Generative AI

for lawyers
Working smarter and faster – while complying with legal
professional obligations in Australia and New Zealand

October 2023

1
Contents

Foreword 3

Introduction 4

AI in the legal profession: 5

from promise to reality

The opportunity for lawyers 7

Current regulatory landscape 16

Professional conduct rules 19

Maintaining confidentiality 20

What’s next? 21

2
Foreword
Artificial intelligence (AI) is revolutionising the Less than a year after ChatGPT exploded onto the
arena of professional services. Lawyers are chief market, a new wave of generative AI solutions
among those gaining opportunities and facing is proliferating. It’s a little daunting to think that
disruption from the emergence of generative these are early days. At Microsoft, we believe
AI tools – which process language and generate it crucial for lawyers to quickly get acquainted
useful text outputs. with the benefits and opportunities that these
technologies present. If machines can increasingly
The ability to generate content – including client produce legal information, and anyone can
correspondence, contract terms, pleadings, hypothetically get “advice” from typing their
legal advice memos and other documentation query into a prompt.
– represents an incredible opportunity for the
legal profession to safely, securely and responsibly
What does this mean for the profession’s future?
speed up routine legal work.
How can law firms and in-house counsel adjust
their business models and safeguard their
unique value? What new skills will be needed
by law graduates and experienced lawyers
alike? And how will this affect the daily duties of
everyone from junior lawyers to senior partners,
and in-house counsel?

In this paper, our goal is to inform lawyers about

the latest AI capabilities and their implications
for professional compliance requirements.
We highlight early industry leaders in generative
AI innovation and examine the status of regulation
in Australia, New Zealand and around the world.

I hope you will uncover insights for your legal

practice and start to use the new generative
AI tools.

Clayton Noble
Head of Legal, Microsoft Australia
and New Zealand

3
Introduction
Today’s lawyers are accustomed to using new deliver smarter legal work, faster. Popularised
tools and technologies to augment their work. initially by OpenAI’s generative AI chatbot
The COVID-19 pandemic accelerated the legal ChatGPT, a new wave of solutions is entering
industry’s digital transformation. Firms embraced the market, enabling professionals to quickly
virtual practices, using cloud-based workflow and produce more content. The technology is still
collaboration tools to communicate and complete in its early days. Currently, Australia and New
tasks. Even the court systems moved online. Zealand have no AI-specific legal professional
regulation, though professional associations and
Yet technology adoption is also being driven by regulators may soon offer preliminary guidance.
commercial imperatives to deliver legal services Legal practice and in-house legal use cases are
faster, more efficiently and at a higher quality still emerging. As so much of a lawyer’s role
in a rapidly shifting market. Large firms, small involves creating, researching, analysing, applying
practices and even in-house corporate teams must or summarising text, the practice of the legal
look to adapt to the rapid change in technology, profession will change.
which will allow for the delivery of legal services
at a faster and more efficient pace whilst also Generative AI already offers lawyers a ‘copilot’
delivering the level of quality expected of the to handle routine documentation and
legal profession. administrative tasks, allowing them to refocus on
the more strategic, high-level and interpersonal
requirements of their role. However, the
technology also comes with potential risks and
ethical issues, which can be managed as will be
highlighted in this paper. As one of the pioneers
of generative AI, as well as a global evangelist for
Artificial intelligence (AI) tools automating the responsible AI, Microsoft provides this paper to
processes of document discovery, due diligence demystify the technology, highlight the benefits to
reviews and standard contract execution have the legal profession and foreshadow changes that
long been widely used in the legal profession. are likely to come.
Sophisticated data analytics and machine learning
are helping legal teams find needles in haystacks No matter what the future holds, advanced human
and make better, faster strategic decisions. Yet capabilities, empathy and insight will continue
lawyers always must balance speed with accuracy to be needed in the legal profession. But viewing
and reliability, remembering their obligations to the AI revolution from the sidelines is not a viable
exhibit high competence and diligence. option. With appropriate generative AI system
selection and governance, lawyers can safely,
Now the arrival of generative AI tools in 2023 securely and responsibly use these technologies in
offers another incredible opportunity for lawyers a way that is fully consistent with their professional
to again increase their productivity – to effectively obligations.

