrial Revolution (4IR). The online training comprises
four modules and intends to highlight the 4IR implications for developing countries,
particularly least developed countries, outline related challenges, and elaborate on how
to capitalize on the potential that 4IR technologies present in these industries, particularly
Artificial Intelligence (AI).
• Department/Division: Directorate of Technical Cooperation and Sustainable Industrial
Development/Division of SME Competitiveness, Quality and Job Creation/SME
Development and Job Creation Unit
• Project Type/Output:E-training
• End Year: 2023
• Project Domain: biopharma, medical and nanotechnologies sectors
• Related Sustainable Development Goals (SDGs): SDG 3 –Good health and well-being,
SDG 9 – Industry; Innovation and Infrastructure
266
• Lesson Learned: UNIDO provided capacity-building support for innovative technologies

and business models, emphasizing the potential of digital technologies like AI and the
opportunity to materialize them through joint ventures, partnerships, and technology
UNIDO
transfer. The combination of investment promotion tools and innovative solutions,
including digitalization, facilitated the identification of a significant number of common
projects with the aim to enhance industrial competitiveness and innovation.
• Media and Links:
o https://hub.unido.org/training-modules-aatbm
• Contact Information: Alejandro Rivera Rojas (A.RIVERA-ROJAS@unido.org)
Project 2: Promoting sustainable bush-processing value chains in Namibia

• Project Description: In Namibia, where agriculture is a primary livelihood but suffers from
low productivity, this project aims to bolster food and income sources by converting
invasive bush species into valuable products like animal feed and charcoal. This
initiative, responding to Namibia's challenges such as water scarcity, waste, and bush
encroachment, leverages digital tools for sustainable land management and biomass
processing. The NGGP plant, part of a strategic action plan, is set to produce high-value
items from Acacia species, enhancing agro-industrial productivity. This will improve local
and regional availability of livestock feed, energy, and other products, promoting better
meat and dairy quality, reducing imports, and boosting exports.
Achievements include:
o Technological solutions for Acacia Bush product development.

o Effective decision-making tools, including maps and innovative software.
o GIS data mapping and analysis for Acacia tree management.
o Automated Acacia detection using AI with satellite and drone images.
Outputs include:
o Advanced mapping, dashboards, and Web GIS tools.

o Updated geospatial data management.
o Automated analysis and mapping processes.
o AI-powered Acacia species detection and mapping process and technology.

Development/ Division of Digital Transformation and AI Strategies
• Project Type/Output: AI processing of satellite and drone imagery, reports, Dataset,
Software tools, Conferences
• Project Domain: Agriculture, Environment, Energy
• Data Source: Geographic Information System (GIS) data—reference maps, satellite
imagery, geospatial technology data, visual example data of Acacia bush trees
• Link to storymap: https://storymaps.arcgis.com/stories/d316418096244bb0b00ece
24082d7632
• The libraries used are open source, the code developed is proprietary.
• Data publicly available: Yes
• Technology/Platform: The project is developed using the R programming language using
RStudio.
267
o Outputs in 2023:
o The Ministry of Industry and Trade of Namibia (MIT), as in kind contribution for the
bush-processing plant in Otjiwarongo, has agreed to supporting the UNIDO Project
with key utilities for the plant, such as electricity and water supply connections.
o The project has contributed to the key content to be presented, and featured as
best practice in applying digital and AI solutions for real sector and socio-economic
development at the UNIDO’s launching of the Global Alliance on Artificial Intelligence
for Industry and Manufacturing (AIM Global) at the World Artificial Intelligence
Conference (WAIC) held in Shanghai, China on 6-8 July 2023 as well as at the AI for
Good Global Summit 2023 in Geneva, Switzerland on 5-7 July 2023.
o A team of local harvesters supervised by the harvesting advisor were hired to collect
bush harvest for trial sessions during the start-up of the bush-processing plant.
o The future recipes for the animal feed were verified and approved by the Ministry of
Agriculture.
o Terms of Reference for the future management company of the production plant was
developed and the Tender for the plant management company/plant operator was
announced by MIT.
o On 24 November 2023, MIT informed about the Ministry’s contribution of NAD
413,000.00 to support the installation of equipment at the plant, which is currently being
covered by the local developer (that erected the hub buildings and infrastructure).
o Impact in 2023:
o Based on a detailed analysis of the viability to convert bush biomass into livestock
feed and charcoal and taking into account environmental and social impacts, UNIDO
proposed a set of suitable approaches, technologies and production processes to
deliver bush-based final products for agricultural, chemical and pharmaceutical
purposes, as well as domestic use. The results of this research published in a Strategic
Action Plan proposes a market-oriented sustainable business model enabling to
establish a bush biomass-based processing and production that leads to higher
value-added products competitive on both local and external markets. This solution
is expected to have a multiplier effect of the pilot plant by a factor of 30-50 in Namibia,
and by 100 of similar plants to be established in the region.
o Thanks to the strong public-private partnership established with the support of UNIDO
between the Government of Finland and a private equity fund, the partners are jointly
establishing a unique pilot production plant that will be the first of its kind to convert
invasive bush species into livestock feed and charcoal – a key step towards establishing
zero-emission bush biomass processing and production in Namibia.
o Innovative digital technologies and know-how for responsible harvesting have been
transferred: specifically, the Machine Learning Model for acacia species mapping that
is based on remote sensing texture image analysis, satellite- and drone-supported
imagery recognition for enhanced performance of the agricultural sector and related
value chains during the COVID-19 crisis and beyond. Machine Learning algorithm was
fine-tuned to provide yield predictions, which will enable the NGGP to produce high
quality bio charcoal and animal feed. A special-purpose animal feed recipe developed
by the project experts will help farmers optimise livestock feeding. The main expected
benefits include the following:
1) Sustainable use of land in harmony with natural processes, improved farming

practices and productivity: reduced land degradation and diminishing invasive
bushes pave the way towards enhanced agricultural activities, which is the major
source of food supply for the local population;
2) Address issues related to water management and shortages, waste generation
and pollution;
268
3) Job creation: technological know-how, hands-on skills and ready-to-use business

models facilitated by the project will provide better opportunities for sustainable
jobs creation among women and men, especially in rural areas.
UNIDO
o Preparations for the operationalization of the special-purpose processing and
production plant are underway following the installation of the procured equipment.
National experts, including technicians, manufacturers, farmers, skilled and semi-
skilled workers, entrepreneurs and their associations, as well as staff of local agencies
and other stakeholders, are undergoing training on the collection and manufacturing
of products derived from bush/Acacia.
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 7 – Affordable

and Clean Energy ; SDG 8 – Decent Work and Economic Growth; SDG 9 – Industry,
Innovation, and Infrastructure; SDG 11 – Sustainable Cities and Communities; SDG 12 –
Responsible Consumption and Production
o UN Partners: UNIDO, EURO Trust Funds

o Government: Ministry of Foreign Affairs of Finland, Ministry of Industrialization, Trade
and SME Development of Namibia, Ministry of Agriculture, Water and Forestry of
Namibia,
o Private Sector Funding
o Civil Society: Trust Fund for Increased Food Security through Agribusiness, Trust
Fund for Trade-related Capacity Building, Walvis Bay Corridor Group, Namibian Meat
Board, Agricultural Bank of Namibia
o Academia: University of Namibia
o UNIDO Open Data Platform: https://open.unido.org/projects/NA/projects/170017

o ProDoc: https://open.unido.org/api/documents/7524581/download/ProDoc%20
Namibia%20170017%20signed.pdf
o Brochure: https://www.unido.org/sites/default/files/files/2019-12/NAMIBIA%20
BAOBAB%202020%20PDF%20WEB.pdf
o Strategic Action Plan for sustainable bush value chains: https://www.unido.org/sites/
default/files/files/2020-02/Namibia_v_2.20-spreads%20%281%29.pdf
o Twitter posts:
- https://twitter.com/UNIDOInnovation/status/1260240671551750145
- Practical Application: Web, Dashboard, AR-based manual on Mobile Devices
https://u nido. maps. arcgis. com/a pps/w ebappviewer/i ndex. html? id= 7 e
3265ab3bbc43b1a5b2a6b2c23dfc3a
269
o UNIDO Success Stories on Promoting Green and Smart Manufacturing, video

recording: https://xfiles.unido.org/index.php/s/cwNWW6792Rmo8M7
• Lesson Learned: The project resulted in a novel machine learning algorithm, a resource
that could be used in the future by further AI projects in Namibia and in regions
facing similar bush encroachment and land degradation issues. The deployment of
the technology has proven to be successful and the reliability of the Acacia-detection
algorithms is high; the algorithms themselves could be training for other vegetation
species and can be applied in other projects addressing issues related to
sustainable biomass processing and detection of eligible species, improving agricultural
productivity and strengthening drought resilience. This technological solution can
produce a multiplier effect in terms of providing a market-oriented sustainable business
model to benefit from biomass for production of competitive higher value added
products, identification of market niches at local and external markets, thereby facilitating
job creation.
This project was especially valuable as an exploration of the applications of Artificial
Intelligence and Machine Learning and the successful collaboration among several
international organizations.
As an added benefit and drawing upon the successful use of GIS in this project, UNIDO
has committed to promoting the creation of a Global GIS Community to raise awareness
on this technology, and to expanding GIS capacity across the organization.
The project methodology holds high potential for subsequent implementation in the
SADC Region, East Africa, Latin America, and even in Europe, where similar invasive
species pose threats to traditional or emerging agricultural activities. International
partners and national counterparts are already expressing strong interest in expanding
industrial development of market-oriented and sustainable value chains within the cattle-
related sector, including leather, textile, wool, and dairy products.
• Contact Information: Farrukh Alimdjanov (F.ALIMDJANOV@unido.org), Kawira Bucyana
(K.BUCYANA@UNIDO.ORG)
Project 3: Strengthening Implementation Design of Artificial Intelligence (AI) Eco-

System in Jordan
• Project Description: To support building the Artificial Intelligence (AI) Ecosystem in
Jordan aiming at creating job opportunities and improving the efficiency and quality
of government services as well as enhancing the comprehensive social and economic
development of different sectors, accelerating economic development, and creating
suitable opportunities for innovation and entrepreneurship thus making Jordan a regional
and enabling center for information technology.
• Project Background: Jordan's entrepreneurial ecosystem, pivotal for job creation and
economic growth, has seen significant development over the last decade, with notable
improvements in global entrepreneurship rankings. The country's efforts in nurturing
technology entrepreneurship have positioned Jordanian startups prominently in the Arab
World, showcasing a high level of education and experience among founders. Despite
these advances, entrepreneurs face challenges such as high taxes, regulatory hurdles,
and a need for legal reforms to foster a more business-friendly environment.
270
In response, the Jordanian government, recognizing the importance of digital

entrepreneurship and the digital economy, established the Ministry for Digital Economy
and Entrepreneurship (MoDEE). This initiative aims to support digital entrepreneurship,
UNIDO
enhance electronic payments, and develop digital skills. Key measures include expanding
broadband access, promoting digital payments, and facilitating access to finance and
global markets. Furthermore, to bolster Jordan's position in the digital economy and
leverage artificial intelligence (AI) for sustainable development, the government, in
collaboration with UNIDO industry stakeholders, created its National AI Strategy and
Implementation roadmap. This strategy, and its subsequent implementation roadmap,
aims to:
o Promote the use of AI in all key economic sectors.

o Build an enabling environment for AI that encompasses the legislative, regulatory and
technological environment.
o Develop a digital infrastructure to keep pace with AI needs and developments.
o Build AI-specialized Jordanian capacities, expertise and skills, and employ knowledge
in developing all sectors.
o Strengthen the role of the public sector in the use of AI and its applications and
build the necessary partnerships with the private sector with the aim of enhancing
productive pathways toward sustainable development.
o Strengthen AI business environment and increase investment and support for AI-
related initiatives and Jordanian startups in the IT sector, providing service-based
solutions.
o Build a well-established system for scientific research, development, application and
experimentation related to AI, and create the right environment for it.
o Raise public awareness and increase confidence in AI in the public sector and all
segments of society.
• Expected Outcome: Artificial intelligence ecosystem in Jordan established and the 2020
Jordan Artificial Intelligence Policy goals achieved.
• Project Type/Output: AI Strategy, Report, Technical Guidelines, and Practical Toolkit
• End Year: 2024
• Project Domain: Trade, Innovation, Digital Transformation, Industrial Technology,
Economic Competitiveness
• Related Sustainable Development Goals (SDGs): SDG 8 –Decent Work and Economic
Growth, SDG 9 – Industry; Innovation and Infrastructure; SDG 17 – Partnership for the
Goals
o UN Partners: UNIDO, European Union

o Government: Jordan
• Impact in 2023:
o The efforts by UNIDO, the Ministry of Digital Economy and Entrepreneurship of Jordan
(MoDEE), and other development partners under the project have contributed to
elevating Jordan's profile as an upper-middle-income country undergoing an AI-
powered transition. According to the Government AI Readiness Index published by
Oxford Insights in 2023, Jordan now ranks 55th out of 193 countries, as compared
271
to 63rd place in 2022 and 81st in 2021. This improvement can be attributed to the
publication of the AI Strategy and Implementation Roadmap for Jordan. Consequently,
Jordan is poised to become a regional leader in terms of AI readiness, opening doors
for additional innovation, investment promotion, and sustainable growth.
o The implementation of this project signifies a significant milestone in promoting AI-
powered digital transformation in Jordan and, ultimately, the development of a robust
AI ecosystem. The project's impact was presented by Mr. Abdelkader Bataineh, Senior
Director of the Policies and Information Department at MoDEE, during the launch of
UNIDO’s Global Alliance on Artificial Intelligence for Industry and Manufacturing (AIM
Global) on July 6, 2023, at the World Artificial Intelligence Conference in Shanghai,
China.
• Lesson Learned: Navigating the governmental political landscape takes time and effort
and thus has to be calculated in the overall timeframe of the project.
• Media and Links:
o https://www.dataguidance.com/news/jordan-ministry-announces-draft-ai-strategy
-and
• Contact Information: Farrukh Alimdjanov (F.ALIMDJANOV@unido.org)
Project 4: Industry 4.0 to foster youth employment in Tunisia and Côte d’Ivoire
• Project background: This project document outlines UNIDO’s proposed approach to
address a major socio-economic problem in Tunisia, as well as in the complementary
country Côte d’Ivoire, namely the high unemployment rates among youth and women. In
the context of the industrial sector, the focal problem lies in the lack of dedicated policy
and mechanisms as well as updated capabilities and infrastructure to benefit from the 4IR
in promoting growth and fostering youth employment. BMZ’s has a Special Initiative on
training and Job Creation with emphasis on technology and innovation development and
value chains promotion, under which the current project document is designed to boost
integration into 4IR ecosystems combined with increased and enhanced employment
opportunities and conditions in selected enterprises in Tunisia and Côte d’Ivoire. The
suggested industrial sector for intervention are: Mechanical (engineering), textile,
pharmaceutical, and agri-food. (Tunisia); ICT and agro-industry (Côte d’Ivoire).
The intervention will foster higher demand for skilled workers by addressing constraints
on the business side that prevents firm growth and cluster development and will support
youth and others to engage in these opportunities by increasing access to technology,
skills, and information about jobs and business prospects. Consequently, the project will
focus on agents working in the targeted clusters, especially youth and women, support
for labour market integration, entrepreneurship, incubators and training structures; and
strengthening of financing institutions (banks, micro-credit institutes, etc.). In addition,
alignment will be sought with the broader framework of relevant national policies.
The project is addressing the main challenge identified in Tunisia, as well as in the
complementary country Côte d’Ivoire, which is the high unemployment rates especially
among youth and women. Regarding the industrial sector, the focal problem to be
addressed is the identification and elaboration of missing points or gaps at dedicated
policy and strategies, capability and infrastructure to adopt and benefit from the 4IR to
promote growth and foster youth employment.
• Specifically, the involved countries share the following challenges on their efforts towards
transition to Industry 4.0:
• The lack of demonstrative institutions which showcase the potential of Industry 4.0
technologies.
• The lack of capacity to quantitatively measure the country’s progress towards Industry 4.0
adoption.
• The lack of relevant courses in education curricula to promote digital skills development.
272
• The absence of industrial accelerators to support the industrial development capacities

of start-ups.
• High unemployment rate among large youth populations.
UNIDO
• A considerable SMEs and labour base but lack technology, innovation, skills and access
to finance thus limiting formal sector growth.
• Inexistent or inadequate regulations, laws, rules and policies in favour of employment
(including youth and women) and SME development.
• Project objective: The main goal of UNIDO interventions is to support Tunisia and Côte
d’Ivoire in increasing youth employment, salaries and labour conditions through shaping
and consolidating a digital economy ecosystem. The envisaged wide impact of this
objective will contribute to a structural transformation of the economy in Tunisia and
set the ground for this transformation in Côte d’Ivoire in the targeted sectors, leading to
sustainable job creation.
• To accomplish this objective the project will carry out a set of activities that contribute to
the following conditions:
o An active governance system, including sound labor regulations.

o A supportive business environment.
o Efficient mechanisms to build 4IR capacity among labor and businesses.
o Enhanced awareness of the opportunities of 4IR.
• The project adopts an integrated approach that seeks to support the four conditions for
employment generation and job security by simultaneously strengthening policy and
regulations, and public-private partnerships that engage business. The project will also
help establish 4IR demonstrations and build institutional capacities for training labour
in 4IR and it will develop a support network among institutions in different countries
promoting 4IR investments, partnerships, and financing. The project has the following
six components, which will be fully implemented in Tunisia as the main and primary
beneficiary, whereas some components will be implemented (with a limited scope) for
Cote d’Ivoire (as the complementary country):
1. Enhanced and conducive business environment for employment generation and

adoption of 4IR technologies and methods within priorities VC and sectors (CIV will
undergo the 4IR reediness assessment; specific strategies will be developed for the
targeted sectors).
2. Establishing a SMART factory within or adjacent to training facilities, development
of curricula, facilities, and methodologies for 4IR integration and adoption (CIV will
benefit from trainings and capacity building/services offered by the smart factory in
Tunisia (i.e. virtual facilities)).
3. Development of dedicated training capacity on 4IR knowledge in vocational centers,
universities and business school (CIV to benefit from the trainings methods and
material adapted from Tunisia, as well as ones developed for CIV)
4. Establishing and strengthening of 4IR institutions and centers in selected country(s) for
shared training, knowledge, experience and activities (CIV will benefit from established
online platform for training, assistance, knowledge and experience sharing).
5. Pilot on adopting and operating 4IR technologies in a selected number of enterprises
(CIV will have a limited number of enterprises that will benefit from diagnosis,
instalment of new technologies, trainings and skills development for staff).
6. Investment promotion network digital transformation facilitation for attracting
investors, partners and finances for acquiring 4IR technologies.
• Artificial Intelligence Component: The component on Smart Factory will allow the use,
training and deployment of AI for increasing competitiveness of the enterprise sector.
In addition, component five, will bring equipment and technologies supported by
AI to enhance production and quality control, as well as to reduce waste and ensure
273
compliance with market standards. Under output four, a number of trainings on AI and
its application in manufacturing were provided to beneficiaries and trainers.
• Project Status: under implementation
• Project Domain:I4.0, Youth Employment, Digital Transformation, Industrial Technology,
Economic Competitiveness
Growth, SDG 9 – Industry; Innovation and Infrastructure;
o UNIDO
o Government: Tunisia, Cote D’Ivoire
o Rafik FEKI (R.Feki@unido.org)
Project 5: UNIDO Global Alliance on Artificial Intelligence for Industry and

Manufacturing (AIM Global)
• Project Description: AIM Global to catalyze international partnerships across
governments, industry, private sector, academia and civil society, committed to the
responsible development and deployment of AI and frontier technologies in industry
and manufacturing. The Alliance strives to redirect the trajectory of technological
advancement towards shared progress and reduced digital divide by advancing
industrial competitiveness, sustainable development, while ensuring fair and responsible
access to AI’s benefits across countries and industries. The Alliance seeks to support
global exchanges on the use of technology, reduce the digital gap, and promote the
safe, inclusive, and sustainable application of artificial intelligence (AI) in industry and
manufacturing. By uniting key players from government, industry, academia, and civil
society, the Alliance aligns with the UN Secretary-General’s call in the "Our Common
Agenda" report for a Global Digital Compact to ensure an "open, free, and secure digital
future for all." It aims to enhance industrial competitiveness and sustainable development
through AI, ensuring its benefits are distributed fairly and sustainably, while addressing
challenges like AI literacy, ethical considerations, and data protection.
• Project Type/Output: Convening Function, Events, Knowledge Products
• End Year: continuous.
• Project Domain: Digital Transformation, AI, IR 4.0, IR 5.0
• Publicly available data: https://aim.unido.org/
• Related Sustainable Development Goals (SDGs): SDG 1 – No Poverty; SDG 7 – Affordable
and Clean Energy; SDG 8 – Decent Work and Economic Growth; SDG 9 – Industry,
Innovation, and Infrastructure; SDG 11 – Sustainable Cities and Communities; SDG 12 –
Responsible Consumption and Production; SDG 17 – Partnerships for the Goals
274
o UNIDO
o Private sector: Huawei (China)
UNIDO
• Relevant Links and Multimedia: https://aim.unido.org/
o Contact Information: Ana Paula Nishio de Sousa (a.nishio@unido.org)
Project 6: Alliance for I4.0 and Smart Manufacturing in Africa (AISMA)

• Project Description: Africa has been left behind during the past industrial revolutions,
despite its affluent population and natural resources. In this context, Industry 4.0 and
smart manufacturing offer significant opportunities and challenges to manufacturing
companies in the African continent. But failure to harness these opportunities and
overcome the challenges will undoubtedly impose considerable risks on the socio-
economic development of African countries, especially for youth at risk of unemployment.
Hence, without efforts to coordinate, develop and improve the existing policies,
institutional and legal frameworks, physical/digital infrastructure, innovation models,
education/ vocational training programmes and funding schemes on the continent,
African manufacturing companies risk falling further behind, exacerbating the global
“digital divide” and lowering Africa’s global competitiveness.
The Alliance aims to address this by unlocking the potential of smart manufacturing in
Africa through raising awareness on Industry 4.0, fostering collaboration among diverse
stakeholders within the Industry 4.0 ecosystem and facilitating the adoption of frontier
technologies in manufacturing companies across the continent.
AISMA will provide a collaboration platform for a wide range of experts from academia,
industry, private sector, policy makers, technology providers, financial institutions or
any other structure relevant to the adoption of Industry 4.0 within the African continent.
The network of experts will focus on exchanging knowledge and expertise, sharing best
practices and forging strategic partnerships on Industry 4.0 related themes such as
industry applications, digitalization technologies, skills development, scientific research
and innovation, standardization and policies/strategies, just to name a few. AISMA will
also offer a number of tools and support functions to its members.
AISMA ultimately aims to promote Africa's integration, generate inclusive and sustainable
development, stimulate job creation, empower people and break the digital divide, in
accordance with the UN Secretary-General's call to establish a Global Digital Compact
that upholds shared principles for an open, free, and secure digital future for all. It
also aligns with the goals and objectives set by the African Union through the Digital
Transformation Strategy for Africa (2020-2030).
UNIDO launched the Alliance for I4.0 and Smart Manufacturing in Africa (AISMA) on 28
November 2023 on the margin of the twentieth session of UNIDO General Conference.
For its launch, several leading experts from various organizations committed to
support UNIDO’s efforts to enhance Africa industrial competitiveness and sustainable
development through Industry 4.0 and smart manufacturing.
The Alliance will contribute to:
o UNIDO’s motto – Progress by innovation

o The Alliance for I4.0 and Smart Manufacturing in Africa (AISMA), aligns with UNIDO's
motto (“Progress by Innovation”), by creating an environment that is conducive for
innovation, as it brings experts, academia, industry professionals and decision makers
from the African continent together, and provide the tools to exchange and develop
innovative initiative and approaches to address the needs of the African continent.
o UNIDO’s vision - a leading platform for knowledge and technology transfer, innovation
and investment
o It aims to raise awareness of Industry 4.0 (I4.0) in Africa, facilitate the adoption of I4.0
technologies, and provide a platform for information sharing, knowledge creation, and
innovative proposals on I4.0 related themes. In addition, it fosters collaboration among
275
diverse stakeholders within the Industry 4.0 ecosystem and facilitates the adoption of
frontier technologies across the continent, aligning with UNIDO's vision and strategic
objectives.
The Alliance will have the following functions:
o Conducting research, surveys and in-depth data analysis on 4ir related issues in the
African context
o Prepare policy briefs for decision makers
o Monitor and evaluate impact of policies and initiatives of I4.0
o Creating a space for communication and exchange of experience knowledge and
data.
o Providing a facility for training, capacity building and skill development
o Facilitating interaction between policy makers, academia, private sector, technology
providers and financial institutions to fine-tune and align policies and initiatives.
o Supporting industry define their specific digitalization & industrialization journey and
accessing/implementing I4.0 methods, technologies and tools
o Identifying mechanisms for financing I4.0 adoption in Africa and facilitating access to
such mechanisms for startups, SMEs and digital entrepreneurs
o Promoting conducive legal and ethical frameworks for I4.0 applications by providing
recommendations for governments and industries
o Promotion and raising awareness on I4.0 themes (impact on employment, gender,
business models, best practices, etc.)

