Law schools tell students to put AI away

The writer is a contributing columnist

These days, we are all struggling to balance the human and the machine in the workplace. Take the training of lawyers. Generative AI can already do much of the grunt work normally given to young, inexperienced attorneys, but if we let it, how will they ever become sage old pros?

America’s top law schools are figuring this out on the fly: for the new academic year, the University of Chicago Law School will ban phones, tablets and laptops from core first-year classrooms, so students cannot use technology shortcuts that “stunt intellectual growth”.

“The whole point is to do things the hard way — because that’s how you learn,” says William Hubbard, chair of the law school’s AI committee. For generations of Chicago law students, that has meant debating cases and legal principles in real time, in front of professors and peers, using what is known as the “Socratic method”. AI is getting in the way of that, he tells me.

“The power of AI tools and the legal profession’s eagerness to adopt [them] are at an inflection point, so we really felt a sense of urgency to act,” he says. But not by banning AI outright.

“We take as a starting point that our students are going to use it, so how do we adapt our curriculum to reflect that, not wish it away?” asks the law school’s dean, Adam Chilton. The school says it wants students to learn “with, without, and about AI”. That means permitting it for things like brainstorming and studying outside the classroom, in legal clinics and in classes that teach about it.

Hubbard referred to a recent study of Chinese students that found the biggest learning losses among those who used AI to shortcut homework. “The key takeaway is not whether you use AI but whether you put in the time,” he told me. The “paradox” of AI in the law is that it is “an amazing productivity tool” in the workplace. “But in education, that gets it completely backward: what matters is that you put in the time”.

For undergraduates, University of Chicago is taking an even stricter approach: they will be barred from using technology in core required social science classes; they must use “analogue” practices like reading texts on paper instead. “We teach students the skills they need for life, including the ability to make decisions if the power goes out,” Jenny Trinitapoli, master of the university’s social sciences collegiate division, says.

The University of California, Berkeley, School of Law, also has a tough new policy that bans the use of generative AI tools “for aid in conceptualizing, outlining, drafting, revising, translating, or editing any work submitted for credit”. “Our default is that when a student is turning in work, it shouldn’t be that of Claude,” law school dean Erwin Chemerinsky tells me (though he adds that individual professors have some discretion). “If AI is doing the thinking and writing for them, we aren’t teaching them how to do it”.

On the other hand, Kiel Bowen, partner at law firm Mayer Brown, thinks that in some cases AI can train better lawyers. Normally, young associates learn partly from “nuggets” of wisdom from senior lawyers. But sometimes, he says, AI is a better “sparring partner” because it can “see around corners”. This is hard even for the most expert lawyer because “we all have our blinders on from our own human experience and from our own professional experience”.

Still, Hubbard thinks there is too much about the law that is quintessentially human to rely overly on machines. “There are core aspects of the practice of law that are not about what’s in the books but what is in the human heart and soul,” he tells me. “There is something very human about the lawyer’s job, someone is in a bad place and they know they have someone in their corner fighting for them.” He hopes AI can help lawyers, including by freeing up time for them to do what really matters to clients — like returning their phone calls.

But who will pay for it? “We need to be thinking about how we can train people to develop judgment in a world in which some of that training might no longer be cost-effectively done by humans,” says Chilton. It’s the same question for all of us: how much AI is too much? And what are we willing to sacrifice?