4
AI in the legal profession:
from promise to reality

Over the past decade, lawyers have used

dedicated AI e-discovery tools to quickly search
large batches of documents. In addition, in large
merger and acquisition (M&A) transactions, AI has
To a degree the future must remain unknown. been used as part of the due diligence process
Artificial intelligence and its effect on the courts, the to review financial records, legal documents,
profession and the law will change the landscape of agreements and any other relevant data.
life in ways we cannot predict.
Predictive coding, natural language processing
James Allsop
and active machine learning have made it easier
Former Chief Justice
to sift through material – going far beyond the
Federal Court of Australia1
capabilities of a manual keyword search. A wave
of bespoke contract management solutions has

Though sometimes used in legal practice as a allowed lawyers to quickly identify key clauses

catch-all term to describe any form of process or and scan for legal issues and risks at a much faster

workflow automation, AI refers to technologies pace than manual review. But until now, many

that perform tasks normally requiring human AI systems and tools have been rudimentary.

cognition, decision-making and judgment. Early In relation to this, Caryn Sandler, Partner and

applications of AI among lawyers centred around Chief Knowledge and Innovation Officer at

data analytics, including making predictions based Gilbert+Tobin, estimates that they led to

15% - 20%
on data or searching large volumes of information.
AI has been present in legal database services such
as LexisNexis, platforms like Nuix, Relativity and
Ringtail, and general search engines such as Bing Efficiency gains at best in the absence of
human review.
and Google.

The promise was hyped up, but the

reality didn’t follow suit.
Caryn Sandler

1
Chief Justice James Allsop, “The Role and Future of the Federal Court within the
Australian Judicial System”, 8 September 2017, Speech to 40th Anniversary of the
Federal Court of Australia Conference

5
Rise of generative AI

Generative AI represents a step-change in artificial

intelligence. The term refers to AI systems capable
of generating new content, such as text, video,
voice or images, in response to human prompts.
This distinguishes it from early AI capabilities
Fundamentally the law is centred on language. And
that involved processing information. The Large
that’s why we can very quickly determine that this
Language Models (LLMs) underpinning the
technology will impact the way we practise law
technology have been trained on a vast amount of
text data. Having an enhanced understanding of Greg Dickason
linguistic, creative and deductive patterns, enables Managing Director of LexisNexis5
AI tools to produce new written works in a matter
of seconds.
For lawyers, including in-house, opportunities to
It is this extraordinary potential that led ChatGPT automate legal work through AI can range from
in late 2022 to acquire 100 million users in just addressing administrative burdens to assisting
over two months, making it one of the fastest- on substantive legal work. Examples include
growing consumer software applications in evaluating information to determine compliance
history. Given the reputation of lawyers as “word with legal, regulatory and corporate standards,
merchants” , studies have found that the legal automating scheduling meetings, estimating time
profession is one of the industries most ripe with to complete legal work based on past experiences
opportunities and also exposed to disruption. and transcribing and summarising the content of
both in-person and virtual meetings.

According to one report,

44%
of legal work stands
to be automated by
generative AI3

A separate analysis by Microsoft and the

Tech Council of Australia has found that

10% 32%
of a solicitor’s tasks augmented by
could be automated generative AI4

2
“AI is coming for lawyers again”, New York Times, 10 April 2023, https://www.nytimes.com/2023/04/10/technology/ai-is-coming-for-lawyers-again.html
3
Goldman Sachs Economics Research, 26 March 2023, https://www.key4biz.it/wp-content/uploads/2023/03/Global-Economics-Analyst_-The-Potentially-Large-Effects-of-Artificial-Intel-
ligence-on-Economic-Growth-Briggs_Kodnani.pdf
4
Microsoft & Tech Council of Australia, Australia’s Generative AI Opportunity, July 2023, https://news.microsoft.com/wp-content/uploads/prod/sites/66/2023/07/230714-Austra-
lias-Gen-AI-Opportunity-Final-report.pdf
5
LexisNexis, AI Decoded: The Growing Influence of Generative AI on the Legal Industry: What Lawyers Need to Know, Episode 1 of AI Decoded, Legal Talk podcast, https://www.lexisnexis.
com.au/en/insights-and-analysis/practice-intelligence/2023/ai-decoded-the-growing-influence-of-generative-ai-on-the-legal-industry

