Generative AI in the Practice of Law

We have all seen the rapid technological advancements in generative artificial intelligence (“Gen AI”) in recent years. These developments present significant opportunities for lawyers to deliver more efficient services; however, they also introduce risks that licensees must not overlook. The use of Gen AI or similar tools must always be guided by our professional responsibilities. These advancements have resulted in the Canadian Courts and Law Societies having to deal with the challenge of how to handle these in the profession of law.

As a short background, Gen AI refers to tools that produce new content such as text, code, music, images etc. using generative models. Large Language Models (LLMs) are a form of Gen AI that generates text in response to user prompts, with ChatGPT being one of the most widely recognized examples.

Federal Court

On December 20, 2023, the Federal Court issued a notice titled “The Use of Artificial Intelligence in Court Proceedings”[1]. This notice required that any document submitted to the Court and prepared for the purpose of litigation must include a declaration if it contains AI-generated content. This declaration would appear in the first paragraph of the document.

The accompanying principles emphasized the need for caution when using AI in the preparation of court materials. The Court also encouraged counsel to consider providing traditional, human services to where there is reason to believe a client may be unfaimilar with, or may prefer not to rely on, AI. In addition, the notice underscored the importance of maintaining a “human in the loop” to ensure accuracy and trustworthiness by having a human double-check any material generated by AI.

On May 7, 2024, the Federal Court issued an updated notice also titled “The Use of Artificial Intelligence in Court Proceedings”[2]. This update clarified what needs to be flagged as produced by AI as the Court gives an example that a declaration is not necessary if an 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. A declaration is required when the role AI plays in the preparation of materials for the purpose of litigation resembles that of a co-author.

In the principles of use of AI, they have introduced a neutrality stance, confirming that the inclusion of an AI-Use declaration, in and of itself, will not attract an adverse inference by the Court.

Additionally, the Civil Rules Committee has establised an Artificial Intelligence Subcommittee[3] which will begin analyzing and advising on proposals concerning the use of AI and technology in the litigation process.

Law Society of Ontario

In April 2024, the Law Society of Ontario issued a white paper titled “Licensee Use of Generative Artificial Intelligence”[4]. This document is intended to help guide licensees as they navigate the use of Gen AI tools and stresses the relevance of rules of professional conduct such as: Competence[5], Confidentiality[6], Supervision[7], Licensee-Client Relationships[8], Fees and disbursements[9] and Discrimination and harassment[10]. As one might imagine, Gen AI tools have the potential to create significant ethical and practical challenges, and practitioners must always be careful when incorporating these technologies into their practices.

The Law Society of Ontario has also developed a quick-start checklist that can be a helpful guide[11] for practitioners when using Gen AI, which can be found online. For anyone interested in implementing any AI or Gen AI in their practice, it is extremely important to review these guidelines to fully understand both the capabilities and the limitations of these tools.

Gen AI in Court

One of the most significant issues associated with Gen AI is its tendency to produce “hallucinations”. These hallucinations are phenomena in which the AI provides a response that incorporates completely fabricated data, presented as if it were accurate. These hallucinations can be particularly problematic in the legal context, and there have already been several notable instances where they have affected legal proceedings.

In the United States case of Mata v Avianca[12], Mata’s attorneys submitted court materials that included fictitious case citations generated by ChatGPT.  Counsel for Avianca stated they were unable to locate most of the authorities cited. ChatGPT assured Mata that all of these cases “indeed exist” and can be found in “reputable legal databases such as LexisNexis and Westlaw”. This led the Mata’s attorneys to reply with an affidavit containing all but one of the decisions, which were all fabricated by ChatGPT. As a result, the case was dismissed, and Mata’s attorneys were sanctioned and ordered to pay a $5,000 penalty.

A similar issue arose in the Canadian case of Zhang v. Chen,[13] which has become an important decision regarding the misuse of artificial intelligence in British Columbia. In this case, counsel filed a notice of application containing only two cases, both of which were nonexistent legal authorities that ChatGPT had “hallucinated.” Counsel explained that they did not intend to generate or rely on fictitious cases and were unaware of the risk that AI tools could produce fabricated authorities[14].

The Court found the lawyer who submitted the notice of application personally liable for the expenses incurred in verifying the fictitious cases. The lawyer was also ordered to review all files currently before the court within 30 days and to advise opposing parties and the Court of any materials that contain any case citations or summaries that were generated by any GenAI tool.[15] The decision highlighted that counsel had failed to apply the Federal Court’s “human in the loop” principle, underscoring the importance of human oversight when using AI in legal practice.

Conclusion

In conclusion, the evolution of Gen AI presents both opportunities and challenges for the practice of law. While these tools have the potential to enhance efficiency, it is crucial for lawyers to exercise caution and maintain transparency with both the Court and clients regarding any AI use. Protecting confidential and privileged information remains paramount.

As the Honourable Justice D.M. Masuhara stated:

“As this case has unfortunately made clear, generative AI is still no substitute for the professional expertise that the justice system requires of lawyers. Competence in selecting and using technology tools, including those powered by AI, is critical. The integrity of the justice system requires no less. “[16]

This serves as a reminder that, although AI can assist in legal work, it cannot replace the judgment, skill, and ethical obligations that are essential to the practice of law.