As I work on my legal research class for the fall, I am faced with the question: To teach generative AI or not to teach generative AI? I think at this point, there is no getting around the fact that I will be teaching generative AI as part of my curriculum.
I will leave the potential for generative AI use in legal writing or other contexts like the workflow within the law library to experts in those fields. My focus will be on vendor-based generative AI in databases such as Lexis+, vLex and Westlaw Practical Law. For now, I will leave generative AI tools such as ChatGPT, MS CoPilot, etc. out of my legal research instruction.
The next step is to think about how I am going to teach AI. As I craft my approach, one of the most challenging questions I am wrestling with is how to help students understand what information or data the generative AI tools are using to draft responses. How am I going to approach this discussion?
To add even more challenge, in the past few years, I have noticed my students’ increasing reluctance to read the secondary and primary sources legal research platforms provide in response to queries. With generative AI, this skill is no less important. Students need to read and analyze any materials generative AI delivers to them. How can I be sure the students take this to heart?
I began to reflect on these basic skills and was drawn back to the skills that make up information literacy. As a librarian at an undergraduate business school for a few years, I had started to approach this topic with my students, and had just begun to gain some traction with professors about the importance of information literacy when I moved back into law libraries. But I reflected – could these information literacy skills help students navigate the brave new world of generative AI assisted legal research?
I pulled up ACRL’s Framework for Information Literacy for Higher Education. Some ideas here just might be helpful this fall. After a bit of searching, I located a Business Research and Services Section (BRASS) guide which maps these ACRL information literacy standards to generative AI. Looking at this tool, I began to think about the Framework and teaching my students about generative AI in legal research tools.
A few of the Framework Standards as BRASS mapped them to AI resonated with me as potentially useful for legal research class this fall.
Standard 1: Authority is Constructed and Contextual
BRASS application to AI: Are responses generated by AI transparent enough to evaluate the source material to determine authority? Does the response give transparency on the tools and research methods used to provide response material, for the purposes of helping to determine authority?
Legal Research: Determining the source material and its authority are crucial to legal research. Can students determine what source material w searched by the generative AI? In a legal research perform, was all content searched? Primary? Secondary? Can we tell?
Standard 2: Information Creation as a Process
BRASS application to AI: Do users understand the AI’s own creation process to transform datasets and source material into generated responses?
Legal Research: Legal research platforms generative AI, similar to other generative AI, can seem quite authoritative. “ChatGPT is built on predictive language modeling, which means it generates natural sounding language, not necessarily factual language. Students should always approach information, from ChatGPT or elsewhere, with skepticism.” See Where does ChatGPT fit into the Framework for Information Literacy? I wonder if students new to legal research will be able to question the answers provided by generative AI in legal research platforms. Or will the format of these responses lend more authority to them than perhaps they warrant?
Standard 3: Information has Value
BRASS application to AI: Are responses generated by an AI influenced by its value as a commercial product? Do the responses reflect any biases or omissions caused by its training datasets?
Legal Research: Legal research databases are definitely influenced by their value as a commercial product – these databases do not provide answers from outside their own content. If content is better on platform A than B, but student goes to platform B and the answer generated is less than complete, how does a student know? The difference here from pre-generative AI searching is how much more authoritative the answer can sound when generated by AI rather than just a list of results. This is similar to bias which may have always been present in legal research platforms, but I wonder if generative AI will bury bias even deeper? Another topic for class discussion.
Standard 4: Research As Inquiry:
BRASS application to AI: Are responses generated by an AI on a specific inquiry enough to help the researchers understand the whole field of inquiry to identify gaps or scope of investigation?
Legal Research: If a case is delivered in response to AI inquiry, will students read and analyze the case? In class, I will have to emphasize not relying on a generative AI summary of a case for analysis. The generated summary may help students narrow down the cases to those which look most applicable to their factual situation. But students need to appreciate that the cases delivered by the generative AI are the start of the research process, not the end.
Standard 6: Searching as Strategic Exploration:
BRASS application to AI: Is the researcher using an AI tool with access to the resources needed to provide relevant answers? Does the AI response point the researcher to further sources?
Legal Research: Do students know which generative AI tool will be most useful for their research? Can they look at results critically and see something may be missing? Does the AI tool point to additional sources to build on the initial query? This is not even getting into the students’ ability to draft effective prompts. That is another area I might have to build into my classes.
One potential upside to these new tools is the opportunity for us, as legal research instructors, to reset, take a step back, and talk with students about information literacy. I hope that this fall, I will be able to address some of the long-standing issues with students’ information literacy skills while introducing new generative AI tools as part of the research process.
What are your thoughts on teaching generative AI in legal research this fall?