Charles Ciumei KC & Wei Jian Chan
Essex Court Chambers
A. Introduction
1. As Generative Artificial Intelligence (“Gen AI”) becomes more prevalent in our lives and legal processes, Courts and other legal institutions around the world have issued guidance on the use of Gen AI. This article summarises guidance from the United Kingdom, United States of America, Canada, Australia, and Singapore, highlighting key themes common to the guidance.
B. Overview of guidance
2. In England and Wales, there is a growing body of guidance on the use of AI.
2.1. The Courts and Tribunals Judiciary issued ‘Artificial Intelligence: Guidance for Judicial Officers Holders’ on 12 December 2023 and updated on 15 April 2025 (the “E&W Courts Guidance”). The E&W Courts Guidance applies to ‘judicial office holders’. It sets out key risks and issues associated with the use of Gen AI and emphasizes that any use of AI by or on behalf of the judiciary must be consistent with the judiciary’s overarching obligation to protect the integrity of the administration of justice.
2.2. The Bar Council issued guidance titled ‘Considerations when using ChatGPT and generative artificial intelligence software based on large language models’ on 30 January 2024 (the “E&W Bar Council Guidance”). The E&W Bar Council Guidance is aimed at all barristers and their chambers. It provides a summary of considerations for barristers if they decide to use Gen AI tools, including the potential for hallucinations, and need to protect confidentiality and privilege.
2.3. The Bar Standards Board also made a blog post dated 8 October 2023 titled ‘ChatGPT in the Courts: Safely and Effectively Navigating AI in Legal Practice’. The Bar Standards Board contrasted two cases: (i) the New York proceedings in Mata v Avianca, in which a faulty AI-generated legal analysis containing fictional citations was placed before the Court, and (ii) comments from Birss LJ that ChatGPT was “jolly useful” and had “real potential” for drafting judgments.
2.4. On 20 November 2023, the Solicitors Regulation Authority, issued a publication titled ‘Risk Outlook report: the use of artificial intelligence in the legal market’, highlighting various aspects of the use of Gen AI, including opportunities, risks, and barriers associated therewith.
2.5. On 6 June 2025, Dame Victoria Sharp sitting in the Divisional Court handed down a judgment in R (Ayinde) v Haringey and Al-Haroun v Qatar National Bank QPSC [2025] EWHC 1383 (Admin), responding to concerns that incorrect propositions of law had been advanced and authorities which did not exist had been cited to the Court in those proceedings. Dame Sharp highlighted regulatory duties applicable to barristers and solicitors, and also the potential criminal, regulatory, and procedural consequences of placing false material before the Court. She also set out examples of Gen AI having been used inappropriately from various other jurisdictions.
3. In the United States, general guidance on the use of Gen AI has not been issued, but there have been individual guidance issued by various institutions.
3.1. The American Bar Association (ABA) issued Formal Opinion 512 on Generative Artificial Intelligence Tools on 29 July 2024, discussing various issues arising out of the use of such tools. These issues include the competent use of such technologies, maintaining client confidentiality, communication with clients about the use of Gen AI, as well as management and billing matters.
3.2. Judge Leslie E. Kobayashi of the US District Court for the District of Hawaii issued guidance requiring disclosure of the use of any Gen AI tool in the preparation of court documents, and certification that the accuracy of any document so prepared has been checked.
3.1. Judge Araceli Martinez-Olguin of the US District Court for the Northern District of California addressed the use of Gen AI in her standing order for civil cases before her (revised 22 November 2023). She emphasized that “counsel must at all times personally confirm for themselves the accuracy of any content generated by these tools”.
4. In Canada, the Federal Court issued a ‘Notice to the parties and the profession: The use of artificial intelligence in Court Proceedings’ dated 7 May 2024. This Notice mandates the declaration if content in material submitted to the Court and prepared for the purpose of litigation was directly provided by Gen AI. Notably, the Notice provides that “a Declaration is not required if AI was used to merely suggest changes, provide recommendations, or critique content already created by a human who could then consider and manually implement the changes”. The Court also urged caution when using legal references or analysis created or generated by Gen AI.
