American law schools are taking sharply different approaches to artificial intelligence, with some restricting technology in classrooms while others are making AI training part of legal education.
For some students returning to campus this fall, the change is immediately visible: laptops and phones are off limits during class. At other schools, students are being asked to use AI as part of their coursework. In some cases, both policies exist side by side as institutions try to preserve traditional legal training while preparing students for a profession increasingly shaped by AI.
At least a dozen US law schools introduced or revised AI-related policies over the summer. The measures range from classroom device restrictions to mandatory courses designed to familiarize students with emerging technology.
The debate reflects a larger question confronting higher education: how can students learn to think independently when AI tools can produce answers almost instantly?
The University of Georgia Law School, for example, introduced an AI policy that generally bars devices