AI and the Evolution of Law and Legal Education

Professor Nicholson Price teaching — Professor Nicholson Price has taught Artificial Intelligence and the Law three times since it was first offered during the Winter 2022 term.

As society pieces together how artificial intelligence (AI) fits into the education puzzle, Professor Nicholson Price invites Michigan Law students to wrestle with questions of how the law shapes AI and how AI shapes the law.

In his seminar Artificial Intelligence and the Law, which he has taught three times since the Winter 2022 term, his students dive into discussions at the frontier of legal thought: How should we think about regulating AI in different contexts? How does the use of AI in medicine or autonomous vehicles, for example, affect liability doctrines? What impact does AI have on the intellectual property system? How do privacy and transparency factor in?

“It’s been great,” he says. “That course has been full every time I’ve taught it. Procedurally, I ask my students to use AI to help them draft some of their responses and tell them that in others they can’t. And I have them reflect on what the experience of using AI tools is like—how that shaped their writing process, things like that—so they can engage with the technology on a more direct level.”

While AI’s surge into the public’s consciousness has been relatively recent, thanks to the launch of tools such as ChatGPT, Price has researched AI for the past decade. He also has collaborated on various projects with faculty in U-M’s College of Engineering; the College of Literature, Science, and the Arts; the Medical School; and the School of Public Health.

Price has a PhD in biological sciences in addition to a JD. He has leveraged his research—which focuses on the use of AI in medicine and how it shapes biomedical innovation—into his teaching at the Law School.

“It’s been really nice to bring together students from very different backgrounds,” he says of the interdisciplinary curiosity. “I’ve had computer programmers in my class. I’ve had people who have interacted with AI systems in medicine. I’ve had engineers, including in the autonomous vehicle space. I’ve had people with no background at all. And to see this cross-cutting interaction of AI across the different areas is quite interesting.”

While he does challenge students to use AI for some of their assignments, this doctrinal class is different from the skills-based “sandbox” class that Professor Patrick Barry teaches.

Instead, students discuss legal scholarship as well as cases, statutes, regulations, and other materials; hear from guest speakers who are active in the field; and even read science fiction short stories such as Isaac Asimov’s "Franchise," a futuristic tale about a presidential election with a single voter.

No replacement for learning

Like everything related to AI, students also have to keep up with the constant changes in technology. Today’s AI could be very different from AI in two to three years.

“I certainly can’t keep up with all the changes,” Price says. “I do my best with changes related to medicine, which is already like drinking from a fire hose. But one of the nice things about teaching a seminar like this is that I have the students do current event presentations, and that helps me learn about what’s going on.”

While AI can play a role in enhancing legal learning, it is far from replacing learning.

Price mentions the popular analogy that using AI to do some of the basic tasks is like using a robot to lift weights for you at the gym. By outsourcing that training to AI, students won’t learn how to think.

“I want to make sure that whatever students are learning about how to use AI, they are not skimping on the ‘how do I do the thinking on my own?’” he says. “At least for the foreseeable future, that’s going to remain absolutely essential.”

Case in point, feedback he receives from students in his seminar shows that AI can fall short of expectations.

“I sometimes ask students to use AI and to write a reflection about it, and those reflections are quite interesting,” he says. “Some students report that they’re really disappointed by the tools. They thought it was going to be tremendously helpful, but it ended up taking them much more time than before and didn’t sound as much like their voice as they wanted it to.”

While he hesitates to label AI as a foundational skill, Price acknowledges that it most likely will show up in legal practice when students start their careers. Therefore, they have to know about it, at least to the extent that they can discern that a model isn’t making something up or how to interact with an AI model in a way that makes it more likely to give them useful results.

“If we move to a phase where it turns out AI can do all the complex legal thinking for you, and you don’t need anybody to check it because it’s always right,” he says, “then we’ve got a lot of other transformational changes to worry about.”