In April 2008, the dean of the University of Chicago Law School shut off the Wi-Fi in his main classrooms.
The last straw: one student shopping for shoes, and the shopping spreading down the row in twenty minutes. The ban just faded, the way rules do when the tide keeps rising.
Last Thursday the same law school went much further, and this time I think it will hold.
The headline everyone ran: no laptops, tablets, or phones in the nine core 1L courses, starting this fall. The dean says students were typing the case name into Claude or ChatGPT and reading the summary back when called on. The Socratic method survived 24 centuries without meeting an interlocutor willing to do the student’s half of the dialogue.
But the real story is the policy itself. The same document builds AI into the required first-year writing course: write without AI first, then use it for research and revision. What Chicago published isn’t a ban. It’s a map, like the framework we use at Alma when we build AI-enhanced courses and degree programs with schools: course by course, decide where AI is banned (or rather won’t help — more on that below), where it’s expected, where it’s encouraged, and where it’s required.
One sentence is worth the whole document: “During the 1L year, the value of effortful struggle, even with concepts that are second nature to experienced lawyers, is paramount.”
When we built Alma, we borrowed a lot from law and business education (which borrowed from Socrates). Instead of just giving every student a personal tutor, we wanted every student to have the opportunity to be cold called, the way 1Ls are. When the AI conversation is the learning experience, you can’t just hide in the back behind your laptop.
And students don’t like it. Being cold called is uncomfortable, and good learning is meant to happen at the border of the uncomfortable. Students get frustrated when the AI drills them and makes them clarify what they mean. Sometimes that’s a prompting problem, the AI overdoing the push. More often the student is resisting mental effort, and our job is to nudge them back toward it, gently. Being taught by AI doesn’t mean the AI will do your work, or hold you to lower standards than a human instructor.
But Chicago’s solution is theirs alone. They can take devices away because their classroom is a room. In online programs, the classroom is the device, and online, async programs are the most accessible serious education we have, usually the cheapest. So the burden shifts from rules to design: assignments done live, in the student’s own voice and on camera, drawing on their own experience and past work. No rule against AI, just an assignment where it won’t help. And if you can use AI to get through that and we can’t catch you, that’s pretty awesome.
The 2008 Wi-Fi ban faded because it was a wall against a tide. This one might hold, because it isn’t a wall. It’s zoning: here we struggle, here we build, and the struggle comes first.