6
The opportunity for lawyers
Most obviously, generative AI can expedite legal
research. A lawyer can type a legal query into a
chatbot and obtain the answer within seconds.
This is much faster than having to manually
The capability of large language models – the sift through legal databases, case summaries
sheer ability to produce content – is something we and legislation. It not only can conduct legal
have never seen before. So I do expect there will be research but can also analyse this information
significant change coming. to identify trends.

Caryn Sandler
Partner and Chief Knowledge and The technology can be used to generate synopses,
Innovation Officer at Gilbert+Tobin summaries and fact sheets, even a rough first
draft of legal analysis for a client. It can help get
lawyers off to a good start in drafting and editing

Generative AI offers a potential gamechanger documents in the right voice. This might range

for lawyers as it does not always require complex from basic client correspondence, form-filling and

systems integrations. Many are easily accessible administration to creating first drafts of contracts,

to firms of all sizes and in-house legal counsel. legal advice and court pleadings.

Microsoft 365 Copilot, for example, works

alongside Microsoft applications commonly used
by lawyers, such as Microsoft Word, Outlook and
Teams, to provide real-time intelligent assistance
by automating activities such as summarising
documents and emails and editing documents.

In broad terms, generative AI has the potential to

serve as a ‘copilot’ – augmenting but not replacing
lawyers in completing routine, repetitive and
time-consuming tasks. Yet crucially, generative
AI systems and tools are only as good as the
material they’ve been trained on, as well as their
probability-based approaches. As such, legal
practitioners must maintain a “human in the loop”
to review, edit, customise and quality-assure
content.

7
Using LLMs, generative AI also offers vastly Microsoft’s Corporate, External and Legal Affairs
updated document comparison and verification (CELA) has started a departmental wide initiative
capabilities. An example is checking large volumes focused on identifying opportunities to leverage
of contracts for risks and compliance, including AI to drive operational efficiencies, increase
potential issues arising from legislation and case productivity and amplify the department’s
law. This helps lawyers stress test their work, impact. The initiative asked for creative ideas
increase efficiency and minimise errors. Tools 6
leveraging AI, which has so far generated over 250
such as Kira, Imprima and Document Intelligence submissions that use existing tools and new ideas
are all gaining popularity as AI becomes standard with three common themes - better knowledge
part of conducting a due diligence review in a management, improve contracting and complying
M&A transaction. 7
with an increasingly complex regulatory
landscape.
Generative AI tools, such as Microsoft 365 Copilot
and Microsoft Azure OpenAI Service are not
designed specifically for any one industry. As 1. Better knowledge management
such, these tools can help entire organisations
implement a technology strategy that meets the
diverse needs of different departments, such as
Human Resources, Finance, Law and Compliance. 2. Improve contracting
In-house lawyers can benefit from generative AI
tools that can also serve other functions within
their organisation, instead of relying on dedicated
legal tools that may not be available or prioritised 3. Complying with an increasingly
as part of an organisation’s overall technology
complex regulatory landscape

strategy.