• Project Type/Output: policies and strategies, standardization, training and skills
development, digitalization technologies, industry applications
• End Year: continuous
• Project Domain: Smart manufacturing in Africa, industry 4.0
• Publicly available data: https://www.unido.org/sites/default/files/files/2024-02/AISMA
_Leaflet.pdf
• Reported as part of 2022 Compendium on UN AI Activities?: No
• Related Sustainable Development Goals (SDGs): SDG 9 – Industry; Innovation and
Infrastructure;
• Relevant Links and Multimedia: https://www.unido.org/sites/default/files/files/2024-02/
AISMA_Leaflet.pdf
• Contact Information: Rafik FEKI (R.FEKI@unido.org)

SDG 1 – No Poverty; SDG 7 – Affordable and Clean Energy; SDG 8 – Decent Work and Economic
Growth; SDG 9 – Industry, Innovation, and Infrastructure; SDG 11 – Sustainable Cities and
Communities; SDG 12 – Responsible Consumption and Production; SDG 17 – Partnerships
for the Goals
3. Related Links
UNIDO Main Web site: www.unido.org
276
UNIDO Open Data Platform: https://open.unido.org
UNIDO IATI Publication (d-portal) https://d-portal.org/ctrack.html?reporting_ref=XM-DAC
UNIDO
-41123#view=main
UNIDO Technical Cooperation: https://www.unido.org/technical-cooperation
General Contact Information: das@unido.org
277
United Nations Institute for Training and Research
Project 1: NVIDIA
• Project Description: The United Nations Satellite Centre (UNOSAT) announced a
collaboration with NVIDIA on training and research activities to promote the use of
Artificial Intelligence (AI) for Earth Observation activities in support of the Sustainable
Development Goals (SDGs), with an initial emphasis on disaster management. This
cooperation framework allows UNOSAT and NVIDIA to benefit from their respective
facilities, resources, and domain experience. The collaboration has two initial priorities: 1)
integration of NVIDIA’s accelerated computing platform within UNOSAT’s infrastructure
to fast‐track research and development of AI for Earth Observation and 2) design and
roll‐out of an e‐learning course on the use of deep learning for flood detection to upskill
data scientists within disaster management agencies worldwide.
• Department/Division: United Nations Satellite Centre (UNOSAT)
• Project Type/Output: Research and Training
• Project Domain: Other – Geographic Information System (GIS)
• Data Source: Copernicus Sentinel-1 satellite imagery, flood maps, exposed population
impact layer
• Technology/Platform: NVIDIA Deep Learning Institute
Health and Well-Being; SDG 6 – Clean Water and Sanitation; SDG 8 – Decent Work and
Economic Growth; SDG 9 – Industry, Innovation, and Infrastructure; SDG 10 – Reduced
Inequalities; SDG 11 – Sustainable Cities and Communities; SDG 13 – Climate Action; SDG
15 – Life on Land; SDG 16 – Peace, Justice, and Strong Institutions; SDG 17 – Partnerships
for the Goal
• Partnership(s)/Collaborator(s)(s):
o NVIDIA
o https://blogs.nvidia.com/blog/2022/06/24/un-satellite-centre-boosts-sustainable
-development-goals/
o https://courses.nvidia.com/courses/course-v1:DLI+S-ES-01+V1
• Contact Information: Einar Bjorgo (einar.bjorgo@unitar.org)
278
Project 2: UNOSAT S-1 FloodAI

• Project Description: The UNITAR-UNOSAT designed, developed, and deployed UNOSAT
UNITAR
S-1 FloodAI: an end-to-end pipeline where Copernicus Sentinel-1 Synthetic Aperture
Radar (SAR) imagery of flood-prone areas are automatically downloaded and processed
by a deep learning model to output flood vector data and update operational dashboards.
Access to timely and accurate data could not only inform the decision-making process to
help optimize the disaster response, but it also has the potential to significantly reduce
the loss of life and mitigate structural damage, particularly in the context of humanitarian
operations, thus supporting both national authorities and international emergency
management organizations for the benefit of affected populations.
• Project Domain: Other – Geographic Information System (GIS)
• Data Source: Copernicus Sentinel-1 satellite imagery, flood maps, exposed population
impact layer, JRC permanent water layer
• Technology/Platform: UNOSAT S-1 FloodAI was deployed on a local GPU at the European
Organization for Nuclear Research (CERN) connected to a 64 TB data storage server and a
high-speed CERN internet connection. The entire infrastructure is currently in the process
of being transferred to a cloud centralized service at CERN built on Kubeflow, a machine
learning platform on Kubernetes. The deep learning model was written in Pytorch. The
operational dashboard is based on an ESRI dashboard linked to a web map.
• Project Updates: Previously reported in 2021 as “Project 3: Flood Mapping – UNOSAT
FloodAI”, is now fully deployed and embedded as one of the UNOSAT Rapid Mapping
Service.
for the Goal
o UN Partners: UN Global Pulse

o Academia: European Organization for Nuclear Research (CERN) , NVIDIA
o https://www.unitar.org/about/news-stories/news/unosat-flood-ai-dashboards-nepal
-creation-one-stop-shop-real-time-evidence-based-decision-making
o https://www.mdpi.com/2072-4292/12/16/2532
o https://ahacentre.org/wp-content/uploads/2022/07/ARMOR-3rd-Ed.pdf
Project 3: Mapping Refugee Settlement and Damage Assessment with Machine

Learning and Remote-Sensing Data
• Project Description: The purpose of this project is the creation of an end-to-end pipeline
that takes high-resolution satellite imagery as input and returns damage assessment
maps in the form of a building footprint together with a damage class label.
279

• Project Domain: Geographic Information System (GIS)
• Data Source: Satellite data, building footprints, damage assessment data
• Data Publicly Available: No
• Project Updates: Previously reported in 2020 as “Project 3: Mapping refugee settlement
with machine learning and remote-sensing data”. The focus is now not only on mapping
refugee settlement, but on building footprints in general and damage assessment.
Health and Well-Being; SDG 9 – Industry, Innovation, and Infrastructure; SDG 10 –
Reduced Inequalities; SDG 11 – Sustainable Cities and Communities; SDG 13 – Climate
Action; SDG 15 – Life on Land; SDG 16 – Peace, Justice, and Strong Institutions; SDG 17
– Partnerships for the Goals
o Academia: European Organization for Nuclear Research (CERN), CERN Openlab,

Wuhan University
o https://arxiv.org/abs/2201.10953v2
Project 4: ML4Floods
• Project Description: UNITAR-UNOSAT partner with Trillium Technologies and FDL Europe
to test and use ML4Floods: an ecosystem of data, models and code pipelines to tackle
flooding with machine learning ML. After a successful testing phase of the methodology,
UNOSAT is implementing ML4Floods into its operations and deploying the tool into the
UNOSAT AI pipeline at CERN.
• Project Domain: Geographic Information System (GIS)
• Data Source: Copernicus Sentinel-2 images, flood maps from Copernicus EMS, UNOSAT
and GloFIMR, JRC Permanent water layer
• Technology/Platform: Google Cloud Platform, Pytorch, pytorch lightning, weights and
biases, GDAL, jupyter notebook, colab tutorials.
for the Goals
280
o Private Sector: Trillium Technologies
UNITAR
o Academia: Frontier Development Lab Europe
o http://trillium.tech/ml4floods/content/intro/introduction.html
o ML4Floods time series segmentation
Project 5: AI and Diplomacy Trainings

• Project Description: The Division for Multilateral Diplomacy (DMD) launched a series of
training activities in the field of Artificial Intelligence (AI) and Diplomacy to the benefit
of the diplomatic community. The aim of this initiative is to delve into the sphere
of AI and Diplomacy, analyzing the impacts, implications, and frameworks in place
needed to navigate the opportunities and intricacies that AI is bringing to the diplomatic
world and overall global context. These trainings allow DMD to support the achievement
of the Sustainable Development Goals (SDGs) and support the reskilling and upskilling
of diplomats and UN personnel to adapt to the future of work and become empowered
and informed global decision-makers.
• Department/Division: Division for Multilateral Diplomacy (DMD)
• Project Type/Output: Training and Research
• Project Domain: Diplomatic Trainings
• Data Source: UNITAR
• Data Publicly Available: n/a
• Technology/Platform: n/a
• Related Sustainable Development Goals (SDGs): SDG4 Quality Education; SDG 8 –
Decent Work and Economic Growth; SDG 9 – Industry, Innovation, and Infrastructure;
SDG 10 – Reduced Inequalities; SDG 16 – Peace, Justice, and Strong Institutions; SDG
17 – Partnerships for the Goal
o n/a
o https://unitar.org/courses/free-face-face-training-introduction-generative-artificial
-intelligence-geneva-based-permanent-11704
o https://event.unitar.org/full-catalog/free-face-face-training-artificial-intelligence-and
-diplomacy
• Contact Information: Amine Mesdoua (amine.mesdoua@unitar.org); Malcolm Olivieri

(malcolm.olivieri@unitar.org)
Project 6: AI Forum “Three Horizons: Governance, Education and the Future of Work”
• Project Description: The Division for Multilateral Diplomacy (DMD) and IE University
(IE) jointly hosted a Forum on Artificial Intelligence (AI) in Madrid. The Forum focused
on various aspects of AI, including governance, ethics, education, and the corporate
world. Entitled, ‘Three Horizons: Governance, Education and the Future of Work’, it
281
brought together around 100 participants from a variety of sectors, including government
officials, academia, private sector specialists, and education and training providers, to
discuss the impact of AI across various domains. UNITAR and IE jointly organized the event
within the premises of the IE Tower in Madrid. This Forum allowed DMD to support the
achievement of the Sustainable Development Goals (SDGs) and spread the knowledge
on pivotal areas for all beneficiaries from different sectors to become empowered and
informed professionals, educators, and global decision-makers.
• Department/Division: Division for Multilateral Diplomacy (DMD)
• Project Type/Output: Training and Research
• Project Status: Developed
• Project Domain: Diplomatic Trainings
• Data Source: UNITAR
• Data Publicly Available: n/a
• Technology/Platform: n/a
• Related Sustainable Development Goals (SDGs): SDG4 Quality Education; SDG 8 –
Decent Work and Economic Growth; SDG 9 – Industry, Innovation, and Infrastructure;
SDG 10 – Reduced Inequalities; SDG 16 – Peace, Justice, and Strong Institutions; SDG
17 – Partnerships for the Goal
o IE University
o https://unitar.org/about/news-stories/news/unitar-and-ie-university-jointly-hosted
-forum-artificial-intelligence
• Contact Information: Amine Mesdoua (amine.mesdoua@unitar.org); Malcolm Olivieri

(malcolm.olivieri@unitar.org)

SDG 1, 3, 6, 8, 9, 10, 11, 13, 15, 16, 17
3. Related Links
https://unitar.org/
Contact Information
Einar Bjorgo (einar.bjorgo@unitar.org)
282
United Nations Joint Staff Pension Fund
UNJSPF
Project: UNJSPF Digital Certificate of Entitlement
Project Description: The UNJSPF “Digital Certificate of Entitlement” (DCE) system is a state-
of-the-art mobile application (App) and web-based authentication solution that can verify the
identity, existence, transactions, and location of the United Nations Joint Staff Pension Fund
(UNJSPF) retirees (approx. 84,000 individuals located in more than 192 countries), providing
a faster, secure, and easier way to meet the requirements for continued benefit payments.
The DCE system allows retirees to use their smart phone, tablet, or other personal device
to submit an electronic proof of their eligibility, in lieu of paper-based submissions. The
solution uses cutting edge biometric/facial recognition, blockchain, AI, and global positioning
technologies to confirm the eligibility of retirees.
The DCE was developed by UNJSPF with the support of the UN International Computing
Center, and piloted in 2020, with retirees from the World Food Programme and the Food and
Agricultural Organization. Its deployment was then fast-tracked in 2021, as a response to the
worldwide disruption of postal services caused by the COVID-19 pandemic.
The DCE introduced innovative processing changes and organizational efficiencies, using a
secure and user-friendly design based on new technologies, making it an environmentally
friendly solution, which significantly reduced the use of paper, mailings, as well as physical
and digital storage.
Aligned with the UN Secretary-General’s strategies on “New Technologies” and “UN2.0”, and
the vision of a digital United Nations, the DCE is part of the implementation of the UNJSPF
strategy in simplifying client experience and modernizing the Fund’s services.
Assurance on the DCE solution is based on independent certifications and audits, including:
• ISO27001 Certification
• Ethics Audit
• Algorithm Audit
NIST verification of the biometric algorithm The DCE solution received the:
• United Nations Secretary-General Award for Innovation and Sustainability; and

• Government Blockchain Association Award for Social Impact.
Furthermore, the DCE solution was subject of a case study published by Gartner Inc., as an
example of digital transformation in the public sector.
283
As of November 2023, over 40 per cent of the UNJSPF retirees’ population has sought enrolment
in the UNJSPF/DCE solution. UNJSPF recently deployed a new “kiosk version”, extending the
ability to use the DCE solution to UN retirees who do not own a computing device.
For more information about the UNJSPF Digital Certificate of Entitlement please visit https://
www.unjspf.org/newsroom/unjspf-digital-certificate-of-entitlement-wins-the-united-nations
-secretary-general-award-for-innovation-and-sustainability
• Project Type/Output: Software tool, Conference

• Project Status: Completed (and in production)
• Project End Year: 2021 (With ongoing expansion/extension of the scope)
• Project Domain: Environment, Health, Human Rights, Poverty, Telecommunications,
Other: Governance & Public Services, Finance & Banking, Supply-chain, Cybersecurity
• Data Source:
o Demographic data = UNJSPF retirees and beneficiaries -> Held and processed on
UNJSPF/UNICC systems covered by UN Privileges and Immunities
o Biometrics, transactions, and locations data = Retirees and beneficiaries’ devices
-> Held on users’ devices and protected by tamper-resistant “Hardware Security
Modules”, with advanced encryption
o In accordance with ISO/IEC 29100/2011/Amd 1:2018, personally identifiable
information are not stored on the blockchain.
The DCE Technology Stack (version 4.0), is composed of:
• UI Screens & User Interaction = React Native/TS/JS

• Custom Application Modules (platform dependent):
o Services
- Face features extraction service: Insightface; PyTorch

- Face verification service: Insightface; PyTorch -> UNICC Biometric AI Solution
(NIST FRVT verified)
- ID Document scanning AI solution -> UNICC Doc Scan Solution
- Real-time liveness and face detection service: ML Kit (Android); iOS Core ML
(Apple)
- Hardware encryption service: Hardware-backed keystore (Android); Apple Secure
Enclave (Apple)
o Platforms: Kotlin/Java (Android); Swift/Objective C (Apple)

o Blockchain:
- Digital Wallet: Hyperledger Aries

- Decentralized PKI: Hyperledger Indy
• Related SDGs: SDG 1 – No Poverty, SDG 3 – Good Health and Well-Being, SDG 4 – Quality
Education, SDG 5 – Gender Equality, SDG 8 – Decent Work and Economic Growth, SDG 9
– Industry, Innovation, and Infrastructure, SDG 10 – Reduced Inequalities, SDG 16 – Peace,
Justice, and Strong Institutions, SDG 17 – Partnerships for the Goals
• Partners:
284
o UN Partners: United Nations International Computing Centre (UNICC)
UNJSPF
o Digital Certificate of Entitlement (DCE) - UNJSPF (https://www.unjspf.org/for-clients/
digital-certificate-of-entitlement/))))
o Retirees and Beneficiaries: the new Digital Certificate of Entitlement App is now live
- UNJSPF
o The Certificate of Entitlement soon in digital format – watch the video - UNJSPF
o Ensuring that the DCE App performs the “biometric test/validation” only locally on
the device (i.e., no transmission/storage of biometrics data outside the control of the
user).
o Inability to use biometric-based authentication mechanisms natively available in the
Android/Apple devices (i.e., SW libraries exposing limited functionalities; Need to
limit App access to extended biometrics information.
o Different approaches between Apple and Android devices in capturing user’s face in
3D.
o Challenging “environmental conditions” (i.e., low lighting; background “noise”),
during biometric profile capturing.
o Relatively limited availability of “modern/secure” devices (i.e., devices with HSM
module) in the user population.
o Communicating user-friendly instructions
• Contact information: Dino Cataldo DELL’ACCIO (dellaccio@un.org)

SDGs 1, 3, 4, 5, 8, 9, 10, 16, 17
3. Relevant Links
https://www.unjspf.org/newsroom/unjspf-digital-certificate-of-entitlement-wins-the-united
-nations-secretary-general-award-for-innovation-and-sustainability
Contact information: Dino Cataldo DELL’ACCIO (dellaccio@un.org)
285
United Nations Office for the Coordination of

Humanitarian Affairs
Project 1: Global Tropical Cyclone Impact Model
Project Description: Tropical cyclones (known as hurricanes or typhoons in certain regions) can
cause significant humanitarian impact, with up to 800 million people affected yearly around
the world. By using machine learning methods and existing forecasts, the impact of cyclones
can now be anticipated, allowing humanitarian organizations to respond more quickly and
efficiently.
The project initially aimed at developing a trigger mechanism for the tropical cyclone early
action protocol of the Philippines Red Cross Forecast Based Financing project, in partnership
with The Netherlands Red Cross’ data and digital initiative ‘510’. A model was developed that
predicts the potential damage of a tropical cyclone before landfall. The model uses ‘inputs’
or ‘features’ for the cyclone track such as windspeed, and rainfall, as well as static features like
topography, housing building material, household relative wealth, and population density.
Based on these inputs, the model then determines the percentage of completely damaged
houses per municipality. The model is currently being used in the OCHA anticipatory action
framework for the Philippines, where pre-agreed financing is released for a pre-agreed action
plan by UN agencies if a certain level of damage is predicted.
In collaboration with the ISI Foundation, the model has now been adapted to work using a grid
(instead of the specific municipalities in the Philippines) using open global data sources, which
helps enable its application globally, and improves its accuracy. The new grid-based model is
currently being extended for use in Fiji to support the OCHA anticipatory action framework,
and will be used in other contexts in the coming year.
• Project Type/Output: Policy framework, Academic paper, Software tool

• Project Domain: Weather
• Data Source:
o Satellite weather data such as wind speed and rainfall

o Topography data such as slope, terrain ruggedness index, elevation, coast length
o Vulnerability data such as housing material, proportion of area that is classified as
urban or rural, Meta’s relative wealth index data
o Demographic data such as population per municipality
286
• Technology/Platform: Model is coded in Python, with analysis in Jupyter notebooks.

Automated model runs are done with GitHub Actions.
• Related SDGs: SDG 10 – Reduced Inequalities, SDG 13 – Climate Action, SDG 17 –
UNOCHA
• Partners:
o UN Partners: OCHA Philippines Country Office, OCHA Pacific Islands Regional Office
o Government: Fiji Meteorological Services, Fiji National Disaster Management Office
o Civil Society: 510 – Netherlands Red Cross
o Academia: ISI Foundation
o GitHub repository: https://github.com/OCHA-DAP/GlobalTropicalCycloneModel

o OCHA Anticipatory Action Framework: https://reliefweb.int/report/philippines/
philippines-anticipatory-action-framework-2022-revision
o Peer review of model: https://centre.humdata.org/peer-review-of-510s-typhoon
-model-and-its-use-in-the-philippines/
o Academic paper: https://nhess.copernicus.org/articles/24/309/2024/nhess-24-309
-2024.html
• Lessons Learned: One key challenge in the project was accessing and processing all
the input data required to train and run the model. Most important is reliable impact
data for historical cyclones. To be useful for model training, the impact data needs to
be geographically specific; in the very least, sub-national. However, this data is often
unavailable, which limits the model’s accuracy. Historical forecasts of tropical cyclones
can also be difficult to find and access in an easily usable format, further limiting the
extent to which the model can be trained.
OCHA and its partners are currently working to extend the grid-based model to be
used globally. This requires finding and assessing the validity of global public datasets
for model training. As the model is applied to further contexts, other country-specific
datasets will be incorporated where possible.
The aim is to make the model available publicly in a relatively easy-to-use format, such
as a web application.
• Contact information: Leonardo Milano (leonardo.milano@un.org)
Project 2: AI based content classification service
Project Description: Manually analyzing a large volume of text data can be a tedious and
time-consuming task. By leveraging Azure Open AI service, the process becomes significantly
more efficient. The service, built using Microsoft Power Platform, enables end-users to feed
in text snippets or excel spreadsheets, receiving categorized content in return. This versatile
service can be configured to analyze diverse content types, including news snippets, project
descriptions, social media posts, or budgetary spreadsheets.
One notable application involves the Climate Team at OCHA, utilizing the service to analyze
historical data within OCHA-administered Country-based Pooled Funds. Project summaries
are input into the service, allowing the discernment of activities specifically relevant to climate
adaptation. This enhances OCHA's ability to optimize the effectiveness and impact of these
initiatives in the future.
Additionally, the Digital Services section at OCHA employs the service to analyze Internet
or Cyber Security incidents, categorizing them into types such as Internet Outage and Data
287
Breach. Leveraging Microsoft Power Automate Flow for real-time classification, the data feeds
directly into a Power BI dashboard used for monitoring purposes.
We are also exploring applications such as automatically sifting through social media postings
to identify those most relevant to humanitarian work.

• Project Domain: Humanitarian Information
• Data Source: The service is generic and can be applied to any context.
• Technology/Platform: Microsoft Azure Open AI, Microsoft Power Platform
• Related SDGs: SDG 13 – Climate Action
• Lessons Learned: We aim to test the service with diverse content types to
comprehensively understand its limitations. In our experience, when utilizing the
service to identify climate adaptation activities, we encountered instances where the AI
generated new categories despite explicit prompts to the contrary. To address this, we
plan to explore the application of fine-tuned models for specific variations. Engaging
in more diverse applications will undoubtedly contribute to a deeper understanding
of the service's capabilities and areas for improvement.
• Contact information: Ikramullah Quraishi (ikramullah.quraishi@un.org), Ranu Gupta
(guptar@un.org)
Project 3: Automated tagging of ReliefWeb Jobs

• Project Description: Leverage a Large Language Model to automatically apply certain
categories to jobs being posted to ReliefWeb - https://reliefweb.int/jobs. The intention is
to reduce human efforts in areas that we believe AI can provide a suitable solution thereby
allowing our human capacity to focus on more impactful activities.
• Project Domain: Automatic tagging of humanitarian jobs being posted on ReliefWeb
• Data Source: Job descriptions are provided by various humanitarian and development
entities – either through an automated feed, by email, or by self-posting. We are using
LLM without any fine-tuning as we have found the results to be more than sufficient ‘as
is’.
• Technology/Platform: LLM is Microsoft Azure OpenAI, Website is Drupal
• Lessons Learned: Over-engineering is an easy trap to fall into. In our case, we were
able to back away from advanced work and costs (e.g. fine-tuning, dedicated services
running at all times) when we simply tested the LLM directly and found the results to
be satisfactory.
• Contact information: Andrej Verity (verity@un.org)
Project 4: ReliefWeb Q&A Chatbot

• Project Description: As a first version, the ReliefWeb Q&A Chatbot will allow a site visitor
to ask questions about the (single) report page that they are viewing. On average, a single
288
report had an attached file of approximately 17 pages. The Q&A Chatbot will retrieve
answers from that material.
UNOCHA
• Project End Year: 2024+
• Project Domain: Humanitarian Information
• Data Source: Reports stored on ReliefWeb. Link to data: https://reliefweb.int/updates
• Technology/Platform: Microsoft Azure OpenAI, AWS Bedrock, Drupal
• Partners:
o Private Sector: Amazon AWS
• Lessons Learned: Trying to provide a Q&A (RAG) across many documents was
challenging when users knew the content extremely well – as they did not find the
results very usual (e.g. missing specific details or nuance). Focusing on a single report
reduces that challenge and thus us time to gather feedback, better understand the
technology and investigate how to expand/grow the solution.
• Contact information: Andrej Verity (verity@un.org)
Project 5: United Nations’ Central Emergency Response Fund (CERF) AI project

summarization tool
• Project Description: We’ve developed an AI summarization tool using Microsoft Azure
OpenAI for the United Nations’ Central Emergency Response Fund (CERF). This tool
efficiently processes Microsoft Word project proposals, extracting relevant information
from various sections to generate a three-sentence project summary. The summary
serves dual purposes: it’s publicly available on the CERF website, and it’s submitted to
OCHA management (including the OCHA USG) for project approval. Traditionally, CERF
program officers manually crafted these summaries, which consumed approximately 5
minutes each. Our tool, however, produces accurate summaries within seconds. Given the
volume of over 500 CERF projects annually, this tool stands to save more than 40 hours
of staff time per year—a significant efficiency gain.
• Project Domain: Humanitarian Financing
• Data Source: Project objective data, project activity data, targeted beneficiary data
(disaggregated by sex, age and persons with disabilities)
• Technology/Platform: Microsoft Azure OpenAI
• Lessons Learned: We learned that creating an AI tool can take several iterations. The
CERF AI project summarization tool required several attempt to create the prompt
that would produce the project summarizations in the correct format. In the future,
we plan to expand the suite of summarization tools which will include an application
summarization tool and a reports summarization tool which are all currently manually
written.
• Contact information: Samir Mahmoud (mahmoud14@un.org)
289

SDGs 10, 13, 17
3. Relevant Links
https://www.unocha.org/
Contact information: Andrej Verity (verity@un.org)
290
United Nations Office for Counter Terrorism
UNOCT
Project 1: Building Knowledge on Counter-Terrorism in the Age of Artificial

