Friday, July 28, 2023
Greetings from SEALS, where I've just left a packed room of law professors grappling with some thorny issues related to ChatGPT4, Claude 2, Copilot, and other forms of generative AI. I don't have answers to the questions below and some are well above my pay grade, but I am taking them into account as I prepare to teach courses in transactional skills; compliance, corporate governance, and sustainability; and ethics and technology this Fall.
In no particular order, here are some of the questions/points raised during the three-hour session. I'll have more thoughts on using AI in the classroom in a future post.
- AI detectors that schools rely on have high false positives for nonnative speakers and neurodivergent students and they are easy to evade. How can you reliably ensure that students aren't using AI tools such as ChatGPT if you've prohibited it?
- If we allow the use of AI in classrooms, how do we change how we assess students?
- If our goal is to teach the mastery of legal skills, what are the legal skills we should teach related to the use of AI? How will our students learn critical thinking skills if they can rely on generative AI?
- How should we keep up with the rapid pace of change?
- How will adjuncts use AI with our students if they are already integrating it into their practice? Alternatively, will adjuncts see the use of AI as cheating?
- If students use papers as writing samples, should there be attestations indicating that they are AI free? Same question for journals/law reviews.
- Can clinicians and others use generative AI to help with access to justice? If so, how can we ensure that the information is reliable and not a hallucination??
- How should schools assess faculty coming up for promotion and tenure? Will junior faculty feel pressured to rely on AI to be more productive?
- Can generative AI be helpful with students with disabilities and neurodivergent students? AI tools can help with creating study schedules, note taking (organizing by topic), time management, summarizing large articles, staying on task, academic support tool, ascertaining how long will tasks take, planning meals and more. If a policy prohibits the use of generative AI in the classroom, should its use be a reasonable accommodation?
- Do we as faculty members have the growth mindset to deal with this change? Or will we teach the way we always do, which may do a disservice to our students. How do we prepare our students to deal with generative AI in practice?
- Do you need a uniform policy or should each professor have their own policy? Should the default policy be that students cannot use it for work that gets academic credit unless the professor has specifically opted in?
- Should we embrace AI especially for students who can’t write? Is using ChatGPT any different from using a calculator? Is it any different from asking a partner for a template so you don't have to start from scratch?
- Should we use more in-class exams? Should they be closed book? Do we need more oral presentations? How might this affect space planning at exam time?
- Should class participation count for more than it already does?
- If you're not familiar with generative AI tools, where should you start?
How many of these questions have you asked yourself, your colleagues, or your dean? If you have some best practices or thoughts, please share them in the comments.