6
Jacqui Jubb, “The benefits of generative AI in law firms”, Lawyers Weekly, 27 June 2023, https://www.lawyersweekly.com.au/newlaw/37591-the-benefits-of-generative-ai-in-law-firms
7
Lauren Croft, “90% of lawyers confirm AI is crucial for M&A”, Lawyers Weekly, 12 July 2023, https://www.lawyersweekly.com.au/newlaw/37714-90-of-lawyers-confirm-ai-is-crucial-for-m-a

8
Consumer versus environment. This makes them more capable
of inputting firm, organisational or client data

enterprise-grade
to develop answers, similar to the distinction
between public and private cloud. For example,

tools
firms such as King & Wood Mallesons have signed
up to the Early Access Program for Microsoft 365
Copilot, which will allow the firm to define the
content that can be accessed by that service and
Some early experiments with generative AI restrict any third parties outside of the firm from
technology have leveraged ChatGPT for various accessing its information.
purposes such as data analysis and content
creation. However, a distinction should be made
between these types of consumer-grade tools
without enterprise-grade data privacy and security
controls, and generative AI technology that
leverages a firm or organisation’s data without
compromising the data privacy and security of
the organisation’s data boundary. In relation to Copilot

generative AI,
A whole new way to work

We should actually be looking to embrace the

potential because it’s coming, like it or not. The
earlier you get your head around it, the more
likely you are to benefit from it. But absolutely it
[generative AI] has to be used responsibly.

Patrick Gunning
Partner at King & Wood Mallesons

Given this, whilst consumer versions of ChatGPT

and other similar generative AI technology aimed
at consumers can access a wide source of publicly
available information, they are not designed for
enterprise level customers. As such, they typically
do not carry advanced data privacy and security
protections such as that offered by Microsoft
Azure OpenAI Service or Microsoft 365 Copilot,
which allow for access to powerful LLMs to use
on an organisation’s own data within a secure

9
Moving up the To cater for the evolving definition of what it
means to be a legal professional, education

value chain
and training models will need a refresh. For
example, junior lawyers may focus their learning
on technological proficiency in addition to
other basic legal skills required such as legal
What does seem clear is that by automating reasoning and drafting capabilities. Generative AI
routine work, generative AI frees up lawyers technology will assist lawyers, not replace the role
to focus on more complex, intellectually or of a lawyer. Firms and in-house legal operations
interpersonally challenging matters. This is of the future may be filled with knowledge
where critical thinking, commercial acumen, engineers, data analysts, technologists, design
human empathy and experience, and a holistic thinkers and transformation experts in addition
understanding of the needs of clients and to lawyers. Traditional billing models will also
stakeholders, may all come into play. be under severe pressure. “Undoubtedly, we are
getting to a world where the billable hour is not
It has been suggested, for example, that going to be representative of what lawyers bring
generative AI would be less able to replace lawyers – and we’re going to have to find different ways to
in areas requiring creativity or risk assessment attach value,” Sandler says.
skills.

Increasingly, lawyers’ human skills are becoming

more important in areas like strategy, commercial
judgment, abstract reasoning, figuring out solutions,
emotional intelligence. Things that we’re probably
not ready yet as a species to outsource to the
machines.

William Howe
Forensic Technology Director at
Clayton Utz

In this way, the work of lawyers will “move up the

value chain” into areas that involve more than
simply providing a legal service – from developing
industry expertise to offering strategic guidance
and building trusted client relationships. More
lawyers may be able to diversify their offerings.

10
Gilbert+Tobin
declares
an AI bounty
Law firm Gilbert+Tobin is an industry The best ideas for using generative AI
leader in using technology to support will likely arise overtime from individual
lawyers’ work, both for client services experimentation through trial and
and internal business transformation. The error. As such, the firm will also explore
firm recently created an AI bounty worth generative AI’s ability to engage in
$20,000 in which lawyers were invited to more advanced legal reasoning, such as
submit ideas on how to use ChatGPT or identifying the weakness in an opponent’s
other LLMs in the workplace. It received case or assessing contract clauses.
106 submissions about how generative AI
could streamline lawyers work – including
summarising content, helping to draft
documents, developing case chronologies My view is that if the technology demonstrates it is
and supporting business development capable of delivering higher-value legal work, we
activities. will have no option but to embrace it in a way that is
safe and effective. It is just too early to tell

Caryn Sandler
Partner and Chief Knowledge and
Innovation Officer at Gilbert+Tobin

Generative Al can assist risk management across industries

Illustrative use case of Generative Al in professional services
Legal Services

1 2 3

Lawyer completes first draft of A GAI tool reviews the document, The tool identifies a potential
a contract. looking for missing clauses by loophole, and suggests clauses to
comparing to other similar contracts. include as a solution.