Intelligence
• Project Description: In 2020, UNOCT’s Global Counter Terrorism Programme on
Cybersecurity and New Technologies collaborated with UNICRI, through its Centre for
Artificial Intelligence and Robotics commenced a research initiative aimed at exploring
the dual nature potential of artificial intelligence (AI) from the perspective of counter-
terrorism.
The initiative explored the dual nature potential of AI, analyzing the concerning aspects of
the advent of AI, such as the possibility of its use with malicious intent by terrorist groups
and individuals, as well as how AI might be leveraged to support counter-terrorism efforts,
in particular in terms of combatting terrorist use of the Internet and social media. This
resulted in the release of two reports “Algorithms and Terrorism: The Malicious Use of
Artificial Intelligence for Terrorist Purposes” and “Countering terrorism online with artificial
intelligence – An Overview for Law Enforcement and Counter-Terrorism Agencies in South
and South-East Asia”.
UNOCT, UNICRI, and OHCHR organized a high-level briefing entitled “Counter-Terrorism
in the Age of Artificial Intelligence: Risks, Opportunities and Safeguarding Human Rights”
on 29 June 2021 as part of the Second Counter-Terrorism Week at the United Nations.
During this briefing the findings of their collective research were presented.
UNOCT and UNICRI organized an expert group meeting to explore the specific
application of AI-enabled social network analysis in counter-terrorism in March 2022 to
develop knowledge and understanding of this application and develop a concept for
further activities in this domain.
• Department/Division: UNOCT’s Cybersecurity and New Technologies programme and
UNICRI’s Centre for Artificial Intelligence and Robotics
Institutions
o UN Partners: UNICRI and OHCHR

o Government: Kingdom of Saudi Arabia, Government of Japan
291
malicious-use-of-ai-uncct-unicri-report-hd.pdf
countering-terrorism-online-with-ai-uncct-unicri-report-web.pdf
• Lesson Learned:
o Challenges:
- Ensuring human rights compliant AI and building public trust in the use of AI by
law enforcement and counter-terrorism agencies, especially given the complex
interaction between human rights and counter-terrorism.
- Absence of consensus on the definitions of both terrorism and AI.
- Insufficient public information related to the level of technological readiness and
the current use of AI tools in the regions of South Asia and South-East Asia.
- Lack of understanding of how malicious actors could use AI.
o Lessons learned:
1. While the use of AI for terrorist purposes is currently not a developed threat, it is
important to not underestimate it.
2. The capacity of all stakeholders to identify and respond to the threat of the
malicious use and abuse of AI for terrorist purposes should be improved.
3. Efforts need to be made to raise awareness of governments and industry partners
about the role of AI in counter-terrorism.
4. It is essential to ensure that law enforcement and counter-terrorism agencies
appreciate the potential human rights impacts of AI, as well as the limitations and
fallibility of AI.
o Future work:
- Supporting the design of human rights compliant models for the use of AI for
counter-terrorism together with UNICRI and OHCHR.
- Building good practices for the use of AI for social network analysis for counter-
terrorism together with UNICRI
- Research on potential uses of generative AI for terrorist purposes and mitigation
strategies
- Enhancing Critical Infrastructure Protection through AI against Cyber-Terrorist
Attacks
- Further research and monitoring of the willingness and future ability of terrorists
to use or abuse AI.
• Contact Information: Balques Al Radwan (balques.alradwan@un.org), Akvile Giniotien

(akvile.giniotiene@un.org)

SDG 16
292
3. Relevant Links
https://www.un.org/counterterrorism/cct/programme-projects/cybersecurity
UNOCT
Contact Information
Akvile Giniotiene (akvile.giniotiene@un.org), Balques Al Radwan (balques.alradwan@un.org)
293
United Nations Office for Disarmament Affairs
Project 1: CCW Group of Governmental Experts on Emerging Technologies in the

Area of Lethal Autonomous Weapons Systems
• Project Description: The Office for Disarmament Affairs (ODA) supports the work of the
Convention on Certain Conventional Weapons (CCW) Group of Governmental Experts
on emerging technologies in the area of lethal autonomous weapons systems (LAWS).
The Group has affirmed eleven guiding principles to guide its work, covering, inter alia,
the applicability of international humanitarian law, the retention of human responsibility
and that human-machine interaction should ensure LAWS are used in compliance with
international law. The 2023 Meeting of High Contracting Parties to the CCW renewed
the mandate of the Group, which will work over the next three years to further consider
and formulate a set of elements of an instrument and other possible measures to address
emerging technologies in the area of LAWS.
• Project Type/Output: Intergovernmental Meeting
• Project Domain: Peace and security, legal, humanitarian
Institutions
• Relevant Links and Multimedia: https://meetings.unoda.org/meeting/group-of
-governmental-experts-gge-on-emerging-technologies-in-the-area-of-lethal-autonomous
-weapons-systems-laws/
• Contact Information: Tania Banuelos Mejia (tania.banuelos@un.org), , Juliana Helou van
der Berg (juliana.helou@un.org)
Project 2: Report of the Secretary-General on Developments in Science and

Technology and their Potential Impact on International Security and Disarmament
Efforts
• Project Description: As requested by United Nations General Assembly resolution
77/43, the United Nations Secretary-General reported to the 78th session of the General
Assembly on current developments in science and technology and their potential impact
on international security and disarmament efforts, including on developments related to
AI. This report has been issued as A/78/408.
• Project Domain: Peace and security
• Related Sustainable Development Goals (SDGs): 16 – PEACE, JUSTICE, AND STRONG
294
• Relevant Links and Multimedia: https://www.un.org/disarmament/topics/scienceandt

echnology/
• Contact Information: Rene Holbach (holbach@un.org) ; Beyza Unal (beyza.unal@un.org)
UNODA
Project 3: Responsible Innovation in AI for Peace and Security
• Project Description: The Office for Disarmament Affairs (ODA) and the Stockholm
International Peace Research Institute (SIPRI) have partnered for a three-year initiative on
responsible innovation in artificial intelligence (AI) for peace and security. This project,
which is funded by a decision of the Council of the European Union (Council Decision
(CFSP) 2022/2269 of 18 November 2022), aims to support greater engagement from
the civilian AI community in mitigating the risks that the misuse of civilian AI technology
can pose to international peace and security. Combining awareness raising and capacity
building activities, it seeks to provide the civilian AI community – especially the next
generation of AI practitioners – with the necessary knowledge and means to engage in
responsible innovation and help ensure the peaceful application of civilian AI technology.
The project covers three work packages that:
1. engage with students – introduce selected Science, Technology, Engineering and

Mathematics (STEM) students from around the world to how the peace and security
risks posed by the diversion and misuse of civilian AI development by irresponsible
actors may be identified, prevented or mitigated in the research and innovation
process or through other governance processes. This through a number of channels
including a series of international in-person workshops that centre a diverse group of
young practitioners from around the world;
2. engage with the AI industry – work with practitioners in industry, as well as professional
associations and standards bodies like the Institute of Electrical and Electronics
Engineers (IEEE) to i) disseminate tailored education materials and engagement
activities to technical professionals; ii) support positive uses of AI for peace and
security; and iii) facilitate dialogue and information sharing between experts from
academia, the private sector and government on how the risk of the diversion and
misuse of civilian AI research and innovation by irresponsible actors can be mitigated;
and
3. engage with educators– work with selected educators and developers of academic
curricula on the development and promotion of educational materials that can be
used to mainstream consideration of the peace and security risks that flow from the
diversion and misuse of civilian AI research and innovation by irresponsible actors
in the training of future AI practitioners (e.g. in courses on AI ethics and responsible
innovation).
Such an approach allows the project to reach the AI community at all levels, including
not only current practitioners but also future generations. It also enables engagement
across academic, industry and other silos, and supports the sustainability of future efforts
by establishing networks that cross these boundaries. It employs the convening power
and experience of UNODA and SIPRI to impact the AI community globally, facilitating
engagement between AI actors from across Africa, Asia-Pacific, Europe, and North
and South America. In-person, virtual and hybrid capacity building and dissemination
elements introduce participants not-traditionally exposed to disarmament issues to the
key concepts in responsible innovation of AI, and fill a gap in existing efforts to promote
responsible development and use of AI, the majority of which pay little to no attention to
the impact of AI research and innovation on the military sphere, or international peace
and security. It also allows for participants to recognize more granular issues, like bias
and inclusion, and consider their implications for peace, security, and disarmament. In
addition to SDGs 4, 5 and 16, the project responds to action 28 of the Secretary-General’s
Disarmament Agenda.
• Project Type/Output: Capacity building and Outreach
295

Gender Equality; SDG 16 – Peace, Justice, and Strong Institutions.
• Partnership(s)/Collaborator(s)(s): Stockholm International Peace Research Institute (SIPRI),
Federal University of Rio Grande do Sul, Malmö University New York University, Sorbonne
University, Tallinn University of Technology Umeå University, Institute of Electrical and
Electronics Engineers (IEEE) (made possible thanks to generous support from the
European Union)
• Relevant Links and Multimedia: https://disarmament.unoda.org/responsible-innovation
-ai/about
• Contact Information: Charles Ovink (charles.ovink@un.org) ; Gaston Collazo (collazog@
un.org)
Project 4: Outreach on Autonomous Weapons and Artificial Intelligence

• Project Description: The Secretary-General, the High Representative and Under-Secretary-
General for Disarmament Affairs and other ODA officials have sought to raise awareness
of the possible implications of autonomous weapons and the weaponization of artificial
intelligence.
• Project Type/Output: Outreach
Institutions.
• Relevant Links and Multimedia: https://www.un.org/disarmament/hrstatement/
• Contact Information: Beyza Unal (beyza.unal@un.org), Charles Ovink (charles.ovink@un
.org), Tania Banuelos Mejia (tania.banuelos@un.org), Juliana Helou van der Berg (juliana.
helou@un.org),
Project 5: Governance of Artificial Intelligence in the Military Domain

• Project Description: The Office for Disarmament Affairs and the Republic of Korea held a
project on governance of artificial intelligence in the military domain. The project provided
an opportunity for diplomatic community to come together with the expert community,
including academia, civil society, and private sector. The project involved exchanging
views around responsible design, development, and use of artificial intelligence in the
military domain, assessed the impact of artificial intelligence on the future of warfare, and
provided potential recommendations in the governance realm.
• Project Type/Output: Capacity-building and Outreach
• Project Domain: Peace and Security
• Related Sustainable Development Goals (SDGs): SDG 16 – Peace, Justice and Strong
Institutions
• Contact Information: Beyza Unal (beyza.unal@un.org) ; Aaron Junhoung Yoo (yoo2@un
.org)
296
Project 6: Artificial Intelligence Tipping Points in the Context of International Peace

and Security
UNODA
• Project Description: Advancements in artificial intelligence and their potential to cause
extreme impact events, either due to their integration into weapons systems or due
to their potential to reach and/or suppress human level intelligence, require specific
attention for multilateral governance. Another area of concern for extreme risks is the
increased resource needs (i.e., energy, minerals, semi-conductors, data, water) and the
potential direct and indirect impacts of AI on the planet. This project aims to examine the
extreme risks artificial intelligence pose in the areas of international peace and security
and explore ways to mitigate such risks by promoting multilateral governance at the
auspices of the United Nations. As part of this project, ODA is identifying specific tipping
points (such as critical safety and security thresholds) for advanced AI systems, with the
assumption that crossing those tipping points might result in irreversible consequences
for the humanity. The project commenced at the end of 2023.
• Project Type/Output: Research
• Project Domain: Peace and Security
• Related Sustainable Development Goals (SDGs): SDG 16 – Peace, Justice and Strong
Institutions, SDG 13 – Climate Action (*due to links with tipping points in relation to
climate risks)
• Relevant Links and Multimedia: None
• Contact Information: Beyza Unal (beyza.unal@un.org); Ulysse Richard (ulysse.richard1@
un.org)

SDG 4, 5, 16
3. Related Links
https://www.un.org/disarmament/
Contact Information
Beyza Unal (beyza.unal@un.org), Charles Ovink (charles.ovink@un.org), Tania Banuelos Mejia

(tania.banuelos@un.org), Juliana Helou van der Berg (juliana.helou@un.org)
297
United Nations Office on Drugs and Crime
Project 1: Improving the monitoring of illicit crop cultivation and drug production
by using artificial intelligence
• Project Description: Jointly with the main drug-growing countries in the world - Colombia,
Peru and the Plurinational State of Bolivia for coca, Afghanistan, Lao PDR, Mexico and
Myanmar for opium – the UNODC Illicit Crop Monitoring Programme (ICMP) uses GIS and
geospatial analysis, satellite imagery and field surveys to monitor the extent and evolution
of illicit crop cultivation and production, as well as the factors driving illicit cultivation. The
crop and socio-economic surveys help Governments in their policy development and in
planning how to tackle illicit drug production.
UNODC cooperates with external partners from academia and other research entities to
continuously improve and develop the methods used in the surveys.
The present project seeks to research and eventually apply (semi-) automated methods
such as deep learning and big data analysis for improving area estimates for illicit crop
cultivation, more specifically the identification of illicit crops and spatial information on
potential agricultural land and potential risk areas.. Moreover, research is conducted
on spectral based yield information, and the early detection of illegal landing runways
applying AI techniques. In the Amazon rainforest, illegally built runways facilitate drug
trafficking and harm the environment.
• Department/Division (if applicable):
• Project Status: Ongoing: Pilot activities have been conducted.
• Project Domain: Agriculture and land use (Plant based illicit drug production, such as the
cultivation of opium poppy and coca bush)
Institutions
• Partnership(s)/Collaborator(s): University of Salzburg, Austria, Digital FAO and Agro-
Informatics Division (CSI);
• Contact information: Lorenzo Vita (Lorenzo.vita@un.org)
298
Project 2: Drugs Monitoring Platform
UNOV/UNODC
• Project Description: The UNODC Drugs Monitoring Platform (DMP) is a unique geo-
coded system that brings together information for collecting, visualising, and sharing
drug seizure data aimed at providing access to near real-time data and insights on drug
trafficking trends. It delivers data using interactive analytical visualisations adapted to user-
specific needs to improve drug threat identification for law enforcement and analysts. The
platform translates qualitative information on individual seizure cases into event-based
data records where single seizures are geolocated and described with standardised
variables, easy to use for analysis of trafficking patterns and improved understanding
of illicit drug markets. The Platform relies on the integration of multi-sourced datasets
stemming from data officially reported to UNODC by Members States and open data and
information from cooperating partners.These data require rigorous deduplication and
quality control procedures prior to data modelling and visualisation.The DMP provides
additional geographical insight on current drug trafficking trends, with approximately
656,000 geo-coded drug seizure data points obtained from countries around the world.
More than ever, there is a critical need to enhance capacities to produce rapid pictures of
threats related to drug trafficking and build better analytical pictures to guide operational
responses and support the development of evidence-based policy. One component of
the DMP project aims to collect and enhance capacities to collect real-time information
through targeted text mining/text analytics applied to data harvested from the internet.
Automation is critical for analysing text-based data efficiently to address the vast quantity
of unstructured data that is generated on a daily basis. For UNODC to process large
quantities of critical information ‘harvested’ in the form of external content, a combination
of Artificial Intelligence (AI) procedures centred around Machine Learning modelling,
the streamlining of data ETL (extract, transform, load) processes together with the
implementation of MLOps have been deployed to ensure the delivery of over 18,000
real-time quantitative data points annually. A series of over 10 newly optimised analytical
dashboards were launched to enhance the DMP user experience and optimise Platform
geospatial content.
• Project Type/Output: Report, Dataset, Software tool (There are a combination of products
for this project which includes a dataset as well as several analytical briefs and software
tool development.)
• AI Approach: Models based on transformers architecture
• Project Start Year: October 2019
• Project End Year: December 2025
• Project Domain: Drug Trafficking trends
• Data Source: webscraped open data on individual drug seizure events from media sites
and official government websites. Initiating the use of information from select social
media sources.
• Publicly Available Data: Yes (The first ever public environment was launched within the
Platform containing a subset of Individual Drug Seizure (IDS) data shared by UN Member
States. All other seizure data content is housed within a closed environment requiring
login credentials).
• Technology/Platform: Python (transformers, spacy), Microsoft Azure Cognitive Services
and other functions, PowerBI, MLOps, CosmosDB.
SDG 5 – Gender Equality; SDG 16 – Peace, Justice, and Strong Institutions
• Links and Multimedia: dmp.unodc.org
299
o Improved more timely response between seizure events and data dissemination
through the DMP.
o Increased geographic coverage and timeliness of seizure data.
o Living process of testing and development. (active learning of machine learning
models required). Models are fine tuned based on a rigorous quality control and
feedback process.
o Not all open data is scrapable, due to data privacy issues.
o Integration of multi-sourced datasets (i.e. open data, officially reported by Member
States to UNODC and other data sharing partners) requires rigorous deduplication
and quality control procedures.
o Additional models developed to monitor trends related to other forms of illicit
trafficking, specifically on tobacco and firearms with updated ETL pipelines.
• Contact information: Francesca Massanello (Francesca.massanello@un.org)
Project 3: Earth Observation and Open Source Intelligence for drugs and crime
analysis
• Project Description: UNODC collaborates with a consortium comprising European Space
Agency, e-GEOS, Dhiria, GAF, Janes and Hensoldt. The partnership (EO4SECURITY)
focuses on developing services for UNODC to investigate environmental crimes and
illicit trafficking. Utilising AI techniques, the consortium processes Earth Observation and
Open-Source Intelligence (OSINT) data gathered from social media, traditional media
and others. The goal is to conduct research on illegal activities like drug production, drug
trafficking and crimes against the environment.
• Department/Division: Division for Policy Analysis and Public Affairs (DPA)
• Project Type/Output: Analytical report, maps, tools and methods
• Project Domain: Drugs and crime
• Data Source: Open-source data and satellite images from various sensors, including
Synthetic-Aperture Radar.
• Data Publicly Available: yes/no, open-source data and commercial satellite imagery.
• Technology/Platform: various
o International organisation: European Space Agency

o Private Sector: e-geos, Dhiria, GAF, Janes and Hensoldt
o EO4Security: from Space, an aid against Environmental Crimes - e-GEOS
• Contact Information: Coen Bussink (coen.bussink@un.org)

SDG 3, 5, 16
300
3. Relevant Links
UNOV/UNODC
https://www.unodc.org/
Contact Information
Peter Erhart (erhart@un.org), Angela Me (angela.me@un.org)
301
United Nations Office for Outer Space Affairs
Project 1: Access to Space for All Initiative and Artificial Intelligence

• Project Description: Access to Space for All is a joint initiative of UNOOSA and space
agencies, research institutions and industry to offer access to space research facilities,
infrastructure and information with the aim of developing technical know-how,
engineering processes and infrastructure in the areas of hypergravity and microgravity,
satellite development and space exploration and promote international cooperation in
the peaceful uses of outer space.
Space technologies, data and applications are key enablers for development, in the
same way access to internet is an enabler. Access to Space for All provides access to
information, educational resources, tools and research infrastructure and facilities thanks
to international collaboration.
The partners of the Initiative are space agencies, research institutions and private
companies.
In the framework of the Initiative, the Office organized a webinar covering the use of
artificial intelligence applications for space technology development. It aimed at providing
applicants to the Access to Space for All hands-on opportunities an overview of what
technologies can be incorporated in their projects. The webinar covered hardware and
software elements. The webinar had speakers from NVIDIA, IBM and the European Space
Agency as well as UNOOSA staff. The Office plans to have a follow up of the webinar this
year.
We have continued to work with NVIDIA on artificial intelligence matters and, as part
of this cooperation, UNOOSA was invited to present Access to Space for All during the
Graphics Processing Unit Technology Conference April 2021 (GTC April 2021).
The objective is to raise awareness about how artificial intelligence can be integrated in
space-technology projects.
• Project Domain: Outer Space Technology
• Technology/Platform: Webinar conducted in MS Teams and GTC conducted in a separate
platform.
• Reported as part of 2022 Compendium on UN AI Activities?: Yes
Industry, Innovation and Infrastructure; SDG 17 – Partnership for the Goals
o Private Sector: Nvidia Corporation, International Business Machines (IBM)
302
o Other: China Manned Space Agency, European Space Agency, German Aerospace
Center, Japan Aerospace Exploration Agency, Center of Applied Space Technology
and Microgravity, Keldysh Institute of Applied Mathematics (Russian Academy of
UNOOSA
Science), Kyutech Institute of Technology, Airbus Defence and Space, Avio and Sierra
Space.
o https://www.unoosa.org/oosa/en/ourwork/access2space4all/index.html
o Access to Space for All and Artificial Intelligence
• Lessons Learned: There webinar was well attended and the participants were very much
engaged, demonstrating the interest for the combination of these two topics.
• Contact Information: Jorge Del Rio Vera (jorge.delriovera@un.org)
Project 2: AI and Climate Action Curriculum

• Project Description: The project consists on the development of a curriculum module on
the utilization of space-based datasets for the development of AI applications on climate
action, targeting girls in developing countries.
• Project Type/Output: Curriculum module and capacity building implementation
• Project Domain: Outer Space and Sustainable Development Goals (SDGs)
• Project Updates: A partnership with Technovation (https://technovation.org/)))) was
launched in 2021 and several webinars presenting the role of space activities for climate
action were organized. Including a Diversity and Inclusion event during the International
Joint Conference on Artificial Intelligence (IJCAI 2021).
Gender Equalit;, SDG 6 – Clean Water and Sanitation; SDG 13 – Climate Action; SDG
15 – Life on Land; SDG 17 – Partnerships for the Goals
o Civil Society: Technovation, implementing through which mentors and girls will use
space data and tools to address the SDGs
o https://technovation.org/
o https://youtu.be/GbX5QgHQSEQ
o https://w ww. technovation. org/n ews- events/i jcai- 2021- tech- for- sustainable
-development/
• Lessons Learned: The work is in progress, working to attract women (age 8-18) to STEM
careers through the incorporation of space related concepts in the curriculum that is run
by Technovation each year. The use of satellite remote sensing data created excitement
and engagement in the audiences that participated in the various events organized.
• The use of space data requires dedicated expertise and time is needed to incorporate
the expertise and knowledge into and accessible curriculum.
• Contact Information: Jorge Del Rio Vera (jorge.delriovera@un.org)
303
Project 3: AI and Climate Action Curriculum

• Project Description: In its resolution 61/110 of 14 December 2006 the United Nations
General Assembly agreed to establish the "United Nations Platform for Space-based
Information for Disaster Management and Emergency Response - UN-SPIDER" as a
new United Nations programme, with the following mission statement: "Ensure that
all countries and international and regional organizations have access to and develop
the capacity to use all types of space-based information to support the full disaster
management cycle".
A number of initiatives in recent years have contributed in making space technologies
available for humanitarian aid and emergency response. Yet, UN-SPIDER is the first to
focus on the need to ensure access to and use of such technologies during all phases
of the disaster management cycle, including the risk reduction phase which is crucial for
reducing the losses of lives and property.
UN-SPIDER works in delivering resources for Member States to facilitate the acquisition
and processing of space remote sensing data.
• Project Domain: Agriculture, Environment, Weather, Space, Disaster Management,
Emergency Response
• Data Source:
o Sentinels, Aqua/Terra, Landsat and other satellite data,

o Composite data from other sources developed by the space community (drought
indices like NDVI, etc).
o Additional data, products and services established by the space and geospatial
communities
o See “Data Sources”: https://un-spider.org/links-and-resources/data-sources

o Python (Jupyter notebooks) and R Studio SNAP, EsriGIS, Quantum GIS, etc

• Project Updates: Continuing the compilation of technical content on data sources, services
and products developed by the space community to support disaster management efforts.
Continuing to present information on activities carried out by the space and the disaster
management communities and technical information on how space-based technologies
contribute to efforts in all phases of the disaster management cycle. Continuing with
the creation of step-by-step workflows in GIS software, Python (Jupyter notebooks) and
R scripts to download, process and visualize Earth observation data for monitoring and
assessing droughts, floods, mudslides, burn severity after forest fires
Communities; SDG 13 – Climate Action
o Other: UN-SPIDER Regional Support Offices https://www.un-spider.org/network/

regional-support-offices
• Relevant Links and Multimedia: https://www.un-spider.org/

• Lessons Learned: The goal of this project is to provide the knowledge, tools, information,
and assessments in the hands of those who can action them for the disaster management
cycle including emergency response.
304
• Since its launch at the end of 2009, it has been gaining recognition as a one-stop-shop:
o to acquire information on what can be done with space technologies to support efforts
UNOOSA
in all phases of the disaster management cycle, and on the mechanisms, products
and services established by the space community to support efforts in all phases of
the disaster management cycle,
o to gain access to data, information and products developed and made available by
the space and geospatial communities, information on software packages and tools
that can be used to process data and to generate relevant products or information,
o To gain access to step-by-step procedures developed in open software to process
data to generate maps that are relevant in disaster management applications,
o To find information on upcoming training opportunities (virtual, presential, academic)
o Nearly a million visits to the tool on an annual basis in the last two years (2020 and
2021)
• Contact Information: Juan Carlos Villagran de Leon (juan-carlos.villagran@un.org)

SDG 4, 5, 6, 11, 13, 15, 17
3. Relevant Links
https://www.unoosa.org/
Contact Information
Jorge Del Rio Vera, Scientific Affairs Officer (Space Technology) (jorge.delriovera@un.org);
Markus Woltran (Markus.woltran@un.org)
305
United Nations Research Institute for Social Development
UNRISD has completed a think-piece series on new technology and human rights (More
details here)
Project: Think Piece Series
UNRISD has completed a Think Piece Series which invited experts from academia, think tanks
and civil society to engage with the topic of linking technology and human rights, and to share
their experience at the front lines of policy-driven research and advocacy aimed at leaving no
one behind in an increasingly digital, automated world.
This Series aimed to provide perspectives on the intersections between new technology and
various dimensions of civil and political rights and economic, social and cultural rights, including
the right to health, work, social protection, freedom of expression and more. It also presents
reflections on how we conceptualize and practice human rights in the face of technology-driven
change on a global scale.
The Series was launched to coincide with the 37th Session of the UN Human Rights Council,
as part of UNRISD’s commitment to promote socially just and sustainable development within
and beyond the UN system. It is also part of the UN system's celebration of the 70th anniversary
of the Universal Declaration of Human Rights.
• First Edition: From Disruption to Transformation
o Tech for Transformative Change? Looking beyond Disruption—Kelly Stetter

o Time for a Fourth Generation of Human Rights?—Changrok Soh, Daniel Connolly and
Seunghyun Nam
o Embracing Human Diversity: Policies and Enabling Factors for Accessible
Technologies—Alejandro Moledo
o Data Frameworks for a Right to Development—Anita Gurumurthy and Nandini Chami
o Big Data and Monitoring Sustainable Development Goal 3: Not Counting Those Left
Behind?—Carmel Williams
o Accounting for the Most Vulnerable: Ensuring Big Data Works for Sustainable and
Inclusive Development—Sabrina Rau and Sheldon Leader
o How IT Threatens Democracy—Kofi Annan
o Technology and Freedom of Expression: Opportunities and Threats through the
Journalist’s Lens—Mariateresa Garrido
o A Feminist Interrogation of Autonomy on the Internet—Jac sm Kee
306
• Second Edition: Tools for Transformation
The second edition of this think piece series on new technologies and human rights
UNRISD
focuses more on responses and possible solutions to issues sketched out in the first
edition. The authors were speakers at our official side event of the 39th session of the
United Nations Human Rights Council on new technologies and human rights held in
September 2018.
o Profiling and Automated Decision Making: Is Artificial Intelligence Violating Your Right
to Privacy?—Tomaso Falchetta
o Legal Literacy: An Essential Complement to Digital and Scientific Literacy—Thérèse
Murphy
o Human Rights and New Technologies: Setting the Agenda for Human Rights-Centred
Innovation—Molly K. Land
UNRISD held an event on new technologies and human rights, co-sponsored by Austria and
Denmark, at the 39th session of the UN Human Rights Council. More detail here.
2. Challenges and Opportunities

