Michael L. Smith (St. Mary’s U School of Law) has posted “Generative AI in the Attorney-Client Relationship: An Exercise in Critical Revision and Client Management” on SSRN. Here is the abstract:
Discussions of generative AI in legal practice and education often assert that this technology will lead to a sea change in legal writing, research, and revision. While some of the more breathless proclamations deserve skepticism, there’s little doubt that this technology may generate new forms of headaches for those in the legal field – particularly in the hands of clients or opposing counsel who attempt to use this technology to save the time, money, and effort required for complex legal tasks.
To that end, this essay proposes an exercise template for law students which illustrates how generative AI technology may be misused or abused. Presenting students with an AI-generated motion and asking them to reason through a scenario in which a hypothetical client demands that they file the motion tests a number of skills. First, students must critically read and revise the motion – noting shortcomings in AI-generated legal writing and identifying the confident mistakes that permeate the output. Second, and perhaps even more importantly, students must think through how to communicate these mistakes to a stubborn client, requiring them to consider client relationships and motivations and to communicate complex information in a simple, concise, and diplomatic manner. Doing so takes the exercise beyond a practical test of doctrine and legal writing, and engages students with deeper questions of empathizing with client needs, developing their professional identity, and preparing for a world in which generative AI will not only be used, but also abused.