5. The Canadian Judicial Council issued ‘Guidelines for the Use of Artificial Intelligence in Canadian Courts’ in September 2024. These guidelines apply to Canadian judges, and serve two objectives: (i) to establish a rationale for a consistent approach to the utilization of Gen AI in Canadian courts and (ii) to shed light on both the opportunities and risks inherent in Gen AI’s potential incursion into court administration and judicial decision-making.
6. In Australia, the Supreme Court of Victoria issued ‘Guidelines for litigants: responsible use of artificial intelligence in litigation’ in May 2024. These Guidelines stress the importance of understanding how Gen AI tools work, that disclosure of the use of Gen AI may be required, and emphasize that the use of Gen AI is subject to the obligations of legal practitioners in the conduct of litigation, including the obligation of candour to the Court and the requirement in the Civil Procedure Act 2010 that submissions have a proper basis.
7. The Chief Justice of New South Wales issued Supreme Court Practice Note 23 on Use of Generative Artificial Intelligence on 28 January 2025. This Practice Note prohibits the entry of material produced on subpoena, or any material that is the subject of a statutory prohibition upon publication into a Gen AI system unless appropriate safeguards are imposed (e.g. to avoid such material being used to train Gen AI models). It also prohibits the use of generative AI when drafting witness statements and expert reports. However, it notably permits the use of Gen AI for the preparation of documents such as chronologies and summaries.
8. In Singapore, a ‘Guide on the Use of Generative Artificial Intelligence Tools by Court Users’ applicable from 1 October 2024, applies to all matters before the Supreme Court, State Courts, and Family Justice Courts. This Guide does not prohibit the use of generative AI tools, but emphasizes that Court users are responsible for the content of their documents, including the need to ensure accuracy and protect intellectual property rights.
C. Key themes
9. As a general trend, legal bodies have been permissive but cautious in relation to the usage of AI. A number of common themes run through the guidance highlighted above:
10. First, although no legal bodies have prohibited the use of AI tools, virtually all bodies issuing guidance have emphasized that professional obligations continue to apply, and that legal professionals remain obligated to ensure the accuracy of their documents. It appears that legal bodies are mindful of the potential for Gen AI models to ‘hallucinate’ and produce incorrect references. They are mindful of the tendency of non-expert users (who do not understand Gen AI) to place too much faith in the output of such systems.
11.Second, various guidance documents have also emphasized the need to maintain confidentiality, and to avoid uploading confidential information to Gen AI systems which do not have suitable safeguards (e.g. against privileged documents being used to train Gen AI models).
12. Third, a number of jurisdictions have taken the step of requiring disclosure of Gen AI usage to promote transparency and as a means of regulating the use of Gen AI, although there is by no means a ‘general rule’ in this regard. In addition to disclosure to the Court and to other parties, lawyers should consider the need to disclose Gen AI usage to their clients where appropriate.
13. Fourth, it is clear that the appropriateness of using Gen AI will depend in large part on context and circumstances. The usage of Gen AI may be appropriate in certain circumstances (preparing chronologies or summaries) but not in others (preparing witness statements or expert evidence).
D. Conclusion
14. Despite the hype surrounding Gen AI as a world-conquering technology, it is in many ways a very expensive solution in search of a problem. The swift promulgation of guidance by Courts and legal institutions worldwide reflects the rapid development of Gen AI technologies. Nevertheless, it is clear that the usage of Gen AI in legal proceedings remains in its infancy, and the frameworks governing its use will continue to evolve over time. Legal institutions worldwide seem to be taking a pragammatic and relatively cautious approach to integrating Gen AI technologies into legal processes. The legal profession is not the only one facing the challenges of understanding the strengths and weaknesses of Gen AI and exploring how best to deploy it: see, for example, a recent paper examining the use of LLMs in the provision of mental health care: Moore, J. et al (2025) Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers, arXiv:2504.18412 [cs.CL].