Great interest in the topic, but difficult to convert into solid funding for holistic and critical
research enquiries.
3. Relevant Links
Contact Information
Paul Ladd, Director (paul.ladd@un.org)
307
United Nations University
Project 1: UNU-EGOV and ISSA joint webinar series on Artificial Intelligence dedicated
to the social security agencies and government services
• Project Description: This is part of a series of webinars on Artificial Intelligence
collaboration between UNU-EGOV and International Social Security Agency (ISSA)
University with whom we are also preparing a report on the use of Artificial Intelligence
in Social Security.
The plan for the webinar is to have an initial 20-minute presentation to researchers and
academics as a framing “academic” presentation. Then multiple practitioners from well-
known international agencies, technology focus on their experience and approaches
in implementing AI and data driven interventions, it can also be of about 20 minutes
Finally, we will have a 30-minute Questions and Answers discussion session. The audience
is very varied. It will very likely attract CIOs and ICT managers from social security
institutions wanting to understand more about what various issues on AI ranging from
AI in healthcare, Conversational Agents, Explainability, Risk, Information protection. On
average about 200-300 people remain online during the webinar, mostly concentrated
in Europe, Africa and Americas and some from Asia region. The webinars are recorded
for later viewing for those who could not connect on the spot.
• Division/Department: UNU EGOV under general ISSA - UNU-EGOV partnership on digital
transformation and capacity development
• Project Type/Output:
• Project Domain: Social Security Agencies, Government Agencies
• Data Source: N/A
• Data Publicly Available: N/A
• Technology/Platform: Webinar via zoom
• Related Sustainable Development Goals (SDGs): SDG 1-12, 16, 17
• Relevant Links and Multimedia: A few recent links
o https://www.issa.int/events/webinar-2024-01-18
o https://www.issa.int/events/webinar-2023-06-29
o https://www.issa.org/event/risk-management-in-the-age-of-ai/
• Contact Information: Moinul Zaber (zaber@unu.edu)
308
Project 2: Understanding the urban space for better governance: use of non-
traditional data for real-time disaggregated decision-making
• Project Description: Local governments around the world need well-curated data on
UNU
the urban spaces to monitor their inhabitants and to understand the impact of policy
interventions. Urbanization plays a critical role in changing the urban environment. Most
developed countries have almost completed urbanization. However, with more and more
people moving to cities, the urban environment in developing countries is undergoing
significant changes. Sustainable development cannot be achieved without significant
changes in building, managing, and responding to changes in the urban environment.
The scarcity of well-curated Spatio-temporal data of urban spaces gives the AI engines
the fuel to build machine learning and data science tools for the policymakers to make
decisions based on evidence of the present. Understanding the issues, visualizing
the challenges, and monitoring progress are keys to achieving SDG goals. However,
collecting traditional data from urban spaces is expensive and therefore not easily
replicable. Hence, by the time the data are prepared, the reality of the decision-making
space changes. Moreover, data collected in most cases are not usable for decision-
making. With the advent of computational capacity, and advances in knowledge streams
such as machine learning, data mining, and statistical inference it is possible to harness
data from heterogeneous sources, organize them in conjunction with the traditional data,
and visualize them in various ways. One such data comes in abundance from the satellites.
These data help monitor the urban spaces' changes in real-time. This project uses satellite
data and computational mechanisms to understand the socio-economic condition of
urban spaces.
The primary focus of this project is the cities in the developing world. The first part of
the project is focused on developing a novel method to classify urban spaces based on
the buildings and their surroundings. The novel method is designed to help prepare the
datasets for state-of-the-art deep learning mechanisms. The second part of the project
focuses on training and designing a novel deep learning mechanism that is suitable for
the urban categorization process. The resulting automated method can detect the socio-
economic condition of urban spaces of the cities in the developing world by detecting
highly formal to highly informal zones with very high accuracy. The ease in data curation
and scalability make the model useful for city planners and policymakers in the developing
world at almost no cost compared with traditional survey-based methods. The third part
of the project is now focusing on including various other data features such as road
conditions, green spaces, urban air and temperature, water bodies, etc. to build more
robust categories.
• Department/Division: United Nations University Operating Unit on Policy-Driven
Electronic Governance (UNU-EGOV)
• Project Type/Output: Academic paper/ Dataset/Policy Framework / Software tool
• Project Domain: Sustainable cities and communities
• Data Source: Satellite, Census, and survey data
• Technology/Platform: R, Python, PHP, JavaScript, FCN-8, U-Net, DeepLabv3+
• Related SDGs: SDG 11 – Sustainable cities and communities; SDG 10 – Reduced
Inequality; SDG 8 – Decent Work and Economic growth; SDG 1- No Poverty.
Academia: University of Tokyo, Japan; University of Dhaka, Bangladesh; Independent
University of Bangladesh, Bangladesh.
309
• Lessons Learned: Satellite data with usable socio-economic categorization methods

and computational models can help categorize urban spaces in real time. This can
complement the traditional survey and census-based data sources that are expensive and
time-consuming. The categorization method helps categorize the city spaces into sixteen
subcategories based on building conditions and surrounding environments. The sub-
categories are then clustered to find four socioeconomic categories ranging from highly
formal to highly informal spaces. After training the deep learning module detects these
spaces with over 90% accuracy. The final model can segment the test part with an average
accuracy of 90.0% for Dhaka, 91.5% for Nairobi, 94.75% for Jakarta, 82.0% for Guangzhou
city, 94.25% for Mumbai, 91.75% for Cairo, and 96.75% for Lima. The methods are novel,
rapid, and scalable for the public policy practitioners relying mostly on traditional survey
and census data for their decision-making purposes. Multiple publications to understand
various aspects of urban environment such as movement, energy use and habitat were
also published that indicated the importance of heterogenous data and machine learning
for public policy.
• Links:
o https://www.mdpi.com/1424-8220/21/22/7469
o https://www.mdpi.com/2071-1050/14/7/4336[1][1][1][1]
o Understanding the Urban Environment from Satellite Images with New Classification
Method—Focusing on Formality and Informality. Sustainability 2022, 14, 4336. https://
doi.org/10.3390/su14074336 https://www.mdpi.com/2071-1050/14/7/4336
o Applying State-of-the-Art Deep-Learning Methods to Classify Urban Cities of the
Developing World. Sensors 2021, 21, 7469. MDPI publications, https://doi.org/10
.3390/s21227469 https://www.mdpi.com/1424-8220/21/22/7469
o Seasonal, Temporal and Spatial Variation of Particulate Matter Concentration in
Bangladesh: A Longitudinal Analysis, 2021 IEEE Region 10 Symposium (TENSYMP),
Jeju, Korea, Republic of, 2021, pp. 1-8, doi: 10.1109/TENSYMP52854.2021.9550877.
o "A Novel Disaster Image Data-set and Characteristics Analysis using Attention Model,"
in 2020 25th International Conference on Pattern Recognition (ICPR), Milan, Italy,
2021 doi: 10.1109/ICPR48806.2021.9412504.Cheng, Q.; Zaber, M.; Rahman, A.M.;
Zhang, H.; Guo, Z.; Okabe, A.; Shibasaki, R. Understanding the Urban Environment
from Satellite Images with New Classification Method—Focusing on Formality and
Informality. Sustainability 2022, 14, 4336. https://doi.org/10.3390/su14074336
o Rahman, A., Zaber, M., Cheng, Q., Nayem, A., Sarker, A., Paul, O., Shibasaki, R.,
Applying State-of-the-Art Deep-Learning Methods to Classify Urban Cities of the
Developing World. Sensors 2021, 21, 7469. MDPI publications, https://doi.org/10
.3390/s21227469
o Cheng, Q., Rahman, A., Sarker, A., Nayem, A., Paul, O., Ali, A., Amin, M., Shibasaki, R.
& Zaber. M., (2021). Deep learning coupled with novel classification method to classify
the urban environment of the developing world. In SIGML 2020. Proceedings of the
2nd International Conference on Signal Processing and Machine Learning , Zurich,
Switzerland, January 23-24 and In Computer Science & Information Technology (CS
& IT) 2021, pp 37-56, DOI : 10.5121/csit2021.110103
o Rizvee R.A., Zaber M. (2021) How Newspapers Portrayed COVID-19. In: Byrski A.,
Czachórski T., Gelenbe E., Grochla K., Murayama Y. (eds) Computer Science Protecting
Human Society Against Epidemics. ANTICOVID 2021. IFIP Advances in Information
and Communication Technology, vol 616. Springer, Cham. https://doi.org/10.1007/
978-3-030-86582-5_5
o A. Zaman, S. B. Rabbani, R. Ridwanul Haque and M. Zaber, "Seasonal, Temporal and
Spatial Variation of Particulate Matter Concentration in Bangladesh: A Longitudinal
Analysis," 2021 IEEE Region 10 Symposium (TENSYMP), 2021, pp. 1-8, doi: 10.1109/
TENSYMP52854.2021.9550877.
o Niloy. A., Arif .M, Nayem. A., Sarker. A., Paul. O., Amin. M., Ali. A., Zaber M. & Rahman. A.
(2021). "A Novel Disaster Image Data-set and Characteristics Analysis using Attention
310
Model," 2020 25th International Conference on Pattern Recognition (ICPR), 2021, pp.
6116-6122, doi: 10.1109/ICPR48806.2021.9412504
o F. Tabassum, H. Islam, A. A. Ali and M. Zaber, "A Complex Network Analysis of
Inland Waterways Port Connectivity of Bangladesh," in IEEE Region 10 Symposium
UNU
(TENSYMP), Dhaka, June, 2020. , 10.1109/TENSYMP50017.2020.9230896
o M. Wahed, R. A. Rizvee, R. R. Haque, A. M. Ali, M. Zaber and A. A. Ali, "What
Can Nighttime Lights Tell Us about Bangladesh?," in IEEE Region 10 Symposium
(TENSYMP), Dhaka, June, 2020. DOI : 10.1109/TENSYMP50017.2020.9230806
o S. B. Rabbani, A. A. Ali and M. Zaber, "Does Electric Prepaid Meters Decrease Payment
Delinquency? Evidence from Data Centric Analysis of Electricity Consumption in
Dhaka, Bangladesh," in IEEE Region 10 Symposium (TENSYMP), Dhaka, June, 2020.
DOI : 10.1109/TENSYMP50017.2020.9230814
• Contact information: Moinul Zaber (zaber@unu.edu)
Project 3: Digital Transformation for the Implementation of BMZ Flagship Projects

on Digitalisation – Lot 3 AI Expertise
• Project Description: Consultancy for GIZ and its partners on the following topics:
a. The application of artificial intelligence (AI) and machine learning (ML) in a sustainable
development context (e.g. in agriculture, mobility, urban development, energy,
financial inclusion and the wider financial system and the public sector), including
the development, deployment and integration of ML systems with other software.
b. Explanations of artificial intelligence and machine learning systems including the
current status and potential of AI and ML that are accessible for a non-expert audience,
in particular in GIZ’s partner countries.
c. Risk assessments of the use of AI and ML in a sustainable development context,
following international best practices and GIZs standards and focussing on issues
including potential discrimination, especially of disadvantages groups, in training
datasets and by AI/ML models, as well as mitigation measures for these identified
risks, including approaches for de-biasing of training data.
d. Sustainability assessments for AI and ML deployments following GIZs the following
three good practice standards 1) Principles for Digital Development 2) criteria for ML
projects and 3) “Self-sustaining Open AI cycle in nine steps”.
e. Analyses of the ecosystems of AI and ML including the mapping and assessment of
stakeholders and analysis of framework conditions in partner countries of German
Development Cooperation, nationally, regionally or globally.
f. Contributions from a technical point of view to policy discussion around artificial
intelligence e.g. data governance, local hosting requirements, data flows etc.
g. Creation of material for and implementation of capacity building measures in GIZ’s
partner countries, especially India, in the cross-section of AI and cyber security,
especially on secure AI systems, the relationship between AI and cyber security and
open AI training data and AI-based cybersecurity.
h. The creation of training datasets for AI and ML in a sustainable development context,
especially the creation of openly available datasets in a manner that ensures the
sustainable maintenance and extension of such datasets.
• Division/Department: UNU-EGOV, United Nations University Operating Unit on Policy-

Driven Electronic Governance
• Project Type/Output: Consulting
311
• Project Domain: Public Administration, e-Government

• Data Source: Participation in calls and other off-site activities like virtual working group
meetings, coaching and mentoring of GIZ’s partners, desk research and drafting of
factsheets and practical reports, on-site consulting
• Partnership(s)/Collaborator(s): GFA Consulting Group GmbH
• Contact Information: Delfina Soares (soares@unu.edu)
Project 4: AI4PA Portugal Artificial Intelligence & Data Science for Public Administration
Portugal Innovation Hub
• Project Description: The AI4PA aims to support the digital transition of public
administration by introducing Artificial Intelligence and other technologies (technically
adequate and socially responsible and oriented towards the common good) to increase
the effectiveness of public policies, as well as capacitating relevant stakeholders at the
central, regional and local public administration and SMEs that offer digital solutions
suited to the needs of the State.
• Project Type/Output: Collaborative network, Testing, Experimentation, Training,
Consulting
• Project Domain: Public Administration, e-Government
• Data Source: public policies
• Related Sustainable Development Goals (SDGs): 11
• Partnership(s)/Collaborator(s): AMA – Agência para a Modernização Administrativa, CVTT-
Iscte - Associação Iscte Conhecimento e Inovação - Centro de Valorização e Transferência
de Tecnologias, AESINTRA - Associação Empresarial de Sintra, AIP – CCI - Associação
Industrial Portuguesa - Câmara de Comércio e Indústria, ANPME - Associação Nacional
Pequenas Médias Empresas, Associação Laboratório Colaborativo para o Trabalho,
Emprego e Proteção Social – LCTEPS (CoLABOR), AUDAX - Centro de Inovação e
Empreendedorismo, Cisco International Limited, Sucursal em Portugal, Comunidade
Intermunicipal do Oeste (OESTECIM), Direção-Geral de Estatísticas da Educação e
Ciência (DGEEC), Esri Portugal - Sistemas e Informação Geográfica, S.A, GEP/MTSSS
- Gabinete de Estratégia e Planeamento, INDEG - Iscte Executive Education, IPPS -
Instituto para as Políticas Públicas e Sociais, MORE CoLAB - Laboratório Colaborativo
Montanhas de Investigação, Município de Sintra, Município de Viseu, UGT -União Geral
de Trabalhadores, UNINOVA – Instituto de Desenvolvimento de Novas Tecnologias
Project 5: Global assessment of responsible AI in cities (AI4Cities)

• Project Description: Provide technical support to develop a thorough assessment of
the opportunities and constraints for local government development to responsibly
implement, use and govern Artificial Intelligence (AI) technologies in cities. The following
activities are included:
o designing and analysing a global survey for cities,
312
o performing desk-based research to capture case studies on the use and governance
of AI in cities,
o drafting a research paper with findings and policy recommendations to address
capacity gaps.
UNU
• Project Type/Output: Consulting
• Project Domain: Local government
• Partnership(s)/Collaborator: International Development Research Centre (IDRC), UN
HABITAT
Project 6: 16th International Conference on Theory and Practice of Electronic

Governance (ICEGOV), TRACK 1: Emerging and disruptive technologies for digital
governance
• Project Description: New technologies emerge continuously, with promises of new
affordances and opportunities for digital governance. Emerging technologies are
still under development and, therefore, not fully formed and matured. These are
often associated with disruption, meaning that, if applied, they carry the potential to
fundamentally change the way digital governance is carried out. Such a change could
alter the nature of the public sector organisations and citizen-government relationships
altogether. Artificial Intelligence, robotisation, data analytics, blockchain, and open
data are only a few examples of technologies that can be considered emerging and
potentially disruptive. This Track invites papers that deal with issues related to emerging
and potentially disruptive technologies for digital governance. The track welcomes
empirical studies of government organisations’ work and experimentation with new
technologies, such as machine learning, algorithmic decision-making, face recognition,
new forms of automation and robotisation, blockchain, and more. The Track also
welcomes conceptually oriented papers that further the field’s understanding of emerging
technologies and disruption for digital governance. Papers that critically discuss the role
of emerging technologies for digital governance and their potential impact on public
sector organisations are particularly welcome.
• Project Type/Output: Conference Session
• Project Domain: Emerging and disruptive technologies for digital governance
• Relevant Links and Multimedia: https://www.icegov.org
313
Project 7: AI capacity needs in cities: insights for strengthening local governments'

AI governance, Roundtable Session
• Project Description: ICEGOV 2023 Roundtable Session
• Project Type/Output: Conference session
• Project Domain: Local government
• Partnership(s)/Collaborator : International Development Research Centre (IDRC), Canada,
UN HABITAT
Project 8: Gender-sensitive AI policy in Southeast Asia

• Project Description: This is a research project conducted in Thailand, Indonesia, Malaysia
and the Philippines, which aims to understand gender and societal risks in AI in the focus
countries, with a focus on critical technologies. The outputs of this research include a
policy report, which makes recommendations to include consideration on gender
discrimination, stereotyping and exclusion in future AI legislation, as well as a training
programme for policymakers in the four countries.
• Division/Department: United Nations University Institute in Macau
• Project Type/Output: Policy research project and training programme
• Project Domain: Women’s rights
• Related Sustainable Development Goals (SDGs): SDG 5: Gender Equality
• Relevant Links and Multimedia: https://unu.edu/macau/blog-post/developing-inclusive
-ai-policy-southeast-asia
• Contact Information: Eleonore Fournier-Tombs (fourniertombs@unu.edu)
Project 9: Gendered implications of Artificial Intelligence on the implementation of

the Women, Peace and Security agenda in Southeast Asia
• Project Description: The mainstreaming of artificial intelligence technologies (AI)
globally has had an important impact on the security of women. While AI can be used
for peacebuilding purposes, it is used unequally across genders, with women much
more impacted than men by the digital divide. At the same time, AI systems have been
shown to pose security risks to women, particularly with relation to online harms (such
as cyberbullying, cyberstalking, doxing, misogynistic hate-speech and other forms of
harassment), dis- and misinformation, and privacy. Finally, as Southeast Asian countries
begin to develop strategies and regulations for AI, they often have not taken into
consideration gender risks in AI, which can be obfuscated under the projected economic
potential of the technologies.
Investment in AI technologies is increasing in the Southeast Asian region, although at
an uneven pace, with some countries positioning themselves as global leaders in AI,
314
while others having a much more limited capacity. It is projected, however, that AI will
add 1 trillion dollars to the GDP of Southeast Asian countries by 2030. In this context,
understanding the impact of these technologies on the Women, Peace and Security
agenda (WPS) is critical to supporting Southeast Asian countries to regulate the
UNU
technologies and mitigate their risks.
This research examined the opportunities and risks of AI from a WPS lens in Southeast
Asia, with a focus on four types of gender biases in AI which will need to be addressed
before the region can fully benefit from new technological developments; discrimination,
stereotyping, exclusion, and insecurity. Understanding these and mitigating them is an
important step in developing a safe and trustworthy AI ecosystem. The research included
interviews with relevant stakeholders to understand the relationship between AI and
WPS as well as a social media analysis of women civil society organisations in the region.
Findings highlight three types of AI and its applications: AI for Peace, Neutral AI, and
AI for Conflict. In each category, there are favourable and unfavourable effects of AI for
gender-responsive peace and women’s agency in peace efforts. These findings were
used to inform an e-learning module on AI and the WPS agenda for women civil society
in Southeast Asia and two critical recommendations are made; mitigating the risks of AI
systems to advancing the WPS agenda and fostering the development of AI tools built
explicitly to support gender-responsive peace in line with WPS commitments.
• Project Type/Output: Policy research project and training programme
• Project Domain: Women’s rights
• Data Source: Primary data collected via interviews
• Related Sustainable Development Goals (SDGs): SDG 5
• Partnership(s)/Collaborator(s): UN Women Regional Office for Asia and the Pacific
• Relevant Links and Multimedia: N/A
• Contact Information: Jaimee Stuart (stuart@unu.edu)
Project 10: Generative AI, Anxiety, and Hope for the Future Among Young Adults
• Project Description: Generative Artificial Intelligence (Gen AI) is rapidly becoming part
of our everyday digital experiences, it is not only embedded in apps in our smartphones
but also in virtual assistants, software, chatbots and within search engines. Research has
shown that young people are already using Gen AI in a variety of ways, and this technology
provides many new and emerging risks and opportunities. For instance, it can be used
as a means of enhancing creativity, with platforms and tools offering the opportunity
to generate artwork, stories, and music. It can also help create personalized learning
systems, tailoring help to young people’s specific learning style and needs. Similarly, for
young people with learning disabilities, Gen AI can enhance the accessibility of digital
platforms, making them easier to use for more individuals. However, Gen AI is not without
its potential risks and research shows that young people are aware of this. Their key
concerns center around Gen AI’s ability to spread misinformation and the potential for
job displacement. Youth also have concerns that society may become overly reliant on
Gen AI technology, the potential privacy concerns it poses and how it can be misused to
cheat within an academic context or by governments and corporations.
Overall, while young people accept that Gen AI is here to stay and will inevitably impact
their future, they may face anxiety in response to concerns over the way AI is changing
315
their future careers and educational opportunities, otherwise known as “AI-anxiety”.

Such impacts highlight the need for greater attention to be given to how generative
AI is impacting young people’s wellbeing, and whether helping them learn how to use
these platforms can prevent their fear and anxiety surrounding it. The aim of this project,
therefore, is to understand young adults' experiences of Gen AI – specifically examining
motivations for using Gen AI technologies, what perceived risks and opportunities youth
believe it poses to their futures, and the association (if any) between perceptions of Gen AI
developmental outcomes. This project is made up of two complementary among young
adults (aged 18 – 25 years old), a survey with youth from a range of countries ranging
from LMIC to HIC, and a series of interviews with young people in the United States. As
such, the research used a mixed methods approach incorporating both quantitative and
qualitative data collection. This approach will enable a deeper understanding of the topic
and more complete evidence.
• Project Type/Output: Policy brief, journal article, report
• Project Domain: Youth wellbeing
• Data Source: Primary data via surveys
• Related Sustainable Development Goals (SDGs): SDG 3, SDG 4
• Contact Information: Jaimee Stuart (stuart@unu.edu)
Project 11: UNU Generative AI Monthly Series of Webinars

• Project Description: An in-depth exploration into ChatGPT and Generative Artificial
Intelligence (AI) is a new monthly series of webinars that, recognizing the rapid
advancements and transformative potential of Artificial Intelligence, aims to cultivate a
comprehensive understanding and foster informed and responsible engagement with
these emerging technologies. United Nations University Institute in Macau (UNU Macau)
will be leading this programme, inviting distinguished scholars and researchers from
the UNU system and an extensive network in academia and broader field, to deliver
presentations on topics related to generative AI. The discussions will encompass a wide
array of themes, from technical advancements to the ethical implications, societal impacts,
and policy considerations surrounding the advanced technology. This series is aiming at
presenting diverse perspectives on how generative AI will impact our collective future on
health, education, environment, climate, UN peacekeeping, and humanitarian work.  
• Division/Department: UNU-IIST Macau
• Project Domain: Education, Responsible AI, Law, Computer Science
316
• Related Sustainable Development Goals (SDGs): SDG 9 - Industry, Innovation and

Infrastructure
• Relevant Links and Multimedia: https://unu.edu/macau/unu-generative-ai-series
UNU
• Contact Information: Serge Stinckwich (stinckwich@unu.edu)
Project 12: UNU-UNESCO-UM Workshop: Ethical AI - Pioneering Progress in the

Asia-Pacific
• Project Description: In the unfolding era of Artificial Intelligence (AI), marked by the
broader Fourth Industrial Revolution (Industry 4.0), intricate questions emerge at the
intersections of ethics, politics, law, and economics. The transformative impact of AI,
particularly generative models like ChatGPT, is automating tasks and generating content,
promising substantial economic gains. Yet, this innovation brings ethical challenges,
from data bias to misinformation. Recognizing the urgency for ethical oversight, UNU-
IIST, in collaboration with UNESCO and the University of Macau, is delighted to host the
international workshop titled "Ethical AI: Pioneering Progress in the Asia-Pacific". This
workshop aims to facilitate multi-stakeholder dialogue, uniting experts, policymakers,
and stakeholders to explore the ethical dimensions of AI and seeks to catalyze collective
efforts toward responsible AI development and deployment in the region.
• Project Type/Output: seminar/meeting
• Project Domain: Ethics of Artificial Intelligence
• Related Sustainable Development Goals (SDGs): SDG 4 - Quality Education, SDG 9 -
Industry, SDG 17 Partnerships for the Goals,
• Relevant Links and Multimedia: https://unu.edu/macau/event/international-workshop
-ethical-ai-pioneering-progress-asia-pacific
Project 13: Baseline assessment of the Artificial Intelligence landscape in Mongolia

• Project Description: The project aims to do a baseline assessment of the AI landscape
in Mongolia to develop AI standards, frameworks, policies, and initiatives at the national
level. Desk research has been be conducted covering existing policies and uses of AI in
Mongolia. Following the AI readiness methodology for the UNESCO Recommendations
for the Ethics of AI (2021), the project develop a pilot AI readiness assessment that will
help Mongolia to develop the complete methodology in the future and prepare for the
national AI roadmap. The work has been conducted by UNESCO and the United Nations
University Institute in Macau to develop a research project analyzing the current state of
AI readiness in Mongolia. The project will inform the Mongolian National commission of
UNESCO.
• Project Type/Output: Policy research project
• Project Status: closed
• Project Domain: Ethics of AI
• Data Source: Primary data collected via interviews
317