10
LSJ Online, “Law Firm Sets $20,000 AI Bounty for Staff”, 1 May 2023, https://lsj.
com.au/articles/law-firm-sets-20000-ai-bounty-for-staff/ [extract from Microsoft/Tech Council of Australia report]
11
Climate-ready
for new tools
Australian law firms are already of environmental law by leveraging
incorporating generative AI technology language AI models available through
to support their clients and staff. For Microsoft Azure OpenAI Service. Without
example, Clayton Utz has developed an AI, such information would usually take
environmental, social and governance months to analyse and, instead, can be
(ESG) tool that tracks trends in climate finalised in a couple of weeks. Clayton Utz
change litigation in Australia and around requires human lawyers to review written
the world. This is an area where the law is work created by the ESG tool.
developing rapidly, with new policies to
regulate carbon emissions, rigorous rules
around climate disclosures and emerging
types of actions between plaintiffs and The humans are firmly in the loop,
defendants.
Yes, there are knowledge management hours
involved, but if we have a whole bunch of junior
The tool developed by Clayton Utz uses
lawyers sitting there typing summaries, is that the best
Microsoft Azure OpenAI Service, which
use of their time? We use the AI to do a lot of that.
gives customers access to advance
language AI models such as OpenAI William Howe
Forensic Technology Director
GPT-4, GPT-3, Codex, and DALL-E
at Clayton Utz
but with the benefit of using it within
a private data boundary and with
responsible AI content filtering and abuse
monitoring. Generative models such as
Microsoft Azure OpenAI Service have
significant potential benefits for lawyers
as customers can tailor Microsoft Azure
OpenAI Service models using their own
dataset and train them to better respond
to their specific prompts.

In the case of the ESG tool, it identifies

trends in environmental case law
by synthesising hundreds of pages
12
Clayton Utz was an early pioneer in Inevitably, Howe believes, law firms will
incorporating AI capabilities. With the create new business models as large
recent rise of AI, it formed an internal language models enable more advanced
AI working group and developed legal semantic searching and in-context
risk frameworks to cover areas such as learning.
confidentiality, intellectual property
and mitigating against generative AI
hallucination. Today, Howe is one of
the leaders of a team of more than 100, Technology literacy is already a core skill for lawyers.
offering client-facing services to in-house
The future is a hybrid model where AI becomes a
counsel and other business functions such
standard part of the technologies lawyers use on a
as compliance and risk.
daily basis. We have an optimistic view of the future
and it’s really exciting to be part of this journey with all
In addition to its ESG tool, the firm has my colleagues.
introduced several products that use
William Howe
generative AI to distil large amounts of
Forensic Technology Director
information. Howe cites use cases such
at Clayton Utz
as e-discovery, data analysis for fraud or
regulatory investigations, even simply
helping employers identify their key
obligations without having to navigate a
thicket of industrial relations instruments.
Another emerging area is risk and
reputation management, where clients
wish to proactively monitor everything
from website content to annual reports
and staff social media posts.

13
Microsoft
365 Copilot
for lawyers
Microsoft 365 Copilot is a sophisticated as such, does not unintentionally leak to
processing engine that leverages individuals outside of the limits set by an
powerful LLMs with Microsoft 365 apps organisation.
to capture natural language commands
to automate a range of activities such as King & Wood Mallesons has signed-
producing content, analysing data and up to the Microsoft 365 Copilot Early
building presentations. Microsoft 365 Access Program. Like for many other
Copilot leverages existing Microsoft 365 organisations, client confidentiality is
role-based controls to limit its access fundamental to King & Wood Mallesons.
only to information that a user already Generative AI relies on training natural
has authorisation to access. It also language models, based on document
does not use customer data (including sets. If those document sets include any
prompts and output responses) to train client information being exposed in a
its foundational language model. Instead, public environment, it restricts a lawyer’s
it is designed so that data does not use of such technology. In relation to
leave an organisation’s boundary and, Microsoft 365 Copilot,

