Use of generative AI is in its early stages, and lawyers are early adopters because of the tremendous potential for Large Language Models (LLMs) such as Chat GPT and Casetext’s Co-Counsel to streamline work. Mark Williams, Associate Director for Collections and Innovation for Vanderbilt’s Alyne Queener Massey Law Library, and Cat Moon, Director of Innovation Design for Vanderbilt’s Program on Law & Innovation, organized a two-day bootcamp to introduce faculty to the potential of LLMs based on generative AI technology.
“Generative AI made a huge leap forward in November 2022—a stunning development that most people thought was years away,” Williams explains. “A technology designed to simply predict the next word of text started exhibiting problem-solving and reasoning ability. We designed a bootcamp to introduce faculty who are not yet using LLMs to this rapidly evolving technology and help faculty who are already testing out LLM tools better understand how they work, identify more ways they can use them to support their work, and avoid the pitfalls.”
Williams and Moon emphasize that learning how LLMs work and using them effectively will be essential to lawyers and legal scholars going forward. A survey published by Lexis/Nexis in April 2023 that included more than 2,300 lawyers and law students revealed 36 percent of the lawyers surveyed and 44 percent of law students have used LLMs either personally or professionally. “AI won’t replace lawyers, but it can help them work much more efficiently,” Moon says. “It has tremendous potential to streamline work and ultimately to increase access to legal services.”
In their two-day bootcamp, Williams and Moon offered practical guidelines to enable faculty to make the time they spend exploring LLM tools productive and begin using them for time-consuming tasks such as developing syllabi for courses, developing tests and quizzes, and drafting emails and meeting agendas. “When faculty can gain time to spend on their scholarly work and developing new courses, that’s a positive outcome of using this technology,” Williams says.
Several faculty members are already exploring or experimenting with using LLMs, and Moon and Williams incorporated their trial uses as demos. “One goal of the bootcamp was to help faculty develop effective prompts that generate usable results and give them a glimpse of how LLMs can streamline empirical research,” Williams says.
Williams and Moon believe teaching faculty and students how LLMs work and how to use them appropriately and productively is essential. The Lexis/Nexis survey indicated that lawyers saw the greatest potential for LLMs in conducting and organizing basic legal research, drafting and analyzing documents, and writing emails. “Students must be prepared to use LLMs in legal research, writing, and analysis in legal practice, and they must understand the ethics and risks of its use,” Moon says.
Moon and Williams are developing a course on evolving uses for LLMs that will be offered to students as an upper-level elective in fall 2023.