• Related Sustainable Development Goals (SDGs): SDG 9 - Industry, Innovation and
Infrastructure, SDG 17 Partnerships for the Goals
Project 14: The Use of Synthetic Data to Train AI Models: Opportunities and Risks
for Sustainable Development
• Project Description: Using synthetic or artificially generated data in training AI algorithms
is a burgeoning practice with significant potential. It can address data scarcity, privacy,
and bias issues and raise concerns about data quality, security, and ethical implications.
This issue is heightened in the Global South, where data scarcity is much more severe than
in the Global North. Synthetic data, therefore, addresses the problem of missing data,
leading, in the best case, to better representation of populations in datasets and more
equitable outcomes. However, we cannot consider synthetic data to be better or even
equivalent to actual data from the physical world. In fact, there are many risks to using
synthetic data, including cybersecurity risks, bias propagation, and simply an increase in
model error. This technology brief proposes recommendations for the responsible use of
synthetic data in AI training and the associated guidelines to regulate the use of synthetic
data.
• Division/Department: UNU Centre, UNU-CPR and UNU-IIST Macau
• Project Type/Output: Policy brief
• Project Domain: All
• Related Sustainable Development Goals (SDGs):
• Relevant Links and Multimedia: https://unu.edu/publication/use-synthetic-data-train-ai
-models-opportunities-and-risks-sustainable-development
Project 15: Building Citizen Science Intelligence for Pandemic Preparedness and
Response: Needs Assessment and Pilot Implementation
• Project Description: No more pandemics — this is the ambitious goal set by the
Independent Panel for Pandemic Preparedness and Response, whose report reviews
the international community’s response to COVID-19 and identifies lessons learned.
COVID-19 has shown that our lack of globally accurate, real-time data on outbreaks is a
matter of life and death. A leading group of scientists and experts, from every part of the
world, must come together urgently to create a new, neutral, and trusted digital system
that can revolutionise how data is gathered and used globally. This project involving
participatory modelling and Artificial Intelligence based simulation, funded by I-DAIR, the
International Digital Health and AI Research Collaborative, is based on the premise that, to
contain the pandemic through accurate, real-time and data-driven measures, it is urgent
to establish collective intelligence capabilities that involve all stakeholders. UNU Macau
is currently involved in a pilot study focusing on Vietnam, Kenya and Brazil and aimed at
318
addressing how to effectively manage patients and resources available in a hospital and
its surrounding communities such that citizens can play an active role in the response and
develop broad disaster resilience in the context of a pandemic crisis. In partnership with
local communities in these three countries, two artefacts has been developed: one role
UNU
playing game and a participatory agent-based simulation.
• Project Type/Output: Policy research project, software development
• Project Domain: Public Health, AI
• Data Source: data collected during FGD and workshops
• Related Sustainable Development Goals (SDGs): SDG 3 - Good Health and Wellbeing
• Relevant Links and Multimedia: https://unu.edu/macau/project/building-citizen-science
-intelligence-pandemic-preparedness-and-response-needs
Project 16: Disinformation and Peacebuilding in Sub-Saharan Africa: Security

Implications of AI-Altered Environments
• Project Description: This research project examines the effects of new AI technologies,
namely generative AI, on peacebuilding and conflict in Sub-Saharan Africa, namely in
DRC, Kenya, and Côté d’Ivoire. In this context, this report aims to further explore the way
in which AI technologies as they currently stand impact peace and conflict, and what
methods might be used to mitigate their adverse effects - through the development
of better tools and the inclusion of peace and conflict considerations in AI governance
frameworks.
• Division/Department: UNU Centre for Policy Research
• Project Type/Output: Report and Seminar
• Project Status: Complete, pending publication
• Project Domain: AI Policy, Disinformation, Peace and Security
• Data Source: Interviews
• Data Publicly Available: Interview summaries to be published in report
• Technology/Platform: NA
• Partnership(s)/Collaborator(s): Interpeace
• Relevant Links and Multimedia: NA
• Contact Information: Eleonore Fournier-Tombs (fourniertombs@unu.edu)
Project 17: Towards a Peacekeeper Diversity Index: Evaluating the Relative

Effectiveness of Peace Support Personnel Gender, Origin, and Status
• Project Description: This study aims to cover a critical methodological gap by testing
the utility of machine learning in identifying potential correlations between UN PSO
demographics and conflict in four case studies: the Central African Republic, the
319
Democratic Republic of Congo, Mali, and South Sudan. Quantitative data relating to the
PSO personnel characteristics was sourced from the UN Peace and Security Data Hub.
The conflict location/intensity information was sourced from the Armed Conflict Location
and Event Data Project (ACLED).
• Division/Department: UNU Centre for Policy Research
• Project Type/Output: Report and Seminar
• Project Status: In progress
• Project Domain: Peacekeeping, Diversity
• Data Source: Peace Security Data Hub
• Technology/Platform: https://psdata.un.org/
• Partnership(s)/Collaborator(s): XCEPT
• Relevant Links and Multimedia: NA
• Contact Information: Eduardo Albrecht (albrecht@unu.edu)

SDG 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17
3. Relevant Link
https://unu.edu/
Contact Information
Serge Stinckwich (stinckwich@unu.edu)
320
Universal Postal Union
UPU
Project 1: State of the Postal Sector Report

• Project Description: Artificial intelligence (AI) is at the forefront of revolutionizing the next
wave of postal services. The "State of the Postal Sector" report, UPU’s flagship publication,
delves into AI's transformative role in fostering enhanced collaboration among postal
sector stakeholders worldwide. Spanning two dedicated chapters, the report explores
the vast, yet to be fully realized, potential of AI within the sector and among UPU member
states. It outlines a strategic roadmap to capitalize on AI, detailing pivotal use cases
that could significantly elevate the efficiency and innovation of postal services at all
development levels.
• Project Type: Report
• Project Status: Published
• Partnership: postal services, regulators and governmental authorities in UPU member
states
• Project Website (link):
o State of the Postal Sector 2023 https://www.upu.int/en/publications/2ipd/state-of-the

-postal-sector-2023
• Contact Information: José Anson (jose.anson@upu.int)
Project 2: Prediction of postal item delivery day on the UPU global track and trace
system
• Project Description: The UPU’s global postal supply chain consists of almost 700,000
postal access points and reaches 95% of the world’s population for the delivery of postal
items; mail, parcels and postal payments.
In the delivery process to the addressee, the provision of reliable information on the
expected delivery date of the postal item is a challenge, especially so for postal operations
in least developing, developing countries and small island developing states.
Using machine learning models trained on postal big data, the UPU’s global track and
trace solution (https://globaltracktrace.ptc.post/ ) which provides international routing
information of the postal item, now includes a prediction engine for the estimated delivery
day. It achieves up to 67% accuracy from posting and up to 87% accuracy on the delivery
leg.
• Project Type: AI Big Data Platform
• Project Status: Active
• Project Domain: Postal logistics
321
• Technology Platform: Hadoop cluster

• Project Website (link): https://globaltracktrace.ptc.post/
• Contact Information: PTC.PostInfo@upu.int, external.relations@upu.int

SDGs 8, 9, 13 and 17
3. Relevant Links
FAQ: https://www.upu.int/en/Contact-us/Postal-shipments
Product: https://globaltracktrace.ptc.post/
State of the Postal Sector 2023 https://www.upu.int/en/publications/2ipd/state-of-the-postal

-sector-2023
322
World Bank Group
WBG
Project 1: Natural Language Processing (NLP) to evaluate the presence of gender

biases in the written decisions of judges
• Project/initiative description: Natural Language Processing (NLP) to evaluate the presence
of gender biases in the written decisions of judges
o Do Judges Favor their Own Ethnicity and Gender? Evidence from Kenya: Leveraging
publicly available information on hundredths of thousands of judicial decisions in
Kenya, the team uses NLP to measure gender attitudes in the language of decisions
and evaluate the presence of gender biases. The paper finds that the written judgments
are on average shorter and less likely to be cited when defendants who are of the same
gender or ethnicity as the judge win their case. This is consistent with in-group biased
decisions being of lower quality. In addition, the findings show that female defendants
are less likely to win the case if the judge exhibits stereotypical or negative attitudes
towards women in their writings.
o Gender Attitudes in the Judiciary: Evidence from U.S. Circuit Courts: this is earlier
work by some of the team members using the same method for the US. The paper
proposes a novel judge-specific measure of gender attitudes based on use of gender-
stereotyped language in the judge’s authored opinions. The authors find that slanted
judges vote more conservatively in gender-related cases. Slant influences interactions
with female colleagues: slanted judges are more likely to reverse lower-court decisions
if the lower-court judge is a woman than a man, are less likely to assign opinions to
female judges, and cite fewer female-authored opinions.
• Department/Division: Development Impact Evaluation (DIME)

• Project type: Research Report
• Project status: completed
• Related SDGs: SDG 16 on Peace, justice, and strong institutions
• Partners: none
• Any publicly available URL links/multimedia: https://openknowledge.worldbank.org/
handle/10986/37108;;;; https://ideas.repec.org/p/cge/wacage/462.html
• Project lead contact: Daniel Chen (dchen9@worldbank.org ); Manuel Ramos-
Maqueda (mramosmaqueda@worldbank.org )
Project 2: Transmitting AI Training: Evidence from Policymakers in Pakistan

• Project/initiative description: Transmitting AI Training: Evidence from Policymakers in
Pakistan
323
o How does AI training impact policymaking? We randomly assign a rigorous “AI for
policy” workshop to deputy ministers in Pakistan and find that deputy ministers shift their
attitudes towards AI and increase funding for digitization, a precursor to AI. During this
randomized evaluation of the training program, we then cross-randomized ministers
to receive AI fairness activism that emphasizes the inescapability of algorithmic bias.
AI fairness activism causes policymakers to report greater costs associated with AI in
policymaking and decrease funding for digitization. Both interventions transmit from
the deputy ministers to their subordinate staff and impact the population. Amid land
record digitization efforts, treated ministers’ jurisdictions reduced delays in handling
land disputes by 33%. AI training increases downstream support for AI in policy
making while AI fairness activism reduces the effect of the training.
• Department/Division: Department/Division: Development Impact Evaluation (DIME)

• Project type: Development Research Report
• Project status: Working paper draft
• Project end year: 2022 (Expected)
• Related SDGs: SDG 16: Peace, Justice and Strong Institutions
• Partners: Pakistan National Institute for Public Policy
• Any publicly available URL links/multimedia: http://users.nber.org/~dlchen/papers/
Transmitting_AI_Training.pdf
• Project lead contact: Daniel Chen (dchen9@worldbank.org ); Sultan Mehmood
(smehmood@nes.ru )
Project 3: Transport Systems Improvement Project (TRANSIP) in Ethiopia using AI

• Project/initiative description: As a subcomponent of the Transport Systems Improvement
Project (TRANSIP) in Ethiopia impact evaluation component, we have developed a video
analysis tool using open-source AI and computer vision algorithms to produce data
on incidents, such as near collisions, that are not normally captured by traditional road
safety data, which tends to focus on accidents and fatalities. Using CCTV footage of over
200,000 minutes from surveillance cameras located around Addis Ababa, we employ AI
to detect different types of objects, such as cars, buses and pedestrians. Furthermore,
we assemble the objects’ trajectories, assign them to location specific movements (such
as jaywalking, east-left-turn, etc.), compute their speed and develop a granular measure
of near collisions between objects. This data will be used to evaluate the road safety
related impact of different road improvements across the city and improve the targeting
of interventions.
• Department/Division: Department/Division: Development Impact Evaluation (DIME) and
the Transport Global Practice Unit
• Project status: ongoing
• Project end year: 2024 (expected)
• Related SDGs: 3.6
• Partners:
• Any publicly available URL links/multimedia:
• Project lead contact: Nino Pkhikidze (npkhikidze@worldbank.org ), Girija Borker (gborker@
worldbank.org )
324
Project 4: Using AI to Analyze Road Traffic Crash Data in Kenya

• Project/initiative description: Data on road traffic crashes is lacking in low-income countries
with official data estimated to capture only 17% of road traffic deaths. Missing data largely
WBG
results from poorly developed administrative data systems that often rely on paper
records where data is not up to date. This project investigates whether data generated
from smartphones and social media usage can help fill gaps in administrative records.
Specifically, the project creates road traffic crash location data from crowdsourced crash
reports posted on Twitter in Nairobi, Kenya. The project scraped 874,588 traffic related
tweets from Nairobi and applied a machine learning model to capture the occurrence
of a crash and developed an improved geoparsing algorithm to identify its location. We
geolocate 32,991 crash reports from Twitter for 2012-2020, clustering them into 22,872
unique crashes. For a subset of crashes reported on Twitter, a motorcycle delivery service
was dispatched in real-time to verify the crash and its location; the results show 92%
accuracy.
• Project status: completed
• Partners: MIT Civic Data Design Lab
o [Academic Paper] https://journals.plos.org/plosone/article?id=1

0.1371/journal.pone
.0244317
o [Brief] https://documents1.worldbank.org/curated/en/609971608546546068/pdf/
Urban-Planning-I-Spy-Crashes.pdf
o [Blog] https://b logs. worldbank. org/o pendata/f inding- missing- data- creating
-actionable-information-solving-development-problems
• Project lead contact: Sveta Milusheva (smilusheva@worldbank.org ); Guadalupe Bedoya

(gbedoya@worldbank.org ) ; Arianna Legovini (alegovini@worldbank.org )
Project 5 : Predictive Performance of household poverty models using Machine

Learning
• Project/initiative description: Household surveys give a precise estimate of poverty;
however, surveys are costly and can only be fielded infrequently. This project aims at
comparing the predictive performance of models based on globally available, spatially
referenced public and private sector data sources that have been used to estimate
poverty. We include daytime and nighttime satellite imagery, Facebook marketing
data, OpenStreetMap data, among other sources. The project trains a machine learning
model to predict levels and changes in poverty relying on ground truth poverty data
across 82,000 villages and 59 countries, spanning Africa, Asia, the Americas, and Europe.
Globally, the model explains over 60% of the variation of an asset-based poverty index
at the village level and over 70% of the variation at the district level; in some countries,
the model explains over 90% of the variation in poverty at the district level. Features
from OpenStreetMaps, nighttime lights, and daytime imagery are most important in
explaining poverty, where some features from Facebook Marketing data—such as the
proportion of active Facebook users with interests in restaurants and luxury goods—are
highly (negatively) correlated with poverty across most countries. Accuracy for predicting
changes in poverty is lower, but the model explains above 25% of the variation in poverty
in some countries. The model performs best in lower income countries and in countries
with more variation in levels/changes in poverty.
325

• Project typeDevelopment Research Reports
• Project status: ongoing
• Project end year: 2022 (expected)
• Partners (internal and external to WBG)
• Any publicly available URL links/multimedia
• Project lead contact: Rob Marty (rmarty@worldbank.org); Alice Duhaut (aduhaut@
worldbank.org)
Project 6: Fostering Long-term Savings linked to the Zambia Financial Inclusion

Country Support Program
• Project/initiative description: The project used a novel AI-based text-messaging-based
intervention to: i) identify the behavioral barriers that lead to low engagement with formal
financial services amongst those using the services; and ii) test strategies to help people
overcome those barriers to increase engagement and financial security. Additionally, it
provides the first evidence (that the team is aware of) on the impact of conversational
machine-learning based, two-way text messaging designed to encourage savings and
improve loan repayment behaviors through Q&A capabilities and efforts to enhance trust
in formal financial products.
• Department/Division: Finance, Competitiveness, and Innovation Practice Unit and
Development Economics Research Unit
• Project type: Impact Evaluation Report
• Project status: Completed. The academic paper is in the process of being finalized.
• Related SDGs: The project had a strong focus on financial inclusion which is positioned
prominently as an enabler of other developmental goals in the 2030 Sustainable
Development Goals, where it is featured as a target in eight of the seventeen goals. These
include:
o SDG1, on eradicating poverty;

o SDG 2 on ending hunger, achieving food security and promoting sustainable
agriculture; SDG 3 on profiting health and well-being;
o SDG 5 on achieving gender equality and economic empowerment of women; SDG
8 on promoting economic growth and jobs;
o SDG 9 on supporting industry, innovation, and infrastructure; and SDG 10 on reducing
inequality.
o Additionally, in SDG 17 on strengthening the means of implementation there is an
implicit role for greater financial inclusion through greater savings mobilization for
investment and consumption that can spur growth.
• Partners (internal and external to WBG): Natsave, Junto Finanzas (acquired by Nubank),
Trinity College Dublin, Swarthmore
• Project lead contact: Siegfried Zottel (szottel@worldbank.org )
326
Project 7: Croatia Business Environment Reform with applied AI

• Project/initiative description: The World Bank has been supporting the Ministry of
Economy and Sustainable Development (MoESD) of Croatia in digitalizing Government
WBG
to Business (G2B) service delivery for business registration and business licensing.
This advisory work is carried out by the Bank at the request of the Government, with
funding by the European Union, and in cooperation with the European Commission's
DG REFORM. The work involved piloting an innovative approach for mapping business
administrative procedures that involved the application of AI/Machine Learning (ML) and
Natural Language Processing (NLP). The developed digital algorithm processed 110 laws
and 1,204 bylaws from the registry of regulations to identify more than 9,000 potential
regulatory requirements in 33 administrative areas, including business authorizations
(permits, licenses, etc.), and minimum conditions. Further, the data cleansing resulted
in a mapping database of ~1,500 business-related administrative procedures. The Bank
also produced a report on recommendations for publishing the mapping online and on
a mechanism to keep the mapping data updated following regulatory changes, and a
roadmap for modernization and digitalization of Government to Business (G2B) service
delivery.
• Department/Division: Finance, Competitiveness, and Innovation Practice Unit
• Project type: Advisory Services
• Project status: Active
• Related SDGs:
o SDG8 (Promote sustained, inclusive, and sustainable economic growth, full and
productive employment, and decent work for all) and
o SDG 16 (Promote peaceful and inclusive societies for sustainable development, provide
access to justice for all and build effective, accountable and inclusive institutions at all
levels)
• Partners (internal and external to WBG): European Union and Ministry of Economy and
Sustainable Development (MoESD) of Croatia
• Any publicly available URL links/multimedia: Assessment of Digital Government to
Business Services Business Environment Reform II Croatia
• Project lead contact: Goran Vranic (gvranic@worldbank.org)
Project 8: Croatia Innovation and Digital Economy

• Project/initiative description: The World Bank is supporting the government of Croatia
by informing strategies for the improvement of the innovation performance of firms, and
supporting the promotion of firm digitalization, as is set out in the sub-component C1.1.2.
Resilient, Green and Digital Economy of the Croatian EU National Recovery and Resilience
Plan (“NRRP”). The relevant sub-objective is the third sub -objective which is to provide
technical input for the design and implementation of a national plan for digitalization and
artificial intelligence.
• Project type (operation, ASA, guidance note, something else): Advisory Services
• Project status: Pipeline
• Related SDGs: SDG 9: Build resilient infrastructure, promote inclusive and sustainable
industrialization and foster innovation.
• Partners (internal and external to WBG): TBD
327
• Any publicly available URL links/multimedia: NA

• Project lead contact: Anita Okemini (aokemini@worldbank.org)
Project 9: Disruptive Technologies for Agile Business Regulation

• Project/initiative description: The Disruptive Technologies for Agile Business Regulation
(Agile & RegTech KPD) was implemented from April 2019 through May 2021 to develop
and pilot agile regulation approaches to unlock investments and enable regulatory service
delivery while safeguarding public policy concerns and mitigating risks. The main result
is introducing a new solution area/workstream within the World Bank Group – Regulatory
Technology (RegTech). The project supported organizing one of the first workshops on
Agile Regulation in cooperation with OECD, European Commission, UK Government
Better Regulation Executive (BRE), and Massachusetts Institute of Technology Internet
Policy Research Initiative (MIT-IPRI). For the agile regulation component, on the onset
of the COVID-19 pandemic in March and April 2020, the project produced an internal
white paper co-authored with OECD on AI and other emerging technologies to inform
regulatory and non-regulatory measures to fight COVID-19. For the integrated regulatory
delivery, the project developed research and publication on the New Technologies for
Regulatory Delivery under the Donor Committee for Enterprise Development (DCED)
• Project type: Report
• Project status: Closed
• Related SDGs:
levels)
• Partners:
o Internal WBG partners: IFC

o External partners: OECD, European Commission, UK Government Better Regulation
Executive (BRE), and Massachusetts Institute of Technology (MIT)
o Use of New Technologies for Regulatory Delivery, available at: https://www.enterprise

-development.org/wp-content/uploads/DCED-BEWG-Use-of-New-Technologies-in
-Regulatory-Delivery.pdf
o Blog and database: Technology helps strengthen countries’ regulatory capacity to
respond to COVID-19, available at: https://blogs.worldbank.org/psd/technology
-helps-strengthen-countries-regulatory-capacity-respond-covid-19
• Project lead contact: Goran Vranic (gvranic@worldbank.org)
Project 10: Business environment guide and diagnostics

• Project/initiative description: This project developed a business environment operational
guide for project teams and a framework for a business regulation diagnostic framework
to inform WBG analytics and facilitate the design of business environment reform
programs. The business environment operational guide provides the overall framework
and chapters for various business environment topics (entry and exit, operations,
328
regulatory technology, informality, gender). The other deliverables (diagnostic framework,

policy note on data-driven business registries, deliver framework for licensing and
inspections, and the determinants of informality note) complement the operational guide
by developing further the topic presented in the guide and providing more detailed
WBG
toolkits. The guidance note on Data Driven Company Registry takes a deep dive into
the frontier developments in the use of data, AI and other emerging technologies
for company and business registration. The note summarizes key regulatory policies,
presents an initial maturity model, and an implementation approach with a high-level
roadmap. The note benefits from lessons learned from successful pilot projects (Denmark
and Greece). It explores AI-augmented real-time company registration and the use of
AI for the prevention of fraudulent behavior. It complements the chapter on Regulatory
Technology Data and G2B in the Business Environment Operational Guide.
• Project type: Report
• Project status: Closed
• Related SDGs:
levels)
• Partners
• Project lead contact: Sylvia Solf (ssolf@worldbank.org)
Project 11: Logistics performance index

• Project/initiative description: Logistics Performance Index 2022: The objective of this
activity is to produce a new version of the Logistics Performance Index (LPI; a biennial
index established by the World Bank in 2007) and its component indicators, using
Machine Learning and Big Data techniques, to make sense of large supply chain tracking
data. This dataset will help policymakers to benchmark a country’s logistics performance
and to develop informed policies and interventions in trade- and transport-related
infrastructure, border management, and logistics services delivery.
The project relies on Machine Learning for two components:
o Deriving country indicators mimicking the scale of the survey bases LPI, using the
information from a large number of supply chain tracking variables.
o In the future, derive logistics related indicators from GIS information (E.g. density of
logistics) or social network data (Twitter, LinkedIn). Research in this area is postponed
until delivery of the main indicator in late 2022.