14
 together to identify and prioritise use
cases to increase efficiencies and client
outcomes using AI technology, such as
Having the various [Microsoft] Azure-based AI Microsoft 365 Copilot. Areas of focus
offerings is attractive to any law firm, frankly, because for the working group include improve
it’s in a closed environment. You can be confident efficiency through transformation of
the information won’t go outside. That’s been an routine processes such as producing
important consideration in our decision to go into materials required for proposals to client
the Microsoft preview program. We see that as a safe
and enabling better ways to communicate
environment and that is the most important thing. We
complex concepts by converting texts to
can’t do any of the use cases without it.
diagrams and pictures.
Patrick Gunning
Partner at King & Wood Mallesons Gunning says Microsoft 365 Copilot
will also be useful in utilising pattern-
King & Wood Mallesons has established matching capability to find the best
a working group that not only involves it recent examples of the firm’s advice on a
lawyers but also other areas of the firm topic, thus allowing for better knowledge
such as business development, finance management.
and innovation. This group is working

15
Current regulatory
landscape
Lawyers seeking to use generative AI tools agencies have signed an Algorithm Charter that
must comply with all applicable laws, as well as covers the ethical design of public services.10 The
specific rules and guidance governing the legal New Zealand Privacy Commissioner has also
profession. So far, AI is an area where technology published guidance to help organisations manage
developers are moving fast and regulators are generative AI.11
catching up, seeking to strike the right balance
between the technology’s opportunities and risks, Microsoft supports the regulatory reform actions
while not stifling innovation as new use cases and of Australian and New Zealand governments
possibilities cascade. and recognises its own responsibilities. Microsoft
has implemented and publicly released its own
In Australia and New Zealand, there are currently Responsible AI Standard, which aligns with
no laws or regulations that apply specifically to AI. the Australian Government AI Action Plan and
However, Microsoft and others are working with identifies six key principles that guide how
Australian and New Zealand governments to help Microsoft develops AI products. These principles
identify and develop risk-based AI regulation to encompass the key concepts of fairness, reliability
address potential gaps in the current legislative and safety, privacy and security, inclusiveness,
landscape and reduce risks of harm. transparency and accountability. Microsoft
understands that responsible AI is a journey.
As part of a soft-law, principles-based approach, As such, the Responsible AI Standard is a living
the Australian Government has published an document that is evolving to address new
AI Action Plan, along with a set of voluntary research, technologies, laws and learnings from
ethics principles that may be used by business within and outside the company.
or government in embracing AI. Key principles 9

include the need for AI systems to be fair, reliable,

non-discriminatory, transparent and explainable
to the people they impact. The government
is also considering requiring organisations to
provide more transparency about how they use
automated systems to make decisions as part
of planned reforms to the Australian Privacy
Act. Meanwhile in New Zealand, government

9
Department of Industry, Science, Energy and Resources, AI Ethics Principles,
accessible at: https://www.industry.gov.au/data-and-publications/building-
australias-artificial-intelligence-capability/ai-ethics-framework/ai-ethics-principles,
10
Office of the Prime Minister’s Chief Science Advisor, “Why is regulating AI such a
challenge?”, 13 July 2023, https://www.pmcsa.ac.nz/2023/07/13/why-is-regulating-
ai-such-a-challenge/
11
New Zealand Privacy Commissioner, Generative Artificial Intelligence – 15 June 2023
update, https://www.privacy.org.nz/publications/guidance-resources/generative-
artificial-intelligence-15-june-2023-update/

16
Ensuring responsible use of AI is not limited To guide this, Microsoft has announced its
to technology companies and governments. three AI Customer Commitments to assist our
Every organisation, including law firms, that customers on their responsible AI journey.
creates or uses AI systems will need to develop These commitments include:
and implement its own governance systems.

