I have been amazed at the knee-jerk and head in the sand reaction of many courts to the use of Gen AI by lawyers appearing before them and the use by judges. Several judges have demanded that lawyers disclose the use of Gen AI and, if so, verify the accuracy of the citations that have been used. One judge even went so far as to demand disclosure of any AI tool, generative or not. And the 5th Circuit is considering not only requiring lawyers to disclose and verify but also requiring non-represented parties to do the same. (Like how are they going to do that?)
Of late, though, there have been more studied reactions to Gen AI in law. There is a welcome recognition of its potential benefit in judicial proceedings by both lawyers and judges.
UK Judicial Council AI Guidance
On December 12, the UK Judicial Council issued its Guidance for Judicial Office Holders in England and Wales for using AI and Gen AI. This document acknowledges the inevitable: judges and lawyers will, and probably should, use these tools. It emphasizes education over edicts. It’s a stance that acknowledges the reality of Gen AI usage and promotes a nuanced understanding of its application.
The Council’s Paper begins by clarifying what Gen AI is and how it differs from broader AI applications. The Council offers definitions to clear up any misunderstanding about the tools. Then, it delves into the mechanics of Gen AI and large language models in a layperson-like manner. It describes in plain language how the models work and what they do.
The Paper discusses the risks and limitations of Gen AI. It notes that the tools are perhaps best used for nondefinitive confirmation rather than to generate new facts. The Council specifically notes that the quality of the responses generated will depend on the quality of input. The Council cautions users to beware of potential biases caused by the tools. The Council also thoughtfully addresses privacy concerns. It provides an outline of the importance of and best practices for maintaining security.
Use of Gen AI in Law
The Paper further notes that legal professionals have long used various AI tools, like Technology-Assisted Review (TAR). It states that there’s generally no need for courts to mandate disclosure of AI or Gen AI usage. But it does caution that lawyers need to adhere to their duty of verifying the accuracy of citations and case holdings. It also recognizes the challenges faced by unrepresented parties in verifying the accuracy of such information. In this regard, the Council suggests that judges may need to play a more proactive role in these scenarios. (something our esteemed 5th Circuit seems to have ignored in its recent proposals).
The document reviews the potential use cases of Gen AI and discusses what it should not be used for. And it doesn’t just stop at outlining the do’s and don’ts; it also provides practical tips for judges to identify the use of Gen AI, such as unusual citations or American spellings in submissions. These insights reflect a deep understanding of the nuances of Gen AI applications in legal contexts.
Overall, the Paper is a good effort to educate UK judges about the risks and benefits of the technology. It’s a practical approach, recognizing the reality of the use of Gen AI.
Chief Justice Roberts on AI
We are also beginning to see these more nuanced reports in the U.S. In his recent end of year Report, Supreme Court Chief Justice Roberts at least discuses the use of Gen AI by the judiciary. While his discussion is a bit sophomoric, he does recognize that despite everything, technological changes do and will transform the work of judges. And he correctly notes that trials today look very different than they did even a decade ago. Lawyers present, and jurors receive evidence and information very differently today, which I have discussed before.
Chief Justice Roberts even mentions the potential benefit of Gen AI for legal research and its ability to increase access to information for lawyers and the unrepresented. He notes the potential accuracy, security, and bias issues associated with GEn AI. But he recognizes that “As AI evolves, courts will need to consider its proper uses in litigation.” He promises several Judicial Conference Committees will be doing that in the future.
While Chief Justice Roberts’ Report was not as detailed as that of the U.K. Judicial Council, it does provide a recognition to U.S. Courts of the benefits of Gen AI. Perhaps more importantly, Chief Justice Robert’s Report is a statement of the inevitability of Gen AI in the U.S. Court systems.
AI in Judicial Proceedings: No More Sticking Your Head in the Sand
Both judicial statements are clear signals to judges: no more sticking your head in the sand. No more pretending Gen AI is evil and should not be used. Given some judicial reactions, that’s the most long-ranging impact of both Reports.