• Project type (operation, ASA, guidance note, something else): ASA, data
• Related SDGs:
o SDG 8 – Decent work and economic growth;
329
o SDG 9 – Industry, Innovation, and Infrastructure
• Partners (internal and external to WBG):
o WBG-internal (all in advisory role): Transport Global Practice Unit, Development

Economics Unit (DEC)
o WBG-external (data partners): Universal Postal Union (UPU); TradeLens; IATA CargoiQ;
MDS Transmodal
o WBG-external (advisory role): UNCTAD
• Any publicly available URL links/multimedia: Publication planned for December 2022)
• Project lead contacts: Christina Wiederer (cwiederer@worldbank.org); Jean-Francois
Arvis(jarvis1@worldbank.org)
Project 12: Economic networks

• Project/initiative description: This project takes stock of existing leads and POC before
proposing a concept note to propose applying ML-inspired methodology to improved
trade costs and trade composition/complexity databases. The concept is to exploit the low
effective dimensionality of actual trade networks and identify with ML the latent features.
This applies to trade composition and trade costs. ML algorithms identify latent geometry
consistently with known elementary economic mechanisms (productivity, gravity).
• Unit/GP leading: Finance, Competitiveness, and Innovation Practice Unit (Trade and
Competitiveness)
• Project type: ASA, data
• Related SDGs
• Partners (internal and external to WBG): TBD , potentially WTO
• Any publicly available: internal presentation available, earlier research working papers
• Project lead contacts: Jean-François Arvis (jarvis1@worldbank.org)
Project 13: Fraud analytics in Algeria

• Project/initiative description: This multidimensional project provides technical support
to Algeria’s Customs Services as part of a broader project regarding barriers to exports in
Algeria. One of the activities consist in developing a machine learning model to predict
the probability that a new-submission is fraudulent. Algerian Customs has shared its
historical database with the Bank to facilitate this activity.
• Department/Division: Finance, Competitiveness, and Innovation Practice Unit and Middle
East / North Africa Regional Unit
• Project type (operation, ASA, guidance note, something else): Technical Assistance
• Related SDGs
o WBG-internal: Development Economics Unit
• Any publicly available: internal presentation available, working papers summarizing

experimentation in preparation
330
• Project lead contacts: Jean-François Arvis(jarvis1@worldbank.org)
Project 14: Fraud analytics Kenya
WBG
• Project/initiative description: This project aims at helping the Kenyan Revenue authority
implement tools to better target fraud of corruption, including AI application similar.
• Unit/GP leading it: Macroeconomics, Trade, and Investment Practice Unit
• Project type: Technical Assistance
• Project status: Proof of concept stage; data acquisition pending.
• Related SDGs
o WBG-internal, Development Economics Unit
• Any publicly available: internal presentation available

• Project lead contacts: Jean-François Arvis (jarvis1@worldbank.org)
Project 15: Fraud analytics Armenia aka smart risk management

• Project/initiative description: This project helps the Armenian revenue authority
implement a fraud targeting tool based on AI.
• Unit/GP leading: Finance, Competitiveness, and Innovation Unit
• Project type: Technical assistance as part of a large EU-funded program on tax reform.
• Project status: Not yet started
• Project start year: 2022-23
• Related SDGs
o WBG-internal: Macroeconomics, Trade, and Investment Practice Unit; Development

Economics Unit
• Any publicly available: None yet

• Project lead contacts: Jean-François Arvis(jarvis1@worldbank.org)
Project 16: Using remote sensing and artificial intelligence to measure trade node
activity
• Project/initiative description: The objective of this project is to demonstrate the use of
alternative tools based on artificial intelligence and remote sensing of domestic and
international trade nodes and corridors, based on quantitative, objective metrics, to
inform policy and country interventions.
• Unit/GP leading: Trade and Regional Integration Unit; Macroeconomics, Trade, and
Investment Unit
• Project type: Research report
• Project status: Active
• Related SDGs
• Partners (internal and external to WBG): DEVELOPMENT DATA GROUP
331
• Any publicly available URL links/multimedia: N/A

• Project lead contact: Daniel Saslavsky (dsaslavsky@worldbank.org); Satya Prasad Sahu
(ssahu8@worldbank.org)
Project 17: Monitoring hate speech and misinformation on social media in Nigeria
• Project/initiative description: The World Bank Development Impact Evaluation Department
(DIME) and the Nigeria team of the UK Foreign, Commonwealth and Development Office
(FCDO) have launched a collaboration to track and understand the spread and impact of
misinformation, fake news, hate speech and divisive narratives in Nigeria. This interactive
dashboard highlights trends in the proliferation of hate speech and fake news on Twitter
and patterns of internet searches containing references to ethnic or religious groups and
general terms associated with politics and potential conflict.
• Department/Division: Development Impact Evaluation Unit (DIME) and Data Analytics
and Tools Unit (DECAT)
• Project type: Research report and national dashboard
• Project start year: September 2021
• Project end year: September 2023
• Related SDGs: 11/16
• Partners (internal and external to WBG): UK Foreign, Commonwealth and Development
Office (FCDO)
• Any publicly available URL links/multimedia: https://datanalytics.worldbank.org/ng-social
-media-analytics/
• Project lead contact: Sam Fraiberger (sfraiberger@worldbank.org); Victor Orozco
(vorozco@worldbank.org)
Project 18: Impact of artificial intelligence in developing countries

• Project/initiative description: The objective of this report is to inform the policy debate
around AI in developing countries and help them prepare for the future. The first part
of this report will provide an overview of the technology, the AI value-chain, and the
supporting infrastructure needed to enable AI applications and solutions. The second
part of the report will explore AI applications in developing countries, including a
review of economic opportunities and risks as well as opportunities and risks in social
development. The third part of this report will present an overview of AI policy options
and different approaches to governance models in a selection of leading AI countries.
The final part of this report will present an approach to support the design of policy
responses in developing countries.
• Unit/GP leading: Digital Development Unit
• Related SDGs: SDG 8, 9, 10
• Partners: Jain Family Institute
• Project lead contact: Rong Chen (rchen5@worldbank.org)
332
Project 19: Artificial Intelligence in the Public Sector: Maximizing Opportunities,

Managing Risks
• Project/initiative description: This report provides a preliminary synthesis of the existing
WBG
opportunities, risks, and building blocks required for implementing and integrating AI
in government operations. The report also highlights policy, governance, and people
aspects necessary for AI implementation. It draws on the accumulated literature, case
studies, and emerging trends to provide guidance to World Bank teams working in this
field.
• Unit/GP leading: Equitable Growth, Finance, and Institutions (Governance Unit)
• Project status: Completed
• Related SDGs: 16
• Partners: GovTech Global Partnership
• Any publicly available URL links/multimedia: https://documents1.worldbank.org/curated/
en/809611616042736565/pdf/Artificial-Intelligence-in-the-Public-Sector-Maximizing
-Opportunities-Managing-Risks.pdf
• Project lead contact: Khuram Farooq (kfarooq@worldbank.org)
Project 20: Risk Assessment Framework to Identify and Classify Ethical Risks from AI
use in World Bank Projects
• Project/initiative description: This initiatives aims to develop a risk assessment framework
to identify and classify ethical risks that might be present in World Bank projects, using the
OECD Framework for the Classification of AI Systems as a basis for AI tasks classification.
Approaches to AI risk assessment frameworks by other multi-lateral development banks
and related institutions will also be reviewed.
• Unit/GP leading: Digital Development
• Related SDGs:
• Project lead contact: Rami Amin (ramin3@worldbank.org)
Project 21: Starting Points for Trustworthy AI at the World Bank, using the WBG’s
Environmental and Social Framework (ESF)
• Project/initiative description: This report explores artificial intelligence in the context of
the Bank’s Environmental and Social Framework (ESF), which has emerged as the primary
lens through which the Bank assesses risks and social impacts. The purpose of this report is
to identify key ethical commitments relating specifically to artificial intelligence within the
ESF and related frameworks at the Bank. Approaches by other multi-lateral development
banks and related institutions to develop trustworthy AI will also be reviewed.
• Unit/GP leading: Digital Development
333

• Related SDGs:
Project 22: Tools for Identifying Human Rights Impact and Ensuring Algorithmic
Accountability in World Bank Operations Using Artificial Intelligence and Big Data
Analytics
• Project/initiative description: Building on the rights-based trust framework proposed by
the WDR and based on the experiences of COVID-19 operations, this proposal is aimed
at ensuring that staff working on World Bank operations have the tools to (i) identify when
AI elements of components might impact human rights, and (ii) ensure that appropriate
and proportionate rights-ensuring “algorithmic accountability” elements are included in
such operations.
• Unit/GP leading: Digital Development Unit; Legal Unit (LEGOP)
• Related SDGs: 16
• Partners (internal and external to WBG)

SDGs 1, 2, 3, 4, 5, 8, 9, 10, 13, 16, 17
3. Related Links
https://www.worldbank.org/en/home
334
World Food Programme
WFP
Project 1: SKAI
• Project Description: A lack of on-the-ground information at the start of a humanitarian
crisis is a major obstacle to a quick, effective response. Following disasters, WFP works to
assess the magnitude of damage, the needs of local communities, and its humanitarian
intervention plans to mobilize resources and coordinate emergency response efficiently.
As part of its frontier innovations portfolio, WFP has been exploring new technologies
that can automate this process and speed up response times. WFP partnered with
Google Research to set up SKAI, a humanitarian response mapping project powered by
artificial intelligence — an approach that combines statistical methods, data and modern
computing techniques to automate specific tasks. SKAI assesses damage to buildings by
applying computer vision — computer algorithms that can interpret information extracted
from visual materials such as, in this case, satellite images of areas impacted by conflict,
climate events, or other disasters. The key to this process is a machine learning model
developed specifically for SKA, which detects damaged buildings by comparing imagery
of the same buildings before and after the disaster. SKAI aims to leverage the power of
artificial intelligence and remote sensing to assess damage within 24 hours after disasters
take place.
• Department/Division: INKA
• Project updates: Since its June 2022 open-sourcing on Github with user-friendly Colab
materials, SKAI has proven its effectiveness in multiple disaster responses. Notably,
GiveDirectly leveraged SKAI to expedite cash assistance to Florida households after
Hurricane Ian in 2022. It also supported post-flooding efforts in Pakistan with the Khyber
Pakhtunkhwa government and CrisisReady in Libya. In 2023, SKAI provided crucial
damage assessments to Mercy Corps in Myanmar after Cyclone Mocha and internally to
the World Food Programme following the Turkey-Syria earthquakes. SKAI continues to
expand its reach, empowering governmental and humanitarian organizations to build
and strengthen disaster management capabilities.
• Project Domain: Emergency, Rehabilitation
• Publicly Available Data : Yes
• Technology/Platform: Deep neural network machine learning model, Google Earth
Engine APP, and Google Cloud Platform.
• Related Sustainable Development Goals (SDGs): SDG 2 - Zero Hunger
335
o UN Partners: WFP
o Private Sector: Google Research
• Lessons Learned: Main challenges were twofold. First, the model did not generalize well
for new types of disasters in new geographical areas. This challenge was addressed by
developing the machine learning model that would be trained for a new disaster using a
small number of labelled images. The new SKAI model uses a semi-supervised learning
technique that reduces the required number of labeled examples. As such, SKAI typically
only needs 200 to 500 labeled examples to achieve high accuracy, significantly improving
the speed at which accurate results can be obtained. The ambition for SKAI is to optimize
its platform to function across a variety of geographic locations, disasters, and damage
types.
Second, we need to research more into the ways how artificial intelligence-powered
damage assessment platforms like SKAI can be operationalized at scale in humanitarian
response. .
To address this challenge, the SKAI team has developed a cloud-based platform that
streamlines the loop of image labelling, model training, model validation and fine-tuning.
This can potentially turn SKAI into a self-service damage assessment platform that can
be used by humanitarian practitioners who have little knowledge of machine learning,
improving the user experience and performance of SKAI.
o GitHub - google-research/skai: SKAI is a machine learning based tool for performing

automatic building damage assessments on aerial imagery of disaster sites.
o SKAI | WFP Innovation
o How AI helped 6x our disaster response speed | GiveDirectly
o Tropical Storm Daniel Causes Mass Flooding in Libya • CrisisReady
o [1910.06444] Building Damage Detection in Satellite Imagery Using Convolutional
Neural Networks (arxiv.org)
o [2011.14004] Assessing Post-Disaster Damage from Satellite Imagery using Semi-
Supervised Learning Techniques (arxiv.org)
o The SKAI isn’t the limit: How WFP uses satellite imagery and machine learning in
emergencies | by WFP Innovation Accelerator | Medium
o Machine Learning-based Damage Assessment for Disaster Relief – Google Research
Blog
• Contact Information: Amine Baha (amine.baha@wfp.org)
Project 2: HungerMap LIVE

• Project Description: Understanding the food security situation requires a thorough analysis
of data that are scattered across different data sources and platforms. HungerMapLIVE
brings together different streams of publicly available information on food security,
nutrition, conflict, weather and a variety of macro-economic data – including from WFP
– all in one place to provide a holistic overview of the food security situation at global,
country and sub-national levels. The resulting analysis is displayed on an interactive map
using advanced data visualization tools. As WFP’s global food security monitoring system,
the HungerMap LIVE:
enhances operational effectiveness by identifying areas that are sliding towards food
insecurity, providing information on shocks, hazards and other drivers of hunger in real-
time, ensuring more informed and timely response to food crises;
maximises efficiency by providing continuously updated data at a lower cost and in less
time, compared to traditional food security monitoring systems alone;
336
ensures continuously updated information in countries of interest, regardless of

accessibility issues and the scale of WFP’s operational presence.
• Department/Division: Research, Assessment and Monitoring Division
WFP
• Project Type/Output: Integrated food security information system
• Reported as part of 2022 compendium on UN AI Activities? Yes
• Project updates: Since 2020, the HungerMap LIVE has been upgraded in several aspects.
Firstly, near real-time food security monitoring systems were scaled up from 15 to over
30 countries to provide continuous updates on the food security situation in countries in
or at risk of food crises.
Secondly, the HungerMap LIVE has been improved by applying a new machine learning-
based predictive model, which has been trained using a much broader set of data and a
refined methodology. The new model provides more precise estimates of the number of
people with insufficient food consumption. A technical paper on the predictive model has
been submitted for publication (currently under peer review, please see the link below).
Furthermore, WFP has been working on new predictive models to forecast food security
indicators up to 1 month in advance, in addition to nowcasting. A technical paper about
the forecastability of food insecurity has been submitted for peer review (please see link
below).
Thirdly, an alerting system has been integrated into the HungerMap LIVE, which signals a
marked deterioration in food intake, in COVID-19 cases, in conflict-related fatalities, and
in vegetation anomaly. These new features aim to capture the impact of key drivers of
hunger and signal improving or deteriorating circumstances in near-real time.
Finally, a new risk classification framework has been developed using key food security
indicators. All data and country classifications are made available through the HungerMap
LIVE platform. In addition, a range of new resources including Global, Regional and
Country Insights and Key Trends are available and updated daily on the platform (please
see the links below).
• Project Domain: Hunger and food security
• Data Source: The HungerMap LIVE combines key metrics from various data sources –
such as food security information, weather, population size, conflict, hazards, nutrition
information and macro-economic data – to help assess, monitor and predict the magnitude
and severity of hunger in near real-time. The HungerMap LIVE system comprises:
Remote, near real-time food security monitoring systems, collecting data on key IPC/CH
indicators every day in over 30 countries experiencing acute food crises and those where
WFP has the largest operations.
Machine learning-based predictive models, providing estimates of the prevalence of
acute food insecurity in countries and areas that are stable and where near real-time data
is not active yet.
The HungerMap LIVE global platform, where users can access the food security
information to monitor areas at risk or deteriorating across a range of indicators. Related
resources include Global, Regional and Country Insights and Key Trends.
• Link to data: https://hungermap.wfp.org/
• Publicly Available Data : Yes
• Predictive model: Python, Docker.
HungerMap frontend systems: AlibabaCloud (FunctionCompute (serverless APIs), API
Gateway (REST APIs), Elastic Compute Service (computing), OSS (object storage), RDS
(relational DB), Elastic Container Instance (managed serverless Docker), cloud CDN,
DataWorks (data integration)), Azure DevOps (automation), React (frontend)
• Data backends: RedHat OpenShift (k8s-based running environment), MSSQL (relational
DB)
337
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG – 17

o UN Partners: Food Security Information Network

o Government: USAID, DEVCO, DLR
o Academia: Institute for Scientific Interchange Foundation
• Lessons Learned: Despite the challenges posed by the COVID-19 pandemic, WFP has
successfully leveraged the existing remote, near real-time food security monitoring
systems to collect data from new sectors such as school-feeding and plans to expand
data collection to other sectors, which will be integrated into the HungerMap LIVE.
For example, WFP is currently working on the integration of gender-disaggregated data
and gender-sensitive data. To this end, questionnaires have been adjusted to collect more
gender-sensitive data. This information will be displayed in the HungerMap LIVE, along
with gender-disaggregated results to better understand intra-household differences in
vulnerability.
In addition, the HungerMap LIVE team is working with relevant WFP units to integrate
climate-related metrics such as data on climate shocks as well as nutrition data (e.g.,
micronutrients, acute malnutrition) to the system.
o HungerMap LIVE: https://hungermap.wfp.org/

o Global Insights and Key Trends: https://static.hungermapdata.org/insight-reports/
latest/global-summary.pdf
o Middle East and Northern Africa Insights and Key Trends: https://static.hungermapdata
.org/insight-reports/latest/rbc-summary.pdf
o Southern Africa Insights and Key Trends: https://static.hungermapdata.org/insight
-reports/latest/rbj-summary.pdf
o Eastern Africa Insights and Key Trends: https://static.hungermapdata.org/insight
-reports/latest/rbn-summary.pdf
o Western Africa Insights and Key Trends: https://static.hungermapdata.org/insight
-reports/latest/rbd-summary.pdf
o Asia and the Pacific Insights and Key Trends: https://static.hungermapdata.org/insight
-reports/latest/rbb-summary.pdf
o Latin America and the Caribbean Insights and Key Trends: https://static.hungermapdata
.org/insight-reports/latest/rbp-summary.pdf
o Country Insights and Key Trends: accessible through the HungerMap LIVE country
pages
o ‘Nowcasting food insecurity on a global scale’ paper: https://www.medrxiv.org/
content/10.1101/2021.06.23.21259419v1
o ‘On the forecastability of food security’ paper: https://www.medrxiv.org/content/10
.1101/2021.07.09.21260276v1
o WFP Hunger Monitoring Unit’s blog: https://mvam.org/
o WFP Hunger Monitoring Unit’s twitter account: https://twitter.com/mobileVAM
o HungerMap Launch WFP Insight: https://www.wfp.org/stories/wfp-launches
-hungermap-live
o HungerMap LIVE mock-up: https://docs.wfp.org/api/documents/WFP-0000131795/
download/
o HungerMap LIVE & products mock-up:https://docs.wfp.org/api/documents/WFP
-0000131794/download/
• Contact Information: Jonathan Rivers (jonathan.rivers@wfp.org)
338
Project 3: MEZA (an Optical Character Recognition system that uses Artificial
Intelligence to digitize handwritten records, speeding up data collection and analysis
processes and allowing decision makers make data-based decisions in a timely
manner.
WFP
• Project Description: Nutrition records for millions of malnourished children lie in remote
health clinics around the world. Many of these clinics record patient data using paper-
based booklets, which may be easily lost or destroyed. Different stakeholders involved
in the fight against malnutrition in affected countries acknowledge that the digitization of
these conventional paper-based systems would increase the efficiency and effectiveness
of malnutrition management efforts.
Meza is a tool powered by artificial intelligence, developed by Charitable Analytics
International to help digitize handwritten data from the deep field. MEZA uses an Optical
Character Recognition technology to rapidly collect nutrition and related health data
from remote, low-resource health clinics, enabling WFP and governments to have the
information they need to provide high-quality, context-specific, and timely nutrition
support. Following a WFP Innovation Bootcamp in June 2018, WFP’s Nutrition Division
of WFP and WFP’s Country Office in the Republic of Congo identified Meza as a potential
solution that could enhance the digitization of beneficiary management systems in the
Republic of Congo.
From November 2018 to December 2022, the tool was tested across 70 clinics supported
by WFP in the Republic of Congo and Rwanda through three pilots funded by the WFP
Innovation Accelerator. The third pilot in Rwanda validated the main key hypotheses
of Meza as it had a high accuracy level, over 96 percent, and had validated its ease of
use by clinic health workers. Since the successful completion of the Meza pilot, it has
expanded to Kenya through a collaboration between WFP Kenya Country Office (CO) and
the government of Kenya. Meza is currently used in 79 health facilities in Kenya, ensuring
accurate and timely reporting on health treatments, actively monitoring over 200,000
health indicators.
• Entity Name: Innovation Accelerator
• Department/Division: Innovation and Knowledge Management
• Project Status: Graduated
• Project End Year: The project has graduated from the portfolio of WFP Innovation
Accelerator after a successful pilot that allowed its adoption in Kenya and made it ready
for use in other Countries where WFP would find it relevant
• Project Domain: Agriculture; Education; Health
• Data Source: MEZA uses nutrition data to refine the services offered to people benefiting
from various interventions implemented by WFP. These data include, for example:
nutrition status of children under 5; nutrition status of pregnant and lactating women
and school feeding programme-related data such as information on student attendance.
• Technology/Platform: The tech used is CAI’s own software.
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG 3 – Good
Health and Well-Being; SDG 4 – Quality Education; SDG 17 – Partnerships for the Goals
• Partnership(s)/ Collaborator(s):
o UN Partners: Indirect collaboration with UNICEF & UNHCR during the Proof of Concept
currently tested
o Private Sector: Charitable Analytics International
339
• Lessons Learned: WFP conducted a performance review in the summer of 2020 and
below are the main lessons learned from the two pilots that were conducted in the
Republic of Congo and which are still relevant to date.
1. Facilitate access to mobile network or internet connectivity:

Clinic workers cited connectivity as one of the key challenges they faced. For instance,
one clinic worker had to travel 10 km to access the network and submit photos of the
logbooks. A potential solution to mitigate connectivity limitations could include ICT
assessments carried out before deploying a digital solution to ensure selected sites
have adequate network connectivity to sync data. If the deployment sites are found
to lack connectivity, one of the below models could be used to ensure regular data
synchronization:
Bringing mobile devices with adequate internet or 3G connectivity to a central location
(for instance, nearby cities or towns) on a weekly or bi-weekly basis; or
Deploying district health officials or WFP staff from field offices as roving agents to
visit sites at regular intervals to take photos and sync data with the Meza server once
back to base with connectivity.
2. Invest in digital literacy and capacity building for frontline health workers:
Health workers at remote locations in the Republic of Congo have rather limited
exposure to mobile technologies. For instance, out of the five clinic workers interviewed
in the post-pilot review, two reported that they had never used smartphones before
participating in Meza. While the training before deployment helped introduce health
workers to smartphones, the volume of submissions and the quality of photos suggest
that some frontline workers need more training than others. Introductory training
needs to be tailored to match health workers’ capabilities and experience using
mobile devices, and follow-up training should be organized at regular intervals to
reinforce learning Health workers’ capacity to record information in the Meza logbook
will contribute to the overall effectiveness of nutrition programmes by minimizing the
risk of human error and enhancing data collection to guide effective decision-making.
3. Build confidence in the data extracted from Meza:
The OCR performed well during the first implementation phase of the Meza pilot as
the logbook template was simpler and offered the space required for health workers
to write using a large and clear script. The logbooks had to be redesigned during
phase two as the CO project leads noticed that the P1 templates only covered 70
percent of the programmatic data needed to produce a meaningful analysis. The
P2 logbooks were designed to meet nutrition reporting requirements. However, the
OCR performed poorly because of the P2 templates’ compact structure as well as
data contents (e.g. words and dates) which were beyond the technology’s decoding
capabilities.
Subsequently, the data extracted from the tool was not adequately analyzed or utilized
in order to determine if it could provide actionable insights - the core value proposition
of the project. During phase one, the Meza project managers from the CO M&E unit
fed data from the OCR into a Tableau dashboard in order to test if it was possible
to derive useful insights. However, staff turnover in the M&E team and the logbook
template re-design during phase two meant that the link with Tableau was broken.
Additionally, no one at the CO had access to the Tableau dashboards or the technical
skills necessary to extract data from the Meza web platform and ingest it into another
analytics platform. As the business had no visibility of the data resting in Meza in
order to share feedback with external stakeholders, there was scepticism towards the
project.
Going forward, this project should consider the following factors to build a high level of
confidence in the data extracted from the tool:
1. Clear specification of the capabilities and limitations of the OCR technology (including
but not limited to the type of characters the OCR can or cannot recognize, types of
340
template that the OCR can read without excessive customizations to the template or
the technology, prediction accuracy rate and the conditions required for the tool to
perform optimally).
2. Close collaboration between the team designing the data collection template and
WFP
CAI, including adequate time to test before deploying to the field.
3. Determination of an acceptable data accuracy threshold by the business in line
with their analytical requirements. This entails close collaboration with and frequent
consultations with the end-users of the data (for example Nutritionists at the CO or
HQ-levels), to understand their data needs and minimum acceptable data quality to
inform programmatic decision-making.
4. Continuous monitoring of data extracted from the Meza web platform vis-à-vis paper
records to identify and rectify any discrepancies. Constant and timely feedback to the
Meza developers will be valuable in enabling them to improve the tool’s performance
and data quality.
5. Identification of a data focal point at the CO-level, alongside training geared towards
building data literacy for all relevant decision-makers involved in the programme.
Alternatively, if capacity at CO is limited, business stakeholders at the HQ or RB-levels
should identify and dedicate the resources required to optimally consume data for
informing decisions.
6. Regular communication and dissemination of reports using data received through
Meza to relevant external stakeholders, including health clinics, to demonstrate the
tool’s value in supporting programme efficiency.
Thanks to these learnings, we have decided to test Meza’s proof of concept in a
controlled environment in collaboration with WFP’s technology division. If the concept
is proved, the solution will be re-deployed to WFP Country Offices again.
• Contact Information: Nicolas Umuhizi (nicolas.umuhizi@wfp.org)
Project 4: Optimus
• Project Description: WFP staff face complex, cross-functional challenges every day, often
with many possible solutions. Whether it’s funding shortfalls, access restrictions due to
rainy seasons, new import regulations, or operational scale-ups, every day it’s something
new. To properly manage the complexity and to enable an agile comparison of potential
alternatives, it is critical to use data and optimization to find the right solutions. Optimus
is a web application that looks at WFP operations end-to-end to support better planning,
helping users identify the most cost-effective solutions using advanced mathematics.
It pulls together a wide variety of data—from beneficiary numbers to sourcing options
and from transport routes to nutritional values. Users can create their own scenarios
or ask Optimus to find optimal plans, taking into account operational restrictions such
as lead times and funding and preferences such as nutritional value targets and local
procurement targets. A user-friendly interface allows users from any functional area to
quickly explore and compare different scenarios.
• Department/Division: Supply Chain/Planning Service
• Project updates: Improved user experience: The tool has been refactored, thus becoming
increasingly user-friendly.
• Project Domain: Supply Chain
• Data Source: End-to-end Supply chain data
341
o Back End: Django 2.x + Django REST Framework DRF 3.x + Python 3.x, optimization
solver: COIN-OR
o Front End: react.js + Redux
o Team development: Microsoft Azure
o Hosting server: AWS

o Private Sector: UPS

o Academia: The Georgia Institute of Technology (USA), Tilburg University (Netherlands)
o Lessons Learned: It is crucial to develop the new tool together with the corporate
IT team to ensure smooth adoption, integration, and implementation of the best
corporate human-centered design practices.
o https://optimus.wfp.org/login/?next=/%3F_ga%3D2.212128916.1670390044
.1631111249-1871728533.1615974575
o https://innovation.wfp.org/project/optimus
o https://www.youtube.com/watch?v=wdEcVj5LTGg
• Contact Information: Koen Peters (koen.peters@wfp.org)
Project 5: Humanitarian Chatbots

• Project Description: In emergency and development contexts, communication with
affected communities is crucial. Having access to accurate, tailored information and
engaging in a dialogue contributes to the resilience and ability of people to cope with a
crisis. Since 2016, WFP has been working on the development and rollout of humanitarian
chatbots to help deliver vital information to the people in urgent need. This technology
improves WFP’s outreach to populations in hard-to-reach areas using a mobile device
or a computer and complementing existing communication channels and WFP’s food
security monitoring systems. This technology has proven to be particularly useful during
COVID-19, as communication with affected communities has become even more crucial
in times of unprecedented uncertainty.
Chatbots are highly customizable. For example, chatbots can be deployed as part of an
assets creation programme to provide information about the use of assets as well as to
collect feedback from the users. As part of Complaint Feedback and Mechanism (CFM)
systems, chatbots have been useful in offering access to information 24/7. Moreover, the
analytics produced by chatbots can inform further interactions; WFP staffs can monitor
errors and improve the technology and user experience over time. As such, chatbots are
highly user-friendly and do not require a high level of computer literacy, making them
accessible to large segments of the population.
• Department/Division: Research, Assessment and Monitoring Division
• Project Updates: In response to the challenges posed by COVID-19 in the Latin America
and Caribbean region, WFP rapidly scaled up humanitarian chatbots to provide critical
information to vulnerable populations on COVID-19 measures, WFP assistance, and l
health-related tips on breastfeeding, nutrition, anemia, among others. Harnessing the
342
ChitChat technology and building on the best practices from previous pilots, WFP
has thus provided easy access to accurate health and safety information for affected
communities. Chatbots have been an integral part of WFP’s humanitarian response in
Peru, Colombia, Nicaragua, and Guatemala during the COVID-19 pandemic outbreak
WFP
and in the face of increasing migration from Venezuela to Peru and Colombia. Chatbots
were integrated into the current Complaint Feedback and Mechanism (CFM) systems.
In addition to the helpline which receives the high volume of calls, chatbots powered by
artificial intelligence can answer the most frequently asked questions and are available
24/7. Users can navigate the chatbot to find information , for example, about WFP,
different social protection programs and selection criteria, and cooperating partners
on the ground. This builds trust, enhances transparency and counters misinformation,
improving the way WFP serves communities facing hunger.
• Project Domain: Agriculture, Health, Weather, Other: Hunger and food security
• Data Source: The data source for the chatbots are stored in the chitchat main site and all
differ in content. No personal information is shared, currently only informational material
is included. Configuration based chat system. No additional data source consumed or
produced.
• Technology/Platform: Chitchat technology - Java, PostgreSQL
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG 17 –
o Government: Government of Peru, Government of Nicaragua, Government of