Sharing Microsoft’s Creating an AI Assurance Supporting customers as

learnings about developing Program to help ensure they implement their own
and deploying AI that AI applications that AI systems responsibly,
responsibly. our customers deploy including through our
on Microsoft platforms partner ecosystem.
meet the legal and
regulatory requirements for
responsible AI.

Microsoft also provides in its customer infringement by output generated by Microsoft’s

agreements a Copilot Copyright Commitment, Copilot or Bing Chat Enterprise products,
which provides that if a third party sues a Microsoft will defend the customer and pay any
commercial customer for copyright or other IP resulting adverse judgments or settlements.

17
Europe’s
AI Act
Some indication of what future regulation and banned. Any AI systems that pose a
could look like comes from the European fundamental risk to safety and security
Union’s (EU’s) proposed AI Act, one of the will be classified as ‘high-risk’. Though
first comprehensive laws of this nature permitted in the European market, these
in the world.12 This regulation follows a will be subject to additional compliance
risk-based approach, ranking AI systems requirements, including an obligation to
based on the level of risk they pose be registered in an EU database.
to users. For example, AI systems that
contravene European Union values, such
as violating fundamental human rights
will be deemed as an ‘unacceptable risk’

12
EU AI Act: first regulation on artificial
intelligence, 14 June 2023, https://
www.europarl.europa.eu/news/en/
headlines/society/20230601STO93804/
eu-ai-act-first-regulation-on-artifi-
cial-intelligence

18
Professional conduct rules
there is nothing inherently improper about using
a reliable artificial intelligence tool for assistance”.
Nonetheless, he found there was a “gatekeeping
role on attorneys to ensure the accuracy of
We have an obligation to review everything that is their filings”.
produced using Large Language Models and verify
it – as we would across any other legal work. In addition to any disciplinary consequences
from breaching their professional obligations
Caryn Sandler
Partner and Chief Knowledge and of competence, a lawyer may also be liable for
Innovation Officer at Gilbert+Tobin negligence. It is a lawyer’s responsibility to use
any content produced by an AI tool or system
with reasonable care and skill. A client suffering
In the absence of AI-specific regulations, lawyers
loss from negligent use of AI by a lawyer could
need to abide by professional conduct rules
potentially be entitled to compensation.13
in using AI tools. Uniform Conduct Rules have
been adopted in the Australian Capital Territory,
It is equally possible to imagine a day when there
New South Wales, Queensland, South Australia,
is a positive duty on lawyers to use the latest,
Tasmania, Victoria and Western Australia. The
technology to deliver legal services.
Northern Territory has established its own rules,
as has New Zealand. Each of these emphasises the
responsibility of lawyers to deliver legal services
competently and diligently and to act in the best
interest of the client.

One could argue that subject to appropriate

A lawyer using generative AI should deliver ‘supervision’, it is a solicitor’s duty to consider using
at least the same quality of legal service they ChatGPT as part of our obligation to act in the best
would otherwise be obligated to provide. At a interests of our client.14
minimum, there is need for human oversight and
independent legal analysis to deliver services Genevieve Collins
competently and ensure the technology is not Chief Executive Partner at Lander & Rogers
‘hallucinating’ – that is, simply delivering wrong or
misleading information. Given this, generative AI
In the United States, the American Bar Association
should be seen as a copilot, not an autopilot.
has approved changes to professional standards
making clear that lawyers have a duty to maintain
The notorious example of two New York lawyers
their knowledge and skill by keeping pace with
who included fake citations from ChatGPT as part
changes in the law and its practice, “including
of a personal injury case stands as a cautionary
the benefits and risks associated with relevant
tale. In a written opinion, the judge noted that
technology”.
“technological advances are commonplace and
13
Michael Guihot and Lyria Bennett Moses, Artificial Intelligence, Robots and the Law (LexisNexis Butterworths, 2020) 115. 19
14
Michael Pelly, “Law firms say ChatGPT an ‘opportunity, not a threat’”, Australian Financial Review, 9 February 2023, https://www.afr.com/companies/professional-services/law-
firms-say-chatgpt-an-opportunity-not-a-threat-20230208-p5cj2j
Maintaining
confidentiality
A legal practitioner’s duty of confidentiality is consumer tools such as ChatGPT that may use
part and parcel of the fiduciary relationship with inputted data to train and improve the algorithm’s
their client. It is also reflected in professional workings could potentially result in disclosure of
conduct rules across Australia and New Zealand. confidential information that is input.15 Therefore,
For example, under the Uniform Conduct Rules, a lawyers need to consider the data protection
solicitor must not disclose any information that is commitments and controls associated with an AI
confidential to a client and acquired during their system before using it with confidential firm or
engagement unless an exception applies. client data.

