Chicago bans laptops for first-year law students: betting that human judgment, not AI, still earns the degree

By FindLaw · July 15, 2026.

The University of Chicago, one of the fourteen most prestigious law schools in the U.S. (the so-called T14 group), has published "Rethinking Legal Education in the AI Era," its official strategy for incorporating artificial intelligence into the training of lawyers. The document, signed by Dean Adam Chilton, is the result of three years of work by an internal committee created shortly after the emergence of ChatGPT in 2022. The school sums it up with a phrase that functions as a program: teaching students to think "with, without, and about" AI.

The concrete facts are more striking than the statement of intent. In the core first-year courses—Civil Procedure, Torts, Elements of the Law—laptops, tablets and phones are banned in the classroom, with the sole exception of one "scribe" student who takes notes for the rest. Exams will be taken without internet access. Foundational legal writing will still be taught first without AI, and only afterward will layers of assistance be introduced, aligned with what students will encounter in their summer internships. For upper-level courses, the most significant change is that every student will have to orally defend before their professor the substantial research paper they submit, an explicit safeguard against replacing one's own writing with that of a model. At the same time, the school has purchased licenses for the leading legal AI tools for its legal clinics, where students provide real assistance to people who need access to justice.

What's interesting is not that a university regulates the use of AI—nearly all of them already do—but the deliberate asymmetry of its approach: the earlier and more foundational the learning, the more AI is restricted; the more advanced and applied, the more it is encouraged. It is a pedagogical hierarchy with an implicit thesis behind it: legal judgment cannot be outsourced from day one, because if it is, nothing remains to verify later on. The oral defense of the SRPs (substantial research papers) is the most revealing piece of the plan: it does not try to detect AI use with forensic software, but with the oldest question in the profession—can you defend in person what you have written?

This connects directly with something we have already noted when analyzing AI's impact on the legal sector: the law firm pyramid is narrowing at the base. The research, first-draft writing and document review work that traditionally trained junior lawyers is precisely what generative AI does better and cheaper. The underlying problem UChicago Law is confronting, though it doesn't put it in these terms, is a talent supply-chain problem: if students delegate to AI the tasks that once served to develop judgment, the profession loses its pipeline of judges, partners and prosecutors with judgment of their own within a generation. Banning laptops in the first year is an almost anachronistic gesture, but it works as a deliberate firewall at the stage where that judgment is forged.

In the short term, this is uncomfortable honesty about a hard transition: law firms are already reducing the hiring of junior associates because much of their historical work is automatable, and law schools will have to redefine what "being ready to practice" means when much of the entry-level work disappears or is transformed. UChicago Law does not pretend this problem doesn't exist: it acknowledges that its graduates will use AI in their summer internships and in their professional lives, and that is why it is also building AI labs and licenses for the clinics, rather than simply banning it.

In the long term, however, the design points in the direction we uphold at Zendoric: AI does not eliminate human value, it shifts where it resides. If the lawyers who emerge from programs like this master both the responsible use of AI and the judgment that a model cannot reliably fake, it opens the door for quality legal work—including the access to justice that is today prohibitively expensive for millions of people—to become cheaper and more widely available without degrading its reliability. Legal clinics with AI geared toward free legal aid are, on a small scale, a rehearsal of that abundance: more capacity to help, without human judgment ceasing to be the final filter.

The risk, of course, is fragmentation. One T14 school making a move does not guarantee the rest will, and a patchwork of standards—some schools demanding critical mastery without AI, others handing out the degree to whoever simply knows how to use an assistant well—could erode trust in the law degree credential itself. It remains to be seen whether bar associations and admission exams to the profession will follow the same criterion Chicago is applying in the classroom, or whether legal training will split into different speeds depending on which school—and which generation of lawyers—decided to take the problem seriously in time.