The following is a guest post by Rick Bales (Ohio Northern)
I have long been a fan of experiential learning, and have tried to incorporate group-based problems into my courses as much as possible. But writing good problems is hard work, and problems borrowed from elsewhere do not always fit the specific pedagogical goal I might have for each class period.
Over the last eight months or so, however, I’ve been using ChatGPT4 to help me draft custom-made problems. In nearly every class last semester, we spent the first half of class discussing the reading material, and the second half doing a group-based exercise demonstrating the material from the first half. This summer I’m teaching in China, and AI has helped me customize the exercises for my Chinese audience.
The best part: with the help of AI, I can create an exercise in under an hour, and often in substantially less time than that. I can build this into my normal prep time, whereas before, I would need a ton of extra time to essentially redesign the course.
For example, in my Employment Law course, after covering a unit on employment contracts, I wanted my students to be able to identify and correct problems in a poorly drafted employment contract. (Earlier exercises had them draft specific contract clauses; later in the semester, I will have them draft entire employment contracts.) Here’s the prompt I gave ChatGPT4:
I want to give students a copy of a badly drafted employment contract and have them identify the items in the contract that are badly drafted. The contract should contain lots of ambiguities. It should be unclear whether it is a term contract or an at-will contract. The circumstances under which the employee can be fired should be unclear. The contract should be ambiguous about what happens if the employee wants to quit. Please draft the badly drafted contract, formatted as if it were a real employment contract. Separately, please identify the specific problems with the badly drafted contract.
Within seconds, ChatpGPT4 gave me exactly what I had asked for. The problems and ambiguities were representative of what you might expect to see in an employment contracted drafted by a nonlawyer. I spent a few minutes editing it, then copied-and-pasted the problem from ChatGPT4 into Word and adjusted the formatting to fit my preferences. Creating the exercise took me less than 15 minutes from start to finish.
Here is a second, more complex example. In China this summer I am teaching an ADR course. A large proportion of my students here will leave law school to work for international law firms doing cross-border work, or for large Chinese companies doing business with non-Chinese companies. I wanted to create a negotiation problem that they could work in one class period that would give them a taste of what they might see in law practice. Here is the prompt I gave ChatGPT4:
Please create a negotiation problem for students. The problem should have four complex issues to be negotiated. For each issue, the parties’ initial demand should be outside the range of potential agreement, but the acceptable outcomes for each party on each issue should overlap sufficiently so that agreement is possible. The dispute should be a commercial dispute between an American company and a Chinese company. Please create (1) a detailed set of general facts for both parties in narrative form, (2) a detailed set of private facts for each party in narrative form, and (3) a teacher’s guide identifying the issues to be negotiated and the range of possible agreement. Make all names gender neutral.
Again, ChatGPT4 gave me what I asked for. I could have specified the particular issues to be negotiated, but I wanted to see what ChatGPT4 came up with – and because this is an ADR rather than substantive-law course, I did not have any particular substantive issues in mind. Though the issues ChatGPT4 came up with were not really all that complex, they were suitable for my purpose. And without me asking for it, ChatGPT4 created two roles for each party: a CEO and a chief negotiator.
That gave me an idea: I could introduce extra complexity by creating some discord between the two members of each team. I gave ChatGPT4 a follow-up prompt:
Consistent with everything above, please create a set of private facts for each participant (the CEO for each party and the chief negotiator for each party) creating some sort of conflict between the CEO and the chief negotiator for each party.
I read through the problem to ensure there were no glaring errors or weird facts, and everything looked good. I spent about 15 minutes proofing everything, transferring everything to a Word document, and formatting it to my liking. I then spent another 10 minutes splitting the material into separate documents: (1) a set of general facts that would go to all students, (2) private facts for each side, (3) private facts for each character, and (4) a “teacher’s guide” summarizing all the conflicts and the room for resolution – a total of 8 documents. The entire drafting experience took less than an hour from start to finish.
I have enjoyed using ChatGPT4 to help draft problems for class. I have even more enjoyed the effect this has had on my teaching and my students’ learning. Students are much more engaged in the doctrinal part of each class because they know the material will be important to the exercises they will do the second half of class. Students love the exercises because students are naturally competitive and they can see how the exercises are preparing them for what they will be doing in law practice.
Playing with ChatGPT4 is an ongoing experiment for me. Writing a prompt that formerly took a half hour of trial, error, and a half dozen follow-ups now takes 5-10 minutes with one or two follow-ups. I’m taking notes as I continue to experiment, and will be writing up my results for the next Teaching Issue of the St. Louis University Law Journal.
-Rick Bales
Recent Comments