In the early days of cloud computing, for example,

client confidentiality was one of the key concerns
for law firms when considering the move to the
cloud. The issue was whether use of a third-party
cloud provider such as Microsoft to store and
process confidential client data would amount, in
legal terms, to a disclosure requiring prior consent.

However, based on relevant guidelines published

by industry and regulatory bodies in different
Australian jurisdictions, as well as analysis by the
Office of the Australian Information Commissioner
(OAIC), the legal profession has widely deployed
cloud computing to process and store confidential
client data. Client consent has not been
considered a requirement, as long as the nature of
the cloud service and contractual arrangements
between the parties are sufficiently protective.

Professional confidentiality obligations are clearer

with respect to generative AI than the early
days of cloud computing. Publicly accessible

15
Karen Sloan, ‘A lawyer used ChatGPT to cite bogus cases. What are the ethics?’,
Reuters (online), 31 May 2023 <https://www.reuters.com/legal/transactional/
lawyer-used-chatgpt-cite-bogus-cases-what-are-ethics-2023-05-30>

20
What’s next?

The speed of change is astonishing... I’ve been in

technology a long time, and this is probably the
most exciting thing I’ve seen in my career. What
makes me excited, and actually a bit uneasy, is the
thought that in five years, we will look back on today
and think ‘that was only the beginning…16

Claire Smith
Partner, Clayton Utz

Within the legal profession, there are markedly

varying levels of digital literacy. Among law firms,
the market is competitive and clients expect
the use of the latest technology capabilities
where available. Many firms have major change
The law is not immune in any way and there will
management programs in place – with lawyers
be heaps of industries disrupted by this technology.
engaging with innovation teams to improve
Some of the more routine areas will experience the
processes and solve pain points.
most disruption in the shortest period of time.

In-house legal teams are also presented Patrick Gunning

with powerful opportunities to benefit from Partner at King & Wood Mallesons
enterprise AI systems and tools implemented by
their organisation.
Gunning says that in relation to generative AI
Generative AI promises to ripple across these Some organisations will be happy to lead while
divisions – placing exciting new capabilities in the others sit back until there are established industry-
hands of every lawyer and organisation. Research wide parameters and frameworks. However, the
indicates that rapid advancement of generative AI capabilities
indicates this is an area that organisations must
prioritise to maintain a competitive advantage.

40%
of law firms
The practice of law is changing fast...
are already
experimenting
with AI.17 16
Lauren Croft, “Clayton Utz accelerates ESG work with ChatGPT”, Lawyers
Weekly, 21 February 2023, https://www.lawyersweekly.com.au/newlaw/36705-
clayton-utz-accelerates-esg-work-with-chatgpt
17
Ellie Dudley, Half of Australian lawyers fear AI will take their jobs, research
reveals, The Australian, 23 August 2023.
21
©2023 Microsoft Corporation. All rights reserved. This document is provided “as-is.” Information
and views expressed in this document, including URL and other Internet website references, may
change without notice. You bear the risk of using it. This document does not provide you with
any legal rights to any intellectual property in any Microsoft product. You may copy and use this
document for your internal, reference purposes.
22