Guatemala
o Academia: Centre for Innovation of Leiden University
• Lessons Learned: Moving forward, WFP will optimize the chatbot builder interface to
allow the development of more sophisticated chatbots, to support WFP’s global field
operations as well as the wider humanitarian and development community.
WFP will continue developing ad-hoc chatbot scripts leveraging existing key features
such as voice recognition, languages settings, channel integration, adapted to particular
country and operational contexts. Building on lessons learned from previous pilots, WFP
will continue exploring platforms which are more familiar to and widely used by local
communities such as WhatsApp.
o Infochatea (Peru): https://infochatea.per.wfp.org/

o Nutrechatea (Peru): https://nutrechatea.per.wfp.org/
o Lineachatea (Colombia): https://lineachatea.col.wfp.org
o Bonochatea (Nicaragua): https://bonochatea.nic.wfp.org/
o Misegurochatea.(Guatemala): https://misegurochatea.gtm.wfp.org/
o INFOchatea_dashboard: INFOchatea dashboard - Tableau Server (wfp.org)
o Bonochatea_dashboard_final: Bonochatea_Dashboard - Tableau Server (wfp.org)
o WFP Hunger Monitoring Unit’s blog: https://mvam.org/
o WFP Hunger Monitoring Unit’s twitter account: https://twitter.com/mobileVAM
o Humanitarian Chatbots video: https://d ocs. wfp. org/a pi/d ocuments/W FP
-0000131792/download/
o Humanitarian Chatbots photo: https://d ocs. wfp. org/a pi/d ocuments/W FP
-0000131793/download/
o Humanitarian Chatbots: Kenya Pilot https://www.youtube.com/watch?v=A
SHROjd008s
• Contact information: Rossella Bottone (rossella.bottone@wfp.org)
343
Project 6: Voice-to-text AI phone survey tool

• Project Description: WFP’s field numerators may not be able to conduct face-to-face
household nutrition surveys due to the COVID-19 restrictions. In addition, it is costly to
conduct surveys through alternative means like call centres. We wish to use an artificial
intelligence-powered IVR solution to automate the beneficiary surveying process. The
project is set up to (1) test the commercially viable IVR service embedded with AI-powered
speech recognition technology; (2) fine-tune open-sourced speech recognition models
using audio training data from populations that best represent the demography of WFP
beneficiaries in terms of age, gender, region, accent, dining behaviours, and etc.; and
(3) compare the performance of the two technologies. Shall the technologies be proven
effective, the project will aim for a corporate-wide adoption before scaling to the wider
humanitarian and development sector.
• Department/Division: Ethiopia Country Office, Nutrition Division
• Project Updates: The project team is currently collecting, transcribing and verifying audio
data from native Amharic speakers based in the Amhara region.
• Project Domain: Nutrition, but with possibility to expand to other Project Domains
• Data Source: Proprietary audio data that is transcribed into text.
• Link to data: No
• Lessons Learned: Strong data protection provisions will need to be implemented in the
future, as the project may need to use personal identifiable information such as audio
data from users to enhance its functionalities based on historical data.
• Contact Information: Filippo Dibari (filippo.dibari@wfp.org)
Project 7: Child Growth Monitor

• Project Description: Malnutrition is a global health crisis and the leading cause of death
among children under five. To be able to efficiently treat malnutrition, it must first be
detected. This detection requires anthropometric measurements of weight, height, and
middle-upper arm circumference. However, measuring children accurately is a challenge,
especially in the Global South, due to limited resources, unreliable, bulky traditional
measurement hardware and/or lack of trained enumerators. Child Growth Monitor
(CGM), an AI-powered smartphone app launched by the German non-profit organization
Welthungerhilfe, was created to overcome these issues. With CGM, we aim at developing
a non-profit, open-source, mobile solution that leverages artificial intelligence, especially
computer vision machine learning technology, to enable frontline-healthcare workers
to quickly and accurately measure children under five years, using a smartphone. Our
measurements will be used to support the diagnosis of malnutrition of children according
to the WHO growth standards and provide quick and accurate data in the most ethical
way to organizations working on the UN SDGs and in public health.
• Project Type/Output: Academic paper, Dataset, Software tool
• Project End Year: Founding of a non-profit open-source software business in 2022 for
sustainable providing the software and services.
344
o CGM was re-positioned as a COVID-19 pandemic response tool;

o 04/2020 Data tagging tool;
WFP
o 05/2020 App runs on three additional devices;
o 08/2020 Results are now delivered in just 6 minutes, down from 30 minutes and the
application started measuring reliability of measurements;
o 01/2021 Minimum-viable product was rolled out;
o 03/2021 Assessment carried out by the Boston Consulting Group confirmed the
feasibility of the solution, business model and the ability of the project team to create
the product as planned (BCG Gamma 06/2021);
o Plans of testing CGM in further counties: Nepal, Bangladesh, Uganda and Zimbabwe
with BMZ, field test in Namibia with UNAM and UNICEF;
o First discussions conducted with partners for validation study;
o Studies on Racial Bias in artificial intelligence and COVID-19 are being conducted in
collaboration with Microsoft;
o We were able to build up a solid team and restructure the development team into 3
different units.

• Data Source: All the present data was collected from two states of India, Rajasthan (Baran
district) and Madhya Pradesh.
(Chatarpur and Sheopur districts), during 2017-2019, using the Child Growth Monitor
phone app developed by Welthungerhilfe. We focus on data collected for children
who can stand (usually two-five years of age) for this work. The data was collected in
the regional Anganwadi centers, which are a type of rural child care center in India. The
data collectors were mostly young adults (20-30 years old) and received a four-day data
collection training.
After receiving consent from their legal guardians, children were asked to stand in front
of a solid-colored wall. If needed, a white banner was placed behind the child to replicate
a wall. All the videos were recorded using the Lenovo Phab 2 Pro phone, which has
a time-of-flight sensor to capture point cloud data at 1920x1080 resolution with three
frames/second. The point cloud videos were converted into depth images in the data
processing stage. For each child, the data collector used the phone app to collect three
point cloud videos: (a) front video: where the child is facing the camera, (b) back video:
where the child’s back is facing the camera (Figure 1a), and (c) 360 video: where the child
was asked to spin slowly to capture a 360-degree view of the child. The data collector
decided the length of these videos; usually, the front and back videos were 2-4 seconds
long, while the 360 videos were 5-8 seconds. (Note: For front and back data, a single
image would have sufficed, however as children move frequently, we opted for videos.)
The data collector ensured that the child’s head to toe was fully visible in each video.
Next, manual measurements of the ground truth weight, height, and mid-upper arm
circumference (MUAC) were taken, using the standardized weight machine, height board,
and MUAC tape, respectively. On average, it took 15-20 mins to collect data for a child,
involving consent forms, digital videos, and manual measurements. In case the child did
not co-operate, they move to the next child. The child and/or guardian did not receive any
incentive for participation. Overall, data was collected for 3887 children, and the age-wise
distribution of the point cloud video dataset is shown in Table I.
345
TABLE I – AGE-WISE DISTRIBUTION OF TRAINING AND TEST VIDEOS
Age Total Training Test

2-3 1030 712 318
3-4 1370 895 475
4-5 1487 974 513
Total 3887 2581 1306

o Lenovo Phab 2 Pro / Huawei P30 Pro Android Smartphones

o Microsoft Azure
o Python
o Tensorflow
o Jupyter Notebooks
o More details can be found in the cgm-* GitHub repositories https://github.com/orgs/
Welthungerhilfe/repositories

Hunger; SDG 3 – Good Health and Well-Being; SDG 17 – Partnerships for the Goals
o Partners in Measurement Prediction (AI and ML) (Support with research and
development of the approach to measurement prediction): Technische Hochschule
Ingolstadt, Leipzig Research Center for Civilization Diseases, Tilburg University,
Microsoft, Open Sorce Community
o Partners in Implementation in the Field (Distribution of mobile application on
commodity smartphones to field workers, as well as further IT Integration and
support): UNICEF, Government of Madhya Pradesh, Action Contre La Faim, Fight
Against Hunger Foundation, Clifford Chance, Msg Advisors, Global Nutrition Cluster,
University of Namibia, SMART
o Partners in Tech (Collaboration with partners to bring the solution from specific IR
phones to commodity smartphones): Microsoft, SONY, PHAT Consulting
o Partners is Endorsement & Funding (Endorsement of mobile application in the field,
dissemination of data, demonstration of usage benefits and rewards): WFP, Deutsche
Telekom, Microsoft, GSMA ,Boston Consulting Group, Tereska Foundation, Happel
Stiftung, Government of Madhya Pradesh, Munich RE, Federal Ministry for Economic
Cooperation and Development (BMZ) ,Deutsche Gesellschaft für internationale
Zusammenarbeit (GIZ)
o https://arxiv.org/abs/2105.01688
o https://childgrowthmonitor.org/
o http://github.com/Welthungerhilfe/
o https://twitter.com/ChildGrowthMon
o https://www.youtube.com/watch?v=f2doV43jdwg
o https://www.youtube.com/watch?v=PAvGwHqgr8k
o https://www.youtube.com/watch?v=Ni9PlO00cZ0
o https://www.youtube.com/watch?v=FfYxIkp_vw4
346
o https://www.youtube.com/watch?v=RuluVPJLTEA
o https://w ww. itu. int/e n/I TU- T/A I/2 018/D ocuments/P resentations/J ochen% 20
Moninger.pdf
WFP
1. Well-aligned partnerships are essential to driving the innovation process.

2. A stage-gate project approach with well-defined milestones can minimize risks and
is key to funding a complex and long-term innovation.
3. Ensuring funding for Software and Machine Learning Engineers, Data Scientists
and Software Development may be challenging in the current system that tends to
prioritize emergency response and short-term impact over driving game-changing
innovations involving a higher risk and uncertainty.
User-centric design and agile organizations and processes are required to achieve long-
term sustainable solutions.
• Contact Information: Markus Pohl (Head of Project) / Markus Matiaschek (Head of Tech)
(Markus.Pohl@welthungerhilfe.de / mmatiaschek@childgrowthmonitor.org,)
Project 8: Combination of Drone and Sentinel-2 data for crop type mapping over
areas of resilience interventions
• Project Description: Crop land and type mapping is crucial in the assessment of
agricultural production of a country. It is also a critical prerequisite for the monitoring and
assessment of changes in agricultural livelihood resources, which is particularly important
in areas affected by conflict, natural disasters and other disruptions, allowing staff to timely
allocate resources and deliver food assistance. However, the collection of ground data
tends to take considerable time and may be particularly expensive and hard to carry out
in emergency contexts.
The current project explores innovative methodologies for making the crop type mapping
process more efficient and cost-effective. A combination of images from drones and
satellites such as Sentinel-2 open up new possibilities for obtaining accurate data from
the ground. WFP is testing the use of drones to capture images from much larger areas,
compared to conventional data collection methods which use enumerators or smartphone
applications which lead to relatively small sample sizes. The drone images are then
classified into different crop types with high or sufficient accuracy. The ultimate goal is to
use the drone data to train the Sentinel-2’s machine learning model to recognise different
types of crops over larger unsampled areas. As a result, this new process combining
artificial intelligence with drone and satellite imagery data will substantially reduce field
work and automate the process of crop type mapping.
• Department/Division Research, Assessment and Monitoring Division
• Project Type/Output: Academic paper, Software tool
• Project updates: A new ground data capture experiment using drones was carried out in
Mozambique in 2021, which resulted in a more extensive field data collection.
• Data Source: Sets of drone images covering about 0.5Km2 each, coupled with smartphone
acquired samples of labelled field perimeters. Sentinel-2 data covering the sampled
areas.
• Technology/Platform: The main technology used to develop the project is a Jupyter
Notebook containing Python code. Keras and TensorFlow libraries power the artificial
347
intelligence behind the project. Finally, the technology includes convolutional neural
networks pre-trained on computer vision datasets such as ImageNet, among others, to
transfer knowledge applicable to classification of drone data.
• Related Sustainable Development Goals (SDGs): SDG 2 – Zero Hunger; SDG – 17
o UN Partners: World Food Program

o Government: National Disaster Management Institute, Ministry of Agriculture,
Mozambique
• Lessons Learned: It is important to collaborate with the local government when

deploying new technologies. Since the early-phase, the project has been implemented
in collaboration with the Government of Mozambique and the Mozambique National
Disaster Management Institute, which have contributed to a safe and effective deployment
of drone technology. Currently, drones are helping local authorities in different activities
including crop mapping, flood modelling, and damage assessments).
• Contact Information- Artur Nowakowski, Rogerio Bonifacio (artur.nowakowski@wfp.org,
rogerio.bonifacio@wfp.org)
Project 9: AI4Human – AI for Humanitarian Applications (DLRR and WFP)
Project 9A: AI for Human
• Project Description: AI for Human is an artificial-intelligence-assisted building damage

detection and classification tool used in the aftermath of natural disasters such as floods,
cyclones and earthquakes. The project was focused on the development of methods
and processing environments to detect changes in infrastructure and buildings by using
machine learning techniques. The main focus is on the demand-driven development
of existing deep learning methods and the provision of software for humanitarian
organisations, to speed up response times in emergencies. WFP is particularly interested
in exploring and testing the use of artificial intelligence-enabled procedures, for instance,
for detecting infrastructure damages caused by natural disasters. WFP will provide
the datasets used for the case study The project will also evaluate and illustrate the
opportunities and and limitations pertaining to the machine learning methods in this
specific technology.
Project 9B: Fusion of Remote Sensing and Web-based data sources
• Project Description: This project explores the systematic development and evaluation of
national and global data and metadata published on the Internet to assess its usability for
the derivation of crisis-related information. National and global databases published on
the Internet and blogs, newsfeeds and social media contributions are a complementary
source of information on satellite imagery and derived products. Therefore, the acquisition
and aggregation of such data and the integrated presentation of these two sources of
information (e.g. in mapping products of the ZKI or web-based services) can greatly
contribute to an overall analysis. The fusion techniques will be developed and tested
grounded on past or ongoing crisis situations. For example, humanitarian emergencies
such as floods in Mozambique in 2019 can be included in the model (from M1). Data on
refugee camps and their dynamic developments will also be added systematically (from
M13). Finally, analyses of food security issues such as post harvest loss estimation in Africa
can also potentially enhance the model.
• Department/Division: Emergency Operations Division
• Project Type/Output: Dataset/Software tool
• Project Duration: 2019- 2021
348

• Project updates: Currently developing algorithm for building damage detection and
classification including road damage assessment.
WFP
• Project Domain: Environment, Human Rights, Telecommunications, Weather,
Humanitarian Emergencies related to natural disaster and/or conflict
• Data Source: Satellite data, labelled training datasets for damaged (and undamaged)
buildings, road, other infrastructure, Twitter data related to crisis information.
• Technology/Platform: Python, Jupyter Notebook, Linux, TensorFlow, NLP
Hunger; SDG 13 – Climate Action
o Private Sector: DLR – German Aerospace Center
• Contact Information: Michael Andrew Manalili (michael.manalili@wfp.org)
Project 10: Fampred

• Project Description: FamPred project is complementary to WFP’s HungerMap LIVE
humanitarian mapping project (https://hungermap.wfp.org).).).). FamPred expands the
HungerMap LIVE’s capabilities to forecast food crises by adding projections on insufficient
food consumption. Artificial intelligence offers new opportunities for forecasting food
insecurity in complex systems, where an approach known as “reservoir computing” is
among the most promising. The project aims to develop a reservoir-computing-based
prediction model using relevant datasets for the identified geographies in HungerMap
LIVE. This model would strengthen WFP’s and national governments’ capacities to predict
crises, deploy resources and prevent food crises outbreaks, ensuring communities at risk
of hunger continue to get the support they need, when they need it.
• Department/Division): INKA
• End Year: 2021
• Project Updates: Refined monthly 30-day provincial and provided 60-day national
forecasts to five countries: Yemen, Syria, Nigeria, Camerooon, and Haiti. The solution
expanded its horizons with new use cases like scenario testing, seasonal forecasts, and
national forecasts.
• Project Domain: Hunger and food security
• Data Source: Satellite data, Economic data, Conflict data
• Technology/Platform: A reservoir computing-based prediction model using well-suited
data sets for the identified geographies
Hunger
o UN Partners: WFP
o Government: DLR
• Links and Multimedia: [2312.00626] Forecasting Trends in Food Security: a Reservoir

Computing Approach (arxiv.org)
349
• Contact Information: Duccio Piovani (duccio.piovani@wfp.org), Amine Baha (amine.

baha@wfp.org)
Project 11: AI Sandbox

• Project Description: WFP’s Artificial Intelligence (AI) Sandbox initiative aims to create a
collaborative environment that provides WFP colleagues and partners with a platform
to experiment, innovate, pilot and scale AI models and AI use cases. As the need for
data-driven solutions to combat hunger grows and AI technology evolves rapidly, the AI
sandbox becomes a crucial tool. It allows us to quickly test and deploy AI systems, giving
WFP teams the power to work more efficiently and make a significant impact on a large
scale.
WFP’s Technology Division (TEC), Innovation and Knowledge Management Division
(INK), and Research Assessment and Monitoring Division (RAM) are collaborating to
enable the scale-up of AI in the organization. Through this collaboration, we strive to
pilot impactful AI use cases that can scale. It encompasses investing in talent by upskilling
and attracting expertise, building a strong foundation to establish an ethical governance
framework, and providing a roadmap for a responsible use. In this direction, we set up a
Responsible AI Task Force to oversee and monitor the implementation of AI initiatives,
ensuring they deliver on their and organization’s goals. On top of that, to ensure we foster
experimentation in AI, a dedidated technology platform and environment, to support AI
pilots, encouraging innovation and exploration in a responsible and safe manner.
• Department/Division: INKA, TEC, RAM
• Project Type/Output: Development Platform
• Reported as part of 2022 compendium on UN AI Activities? No
• Project Domain: Humanitarian AI innovation
• Technology/Platform: Cloud environment
• Related Sustainable Development Goals (SDGs): Any SDG
o UN Partners: WFP
• Links and Multimedia: (1) Publier | Fil d’actualité | LinkedIn

• Contact Information: Jorge Fernandes (jorge.fernandes@wfp.org) , Amine Baha (amine.
baha@wfp.org), Erich Natsubori Sato (erich.natsuborisato@wfp.org)

SDG 1, 2, 3, 4, 13, 17
3. Relevant Link
https://www.wfp.org/
Contact Information
Amine Baha (amine.baha@wfp.org); Kyriacos Koupparis (kyriacos.koupparis@wfp.org);

Bernhard Kowatsch (bernhard.kowatsch@wfp.org); Jean-Martin Bauer (jean-martin.bauer@
wfp.org); Hila Cohen (hila.cohen@wfp.org)
350
World Health Organization
WHO
In recognition of the growing importance of digital health technologies—including AI—the
WHO Member States unanimously adopted the resolution on Digital Health during the 71st
World Health Assembly on 26 May 2018 in Geneva, Switzerland (WHO, 2018). Following this
MS agreed on the Global strategy on digital health 2020–2025 (www.who.int/dhstrategy)
which highlights the importance of artificial intelligence. It is deeply embedded in the 172
implementation actions endorsed by Member States, with specific focus under digital health
governance and human-centred health system. During the opening speech of the 144th session
of the WHO Executive Board on 24 January 2020 in Geneva, Switzerland, the importance
of digital health (and, particularly, AI for health) was reinforced: “the future of health will be
influenced by digital health significantly [and WHO must] embrace it, but at the same time,
WHO should be ahead of the curve in digital health, in order to contribute to global health [...]
working with the International Telecommunications Union to find new ways of using artificial
intelligence to get care to remote communities” (WHO, 2020)
1.1 Background and inception

In response to the accelerating advancements in artificial intelligence (AI) and its potential
impact on healthcare, the World Health Organization (WHO) took an early step, together with
the International Telecommunication Unit (ITU), by establishing the Focus Group on AI for
Health (FG-AI4H) in 2018. This strategic move represented a steppingstone in the intersection of
technology and global health policy to advance in (1) the efforts of benchmarking applications of
AI in health, and (2) to advance the development of guidance documents on ethical, regulatory,
and clinical considerations of development and deployment of AI in health.
The experiences, insights, and collaborative efforts within the FG-AI4H across its 10 pillars
and 25 topic groups set up a positive example that transcended the boundaries of Member
States and united key stakeholders on a shared vision. The FG-AI4H was to serve as pioneering
dedicated to the exploration of AI’s untapped potential within the healthcare landscape. Hence
it was recognized the transformative force AI could exert on healthcare delivery, diagnosis,
and overall efficiency. The FG-AI4H also acknowledged the potential risks associated with
the use of AI in health, including issues of equity, privacy, and the digital divide. As such, the
group set out to not only explore the technical capabilities of AI but also to champion ethical
advancements. Two publications are paramount:
• Ethics and governance of AI for health. WHO's commitment to ethical and responsible AI
in health is manifested in this guidance document. This comprehensive resource provides
consensus principles tailored for AI developers, Ministries of Health, and clinicians.
351
Through interactive workshops and training courses, WHO has actively promoted
the implementation of these guidelines, emphasizing the pivotal role of ethics in the
successful deployment of AI in healthcare.
• Generating evidence for AI-based medical devices. The framework is a pioneering
initiative by WHO intersecting a whole AI lifecycle approach in medical devices. Tailored
for developers, researchers, and policymakers, this framework offers comprehensive
guidance on evidence generation specific to health areas like cervical cancer screening.
By setting clear standards, this framework lays the groundwork for the responsible
introduction of new and emerging AI technologies in the healthcare landscape.
The establishment of the FG-AI4H underscored WHO's commitment to staying at the forefront
of technological advancements, ensuring that Member States were equipped to harness the
potential benefits of AI while navigating its complexities. This forward-looking move signaled
a broader shift in global health policy, recognizing the need to adapt rapidly in the landscape
of healthcare technology. The establishment of the FG-AI4H laid the essential groundwork for
the subsequent formation of the Global Initiative on AI for Health (GI-AI4H) in 2023.
1.2 The Global Initiative on AI for Health (GI-AI4H)

The launch of the GI-AI4H is a landmark moment of harnessing the potential of AI in health. This
transition was and is possible thanks to the commitment of three key United Nations agencies,
the World Health Organization (WHO), the International Telecommunication Union (ITU), and
the World Intellectual Property Organization (WIPO) to enable, facilitate and implement the
adoption of AI-based technologies for health.
The setting up of the GI-AI4H moves beyond the traditional boundaries of individual
organizations. It is a testament to the recognition that addressing the complex challenges
and opportunities presented by AI in health requires a multidimensional and collaborative
approach. By pooling their joint expertise and resources the initiative aims at becoming a
powerful force aligned to two core objectives:
• Firstly, the GI-AI4H is dedicated to ensuring innovation and global accessibility to AI

solutions across healthcare services. In leveraging AI, WHO envisions a future where
innovative technologies bridge gaps in healthcare delivery, making essential services
accessible to everyone, especially in low- and middle-income countries (LMIC).
• Secondly, the promotion of ethical and equitable advancements in health AI technology.
Recognizing the potential risks and disparities associated with AI from its inception in
policymaking to the implementation on the ground, the initiative places a strong emphasis
on ensuring that technological progress is accompanied by principles of fairness,
inclusivity, and ethical considerations. This commitment underscores the intention to
create a healthcare landscape where the benefits of AI are equitably distributed, leaving
no one behind.
1.3 Governance structure of the GI-AI4H

The GI-AI4H aims to create a harmonized ecosystem and avoid the vertical fanged approach.
The governance structure includes a Steering Committee (with equal representation of the
founding UN partners and will be the final decision-making authority); Secretariat (as the
functional and operational team cross-cutting across the founding organizations with devoted
resources and structure); Partners (who are donors and champion countries who will bring
resources to support the three key functions); and Informal Expert Groups (entail screened
individuals with expertise in the relevant topic areas).
352
GI-AI4H governance structure
WHO
Roles and Responsibilities
To support the management of the Steering Committee and Secretariat, under the GI-AI4H
governance structure, the founding partners’ roles and responsibilities are:
• WHO: Leads in normative technical guidance, policies, and standards; facilitates resource
and knowledge sharing; and supports through collaborative research.
• ITU: Facilitates resource and knowledge sharing; strengthens knowledge sharing on AI,
health and ICT policy; supports collaborative research.
• WIPO: Focuses on policies around intellectual property (IP) and AI for health; facilitates
knowledge sharing advocates for ethical AI implementation.
1.4 Three pillars of work at a glance

The GI-AI4H is structured around the three key functions, namely, Enable, Facilitate, and
Implement:
1) The Enable function focuses on developing policy, guidance, legislations, norms and
standards, and the benchmarking framework along with the evidence to support Member
States in utilizing AI for health care systems and services.
o Guidance on the use of AI in crosscutting topics and governance areas.

o Policy + briefs on application of AI for health in thematic areas.
o Toolkits and handbooks for implementation and technical support.
2) The Facilitate function promotes knowledge sharing between countries and partners,
fostering a community of practice.
o Resource mobilization and funding partnership development.

o Knowledge mobilization and communities of practice (global registry of networks for
knowledge sharing of academics, AI in health experts, policymakers, designers and
more).
3) The Implement function supports countries in taking evidence-based AI for health

programs to strengthen health systems and services. This function supports and
emphasizes the creation of sustainable implementation models.
353
o Technical assistance and implementation of toolkits, policies guidance, handbooks

and use cases on AI for health.
o Workshop + curriculums on AI for health.
Three key functions of the GI-AI4H
1.5 Unwrapping each pillar of work

The Enablement pillar serves as cornerstone of the GI-AI4H
Development and release of several technical guidance and policy briefs addressing pressing
global health needs, research gaps and governance challenges in the use of AI for health.
These normative technical products can serve as an overarching guide for Member States
and partners to design, develop, regulate, and deploy AI solutions that bring health benefits
to everyone, everywhere. This involves enabling the formulation of policies that navigate
the complex landscape of AI, ensuring alignment with global health standards and ethical
considerations. Two publications are paramount of the enabling function:
354
• Regulatory considerations on AI for health: launched

in October 2023 and articulating non-binding
recommendations for the health ecosystem, including
WHO
Ministries of Health (MoH), developers, manufacturers,
and the whole tech industry. The document
clusters eighteen recommendations for regulatory
considerations in six topic areas: 1) documentation
and transparency; 2) risk management and AI systems
development lifecycle approaches; 3) analytical and
clinical validation; 4) data quality; 5) engagement and
collaboration; 6) privacy and data protection. Addressing
crucial aspects of AI implementation, the document on
“Regulatory Considerations in AI for health” covers key
topics including documentation, transparency, total
product lifecycle approach, intended use, analytical &
clinical validation, privacy, and data protection. Serving
as a comprehensive resource, this document is intended
for all stakeholders involved in AI solutions in the health
sector.
• Ethics and governance of AI for health – Large Multi-
Modal Models (LMMs): The guidance outlines
recommendations for assisting Member States and
tech developers in sharing the benefits and challenges
associated with using LMMs for health. The publication
provides a good overview of tailored risks and
challenges in various uses of LMMs in healthcare: 5
broad applications - 1) diagnosis and clinical care, 2)
patient-guided use, 3) clerical and administrative tasks,
4) medical and nursing education, 5) scientific research
and drug development.; Tailored risks to health
systems: these are risks associated with the use of LMMs
in healthcare; LMM life cycle approach: finely details the
governance behind the development, provision and
deployment of LMMs as a basis for enshrining ethical
principles and human rights obligations through laws,
policies and regulation. Finally, it provides a visually-
sound "checklist" as a summary of the three points
above in which it frames the actions of both developers
and governments across each development, provision
or deployment phases considering their own set of risks
to be assessed during and thereafter.
The Facilitation pillar emerges as the dynamic and collaborative force within the GI-AI4H.
Convene experts to establish informal expert groups and lead technical global consultations
to assess and analyze current uses of AI for health and advise or recommend actions or steps
to ensure the use of AI for health benefits healthy living and well-being in all countries and
at all ages. Through establishing a global network of experts, the GI-AI4H will promote the
development of normative guidance around governance and policy, foster the development of
an implementation toolkit of AI for health, and support key. By cultivating a dynamic community,
this pillar contributes to the collective effort of advancing health AI on a global scale. This
pillar places Member States and all stakeholders of the health community to facilitate and
accelerate the seamless implementation of the normative technical products. This involves
355
creating platforms for the dissemination of insights, best practices, and lessons learned in the
field of health AI.
The Implementation Pillar embodies the practical application of AI advancements and

recommendations across Member States.
Support guidance implementation by piloting the implementation concept of “AI for Public
Health” in countries and relevant stakeholders, to accelerate countries’ good governance of
digital health, and the appropriate use of AI and other frontier technologies for health.
356
WHO
1.6 First meeting of the GI-AI4H in Riyadh
The first workshop and (inaugural) meeting of the ITU/WHO/WIPO Global Initiative on AI for
Health (GI-AI4H) were held from 30 October 2023 until 2 November 2023 in Riyadh, Saudi
Arabia, at the premises of the King Abdulaziz City for Science and Technology (in the building
"The Garage" on 30 Oct., and in "The Innovation Tower" on the remaining days).
The meeting welcomed in total nearly 400 people coming from 42 countries across 6 regions,
with an adequately large number of attendees (42%) joining virtually. The participants and
speakers consist of a variety of the portfolios e.g., researchers, policy makers and health
practitioners. In addition, more than 100 participants expressed interests in the Ideathon, an
independent session running in parallel to the 4-day meeting agenda.
The four-day meeting featured more than dozens of insightful presentations, four multiple-
disciplinary panels, two breakout working sessions as well as a persona role play case study
loaded with valuable information and innovative ideas for GI-AI4H's future. The meeting was
key in identifying collaborative opportunities and strategies for advancing AI in health within the
GI-AI4H framework. In conclusion, the Global Initiative on AI for Health stands as a pioneering
force in the responsible and collaborative use of AI.

All SDGs.
3. Relevant Link
WHO documents
• Guidance: Ethics and governance of artificial intelligence for health
o Full document https://apps.who.int/iris/handle/10665/341996

o Executive summary https://apps.who.int/iris/handle/10665/350567
357
• Guidance: Generating evidence for artificial intelligence-based medical devices: a

framework for training, validation and evaluation. https://apps.who.int/iris/handle/10665/
349093
• Guidance: Regulatory considerations on artificial intelligence for health. https://iris.who
.int/handle/10665/373421
• Policy brief: Ageism in artificial intelligence for health. https://apps.who.int/iris/handle/
10665/351503
• Press release: WHO calls for safe and ethical AI for health. https://www.who.int/news/
item/16-05-2023-who-calls-for-safe-and-ethical-ai-for-health
• Ethics and governance of AI in health - guidance for large multi-modal models: https://
www.who.int/publications/i/item/9789240084759
• Press release: WHO releases AI ethics and governance guidance for large multi-modal
models: https://w ww. who. int/n ews/i tem/1 8- 01- 2024- who- releases- ai- ethics- and
-governance-guidance-for-large-multi-modal-models
• Press release: WHO outlines considerations for regulation of artificial intelligence for
health. https://www.who.int/news/item/19-10-2023-who-outlines-considerations-for
-regulation-of-artificial-intelligence-for-health
Global strategy on digital health
• English:https://apps.who.int/iris/bitstream/handle/10665/344249/9789240020924-eng
.pdf
WHO online course
• Introduction: Ethics and governance of artificial intelligence for health (based on the
ethics guidance). https://openwho.org/courses/ethics-ai
For more information, please, contact us at: ai@who.int
Alain Labrique (Labriquea@who.int), Sameer Pujari (pujaris@who.int), Ursula Yu Zhao (Zhaoy@

who.int), and Shada Alsalamah (Alsalamahs@who.int)
358
World Intellectual Property Organization
WIPO
Project 1: WIPO Translate

• Project Description: "WIPO Translate is a market-leading translation software for
specialized text. Originally created to translate patent documents, WIPO Translate can
also be adapted and customized to other technical Project Domains. Once trained in a
specific subject area, WIPO Translate has been shown to out-perform other paid and free
translation tools."
• Project Status: Full fledged development
• Project Domain: Intellectual Property
• Data Source: Translation database
• Technology/Platform: Supervised learning and Natural Language Processing (NLP)
Infrastructure
• Relevant Links and Multimedia: https://www.wipo.int/wipo-translate/en/
• Lessons Learned: Project Domain adaptation (we need different models for different
Project Domain). E.g. patent translation is totally different than conference translation.
• Contact Information: Sandrine Ammann (sandrine.ammann@wipo.int)
Project 2: WIPO Speech-to-Text

• Project Description: WIPO speech-to-text (S2T) is a homemade transcription tool for
conferences. It generates an automatic transcript, which becomes available a few minutes
after the meeting. This transcript can be further cascaded through WIPO Translate in
order to generate UN-6 language reports (always synchronized with the original video).
It has been successfully used in other organizations.
• Data Source: Conferences
• Technology/Platform: Supervised learning for automatic speech recognition
Infrastructure
359
• Relevant Links and Multimedia: https://www.wipo.int/s2t/

• Lessons Learned: Adapt to various speakers: in international conferences native and non-
native accents is a challenge.
Project 3: WIPO Brand Image Search

• Project Description: Perform a trademark search by text or image in brand data from
multiple national and international sources, including trademarks, appellations of origin
and official emblems.
• Project Status: Full-fledged development
• Data Source: Global Brand Database
• Technology/Platform: Computer Vision
Infrastructure
• Relevant Links and Multimedia: https://www3.wipo.int/branddb/en/
Project 4: International Patent Classification (IPC)

• Project Description: The International Patent Classification (IPC) provides for a hierarchical
system of language independent symbols for the classification of patents and utility
models according to the different areas of technology to which they pertain.
• Project Status: Full-fledged development
• Data Source: Patents
• Technology/Platform: Classification and machine learning.
Infrastructure
• Relevant Links and Multimedia: https://www.wipo.int/classifications/ipc/en/
Project 5 : WIPO Goods & Services Terms Explorer

• Project Description: A tool designed to provide assistance to trademark applicants when
selecting appropriate Goods and Services terms and their associated Nice classification in
different languages during the filing process. This tool also provides support to trademark
examiners in IP offices in their validation of trademark applications.
360

Infrastructure
• Relevant Links and Multimedia: Global Goods & Services Terms Explorer
WIPO
Project 6: Vienna Classification Assistant

• Project Description: The Vienna Classification Assistant is an AI-driven tool developed by
WIPO to support the use of the Vienna Classification system. This system helps classify
figurative elements of trademarks by automatically suggesting appropriate Vienna
Classification codes. The Vienna Classification Assistant enhances the accuracy and
efficiency of trademark classification processes.
Infrastructure
• Relevant Links and Multimedia: Vienna Classification Assistant
Project 7: WIPO Conversation on IP and Frontier Technologies

• Project Description: The objective of the WIPO Conversation is to provide Member States
with an opportunity to exchange views on various topics regarding frontier technologies,
including AI, and to formulate questions with respect to the possible impact of frontier
technologies on the IP system.
• Project Type/Output: Conference
• Project Updates: WIPO has expanded the scope of the Conversation to cover frontier
technologies as a whole. While AI remains an important component of our work, it will
not be its sole focus either. WIPO will also ensure that these sessions have a practical, as
well as conceptual emphasis – reflecting the fact that frontier technologies are piecing
together a new kind of global economy. The 2022 sessions of the WIPO Conversation
covered the Metaverse and Generative AI.
Infrastructure
• Relevant Links and Multimedia: https://www.wipo.int/about-ip/en/frontier_technologies/
frontier_conversation.html
• Contact Information: Alica Daly (alica.daly@wipo.int)
Project 8: AI and IP Strategy Clearing House

• Project Type/Output: Clearing House
361

Infrastructure; SDG 17 – Partnerships for the Goals
• Relevant Links and Multimedia: https://www.wipo.int/about-ip/en/artificial_intelligence/
policy.html#clearing_house
• Contact Information: Alica Daly (alica.daly@wipo.int)
Project 9: ITU, WHO, WIPO Global Initiative on AI for Health

• Project Type/Output: Interagency cooperation to develop tools and resources to support
the use of AI for health
• Related Sustainable Development Goals (SDGs): SDG 3- Good Health and Well-being;
SDG 9 – Industry, Innovation, and Infrastructure; SDG 17 – Partnerships for the Goals
• Relevant links and multimedia: https://www.itu.int/en/ITU-T/focusgroups/ai4h/Pages/
default.aspx
• Contact Information: Nancy Pignataro (nancy.pignataro@wipo.int)

SDG 3, 9, 17
3. Relevant Links
https://www.wipo.int/portal/en/index.html
Contact Information
Alica Daly (alica.daly@wipo.int) and Nancy Pignataro (nancy.pignataro@wipo.int)
362
World Meteorological Organization
WMO
Project 1: WMO RA VI-led event on Artificial Intelligence to support EW4All

• Project Description: The WMO RA VI-led event on Artificial Intelligence to support EW4All
on May 27, 2023, focused on integrating AI and machine learning in meteorology to
improve early warnings and tackle climate change. It featured notable speakers from
ECMWF, EUMETSAT, and tech giants, discussed WMO's AI initiatives, insights from the
European Meteorological Infrastructure, and relevant case studies. The agenda included
machine learning's application and impact on observation and weather forecasting,
highlighting a collaborative approach in the meteorological community to leverage new
technologies.
• Department/Division: WMO Regional Office for Europe
• Project Type/Output: meeting
• Project Status: Finished
• Project Domain: Environment, Early warning
The ITU/WMO/UNEP Focus Group on AI and Natural Disaster stresses the importance
of adapting to evolving technologies to meet the shifting needs and priorities in the
field of natural disaster management. The European Centre for Medium-Range Weather
Forecasts (ECMWF) notes substantial involvement from commercial sectors in advancing
machine learning. EUMETSAT is committed to delivering the highest quality of data and
adopts a proactive stance in anticipating and understanding future requirements. The
significant role of AI and machine learning is acknowledged, with their development
shaping our strategic choices. Machine learning is crucial for enhancing nowcasting and
short-term weather predictions, offering capabilities like automatic feature detection and
a variety of applications including air quality assessments. Huawei and Google DeepMind
presented their data-driven weather forecast models, showcasing the remarkable
potential of AI in weather forecasting.
• Related Sustainable Development Goals (SDGs): sSDG 11 – Sustainable Cities and
Communities; SDG 13 – Climate Action; SDG 17 – Partnerships for the Goals
o ITU/WMO/UNEP Focus Group on AI and Natural Disaster Management

o ECMWF
o EUMETSAT
o Meteo-France
o Huawei Inc
363
o DeepMind, Google
o Microsoft
o UK Met Office
o Nvidia
• Relevant Links and Multimedia: https://wmo.int/news/media-centre/big-tech-and-artificial

-intelligence-can-support-early-warnings-all
• Lessons Learned: .Artificial Intelligence (AI) has the potential to significantly enhance
weather observation, forecasting, and the dissemination of early warnings, thereby
effectively supporting early warning services. The private sector plays a pivotal role in
realizing the goal of achieving 'Early Warnings for All.' This can be achieved by harnessing
the extensive reach of major technology companies and the capabilities of Artificial
Intelligence..
• Contact Information: Natalia Berghi (nberghi@wmo.int),Kanghui Zhou (kzhou@wmo.int)
Project 2: The Mediterranean and Pan-European Forecast and Early Warning System
against Natural Hazards (MedEWSa) project (Horizon Europe project funded by the
European Commission
MedEWSa is an Innovation action (IA) project led by the World Meteorological
Organisation (WMO) and a consortium of 30 partners. Building on existing tools,
MedEWSa will develop a fully integrated impact-based multi-hazard early warning
system (EWS). It will achieve advances beyond the state-of-the-art in the areas of AI-
based decision support solutions for improved impact prediction, methods for impact
prediction and early warnings, modelling and prediction of multiple hazard dynamics and
their interdependencies. MedEWSa will explore hybrid methods combining data-driven
approaches and the laws of physics, using tools learned from observations and physics-
based model results. Ultimately, it will deploy an analysis to provide integrated, accurate,
and rapid early warnings for the pan-European-Mediterranean-African region. MedEWSa
will directly contribute to the United Nations’ Sustainable Development Goals and align
with the UN’s Early Warnings for All initiative. The project will also develop innovative
financial solutions through risk transer to capital margkes.MedEWSa will also set out
a roadmap for contributing to different standardisation activities to identify significant
opportunities to push contributions into future standards, pre-normative activities, and
open collaborative development environments. In that sense, MedEWSa will benefit
from the participation of academia such as the Justus Liebig University of Giessen (JLU),
WMO, SMEs, national meteorological and hydrological services, national ministries, first
responders and HHI, who actively participate in international standardisation activities,
such as the ITU/WMO/UNEP Focus Group on AI for Natural Disaster Management.
• Department/Division: WMO Science and Innovation Department
• Project Type/Output: research project
• Project Domain: Weather, Environment, natural disaster
• Project Updates:The MedEWSa was launched in Athens, Greece in November 2023.
The kick-off meeting was hosted by the Ministry of Digital Governance of Greece, the
event not only served as a collaborative platform but also underscored the significant role
played by the WMO Regional Office for Europe in steering the project's implementation.
The meeting structure featured four specialized groups led by experts in disaster risk
knowledge, hazard monitoring, warning dissemination, and response capabilities,
collectively aiming to align MedEWSa's scientific and techno-economic developments
364
with the overarching goals of the WMO's EW4All Initiative.This event facilitated
comprehensive discussions, reinforcing the integration of MedEWSa's goals with the
broader vision of the EW4All Initiative, while also exploring synergies with SEE-MHEWS..
WMO
Infrastructure; SDG 11 – Sustainable Cities and Communities; SDG 13 – Climate Action;
o National Observatory of Athens

o Fondazione Centro Euro-Mediterraneo sui Cambiamenti Climatici
o Sveriges Meteorologiska och Hydrologiska Institut
o Slovak Hydrometeorological Institute
o European Centre for Medium Range Weather Forecasts
o Hellenic Ministry of Defence
o Centro Internazionale Di Monitoraggio Ambientale – Fondazione
o Institut National de Recherche pour l'Agriculture, l'alimentation et l'Environnement
o International Red Cross/Red Crescent Centre on Climate Change and Disaster
Preparedness
o Justus-Liebig-Universitaet Giessen
o Barcelona Supercomputing Center-Centro Nacional de Supercomputación
o Fraunhofer Heinrich-Hertz-Institut – Fraunhofer HHI
o GeoSphere Austria
o Mitiga Solutions
o Internet of Things applications and Multi-Layer development
o Kajo s. r. o.
o Convergence
o Ministry of Digital Governance (associated)
o GRNET
o Georgia National Environmental Agency
o Region of Attica Civil Protection Department
o European Central Bank (associated)
o Centro Previsione e Segnalazione Maree del Comune di Venezia
o Egyptian Meteorological Authority
o Italian Meteo Agency
o Department of Interior of Generalitat de Catalunya
o Ethiopian Forestry Development
o Pau Costa Foundation
o South-East European Multi-Hazard Early Warning Advisory System (SEE-MHEWS-A)
represented by the Israel Meteorological Service (associated)
o https://wmo.int/media/news/wmo-leads-new-research-project-early-warning-systems
-mediterranean
• Lessons Learned: The MedEWSa was launched in November 2023. The MedEWSa
project emphasizes the importance of research and multi-stakeholder collaboration in
enhancing Mediterranean and European countries' operational EWS capabilities. It aims
to enhance collaboration, research, innovation, and the dissemination of knowledge and
technologies in support of EU policies addressing global challenges.
• Contact Information: Jon Cox (jcox@wmo.int), Jürg Luterbacher(jluterbacher@wmo.int)
365

SDG 9, SDG 11, SDG 13, SDG 17
3. Related Links
https://public.wmo.int/en
Contact Information
Kanghui Zhou(kzhou@wmo.int)
366
International Telecommunication Union
Telecommunication Development Bureau
Place des Nations
CH-1211 Geneva 20
Switzerland
ISBN: 978-92-61-39001-3
9 789261 390013
Published in Switzerland
Geneva, 2024

S Gen Unact 2023 PDF e

Uploaded by

Copyright:

Available Formats

S Gen Unact 2023 PDF e

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

S Gen Unact 2023 PDF e

Uploaded by

Copyright:

Available Formats

ITUPublications International Telecommunication Union

United Nations Activities on

978-92-61-39001-3 (Electronic version)

Please consider the environment before printing this report.

2023 will go down in history as the year that generative artificial

Companies began to rethink their business models amid projections

Key tracks and trends

Number of projects reporting work on a specific SDG

Number of projects reporting collaborations with partner type

other international organization

0 20 40 60 80 100 120 140

Proportion of project collaboration by SDG

Project Outputs by SDGs

(Doughnut size = number of projects)

Project Status by SDG

Proportion of Projects by Reporting Status

completed development ongoing other

Proportion of projects by SDGs by current status

completed development ongoing other

Comprehensive Nuclear-Test-Ban Treaty Organization��������������������������������������������������������� 1

1. Description of Activities on AI������������������������������������������������������������������������������� 1

The Food and Agriculture Organization����������������������������������������������������������������������������������� 4

1. Description of Activities on AI������������������������������������������������������������������������������� 4

International Atomic Energy Agency���������������������������������������������������������������������������������������12

1. Description of Activities on AI����������������������������������������������������������������������������� 12

International Fund for Agricultural Development�����������������������������������������������������������������25

1. Description of Activities on AI����������������������������������������������������������������������������� 25

1. Description of Activities on AI����������������������������������������������������������������������������� 36

1. Description of Activities on AI����������������������������������������������������������������������������� 44

International Maritime Organization���������������������������������������������������������������������������������������45

1. Description of Activities on AI ���������������������������������������������������������������������������� 45

International Organization for Migration������������������������������������������������������������������������������������ 47

1. Description of Activities on AI��������������������������������������������������������������������������������� 47

1. Description of Activities on AI��������������������������������������������������������������������������������� 50

International Telecommunication Union������������������������������������������������������������������������������������� 52

1. Description of Activities on AI��������������������������������������������������������������������������������� 52

Office of the United Nations High Commissioner for Human Rights������������������������������������� 71

1. Description of Activities on AI��������������������������������������������������������������������������������� 71

Office of the Secretary-General's Envoy on Technology���������������������������������������������������������� 80

1. Description of Activities on AI��������������������������������������������������������������������������������� 80

1. Description of Activities on AI��������������������������������������������������������������������������������� 81

United Nations Human Settlements Programme���������������������������������������������������������������������� 92

1. Description of Activities on AI��������������������������������������������������������������������������������� 92

United Nations World Tourism Organization����������������������������������������������������������������������������� 98

1. Description of Activities on AI �������������������������������������������������������������������������������� 98

1. Description of Activities on AI������������������������������������������������������������������������������� 101

The Joint United Nations Programme on HIV/AIDS���������������������������������������������������������������� 104

1. Description of Activities on AI������������������������������������������������������������������������������� 104

United Nations Conference on Trade and Development������������������������������������������������������� 106

1. Description of Activities on AI ������������������������������������������������������������������������������ 106

United Nations Department of Economic and Social Affairs������������������������������������������������� 110

1. Description of Activities on AI ������������������������������������������������������������������������������ 110

United Nations Development Programme������������������������������������������������������������������������������� 118

1. Description of Activities on AI������������������������������������������������������������������������������� 118

United Nations Department of Political and Peacebuilding Affairs and

1. Description of Activities on AI������������������������������������������������������������������������������� 149

United Nations Economic Commission for Europe����������������������������������������������������������������� 152

1. Description of Activities on AI ������������������������������������������������������������������������������ 152

1. Description of Activities on AI ������������������������������������������������������������������������������ 163

Comprehensive Nuclear-Test-Ban Treaty Organization�� 1

1. Description of Activities on AI�� 1

The Food and Agriculture Organization�� 4

1. Description of Activities on AI�� 4

International Atomic Energy Agency��12

1. Description of Activities on AI�� 12

International Fund for Agricultural Development��25

1. Description of Activities on AI�� 25

1. Description of Activities on AI�� 36

1. Description of Activities on AI�� 44

International Maritime Organization��45

1. Description of Activities on AI �� 45

International Organization for Migration�� 47

1. Description of Activities on AI�� 47

1. Description of Activities on AI�� 50

International Telecommunication Union�� 52

1. Description of Activities on AI�� 52

Office of the United Nations High Commissioner for Human Rights�� 71

1. Description of Activities on AI�� 71

Office of the Secretary-General's Envoy on Technology�� 80

1. Description of Activities on AI�� 80

1. Description of Activities on AI�� 81

United Nations Human Settlements Programme�� 92

1. Description of Activities on AI�� 92

United Nations World Tourism Organization�� 98

1. Description of Activities on AI �� 98

1. Description of Activities on AI�� 101

The Joint United Nations Programme on HIV/AIDS�� 104

1. Description of Activities on AI�� 104

United Nations Conference on Trade and Development�� 106

1. Description of Activities on AI �� 106

United Nations Department of Economic and Social Affairs�� 110

1. Description of Activities on AI �� 110

United Nations Development Programme�� 118

1. Description of Activities on AI�� 118

1. Description of Activities on AI�� 149

United Nations Economic Commission for Europe�� 152

1. Description of Activities on AI �� 152

1. Description of Activities on AI �� 163

United Nations Educational, Scientific and Cultural Organization�� 173

1. Description of Activities on AI �� 173

United Nations Framework Convention on Climate Change�� 200

1. Description of Activities on AI �� 200

1. Description of Activities on AI�� 202

United Nations High Commissioner for Refugees�� 218

1. Description of Activities on AI�� 218

United Nations International Computing Centre�� 227

1. Description of Activities on AI�� 227

1. Description of Activities on AI �� 234

United Nations Interregional Crime and Justice Research Institute�� 251

1. Description of Activities on AI�� 251

1. Description of Activities on AI �� 263

United Nations Industrial Development Organization�� 266

1. Description of Activities on AI�� 266

United Nations Institute for Training and Research�� 278

1. Description of Activities on AI�� 278

United Nations Joint Staff Pension Fund�� 283

1. Description of Activities on AI�� 283

United Nations Office for the Coordination of Humanitarian Affairs�� 286

1. Description of Activities on AI�� 286

United Nations Office for Counter Terrorism�� 291

1. Description of Activities on AI�� 291

United Nations Office for Disarmament Affairs�� 294

1. Description of Activities on AI�� 294

United Nations Office on Drugs and Crime�� 298

1. Description of Activities on AI�� 298

1. Description of Activities on AI�� 302

United Nations Research Institute for Social Development�� 306

1. Description of Activities on AI �� 306

1. Description of Activities on AI�� 308

1. Description of Activities on AI�� 321

1. Description of Activities on AI�� 323

1. Description of Activities on AI�� 335

1. Description of Activities on AI�� 351

World Intellectual Property Organization�� 359

1. Description of Activities on AI�� 359

1. Description of Activities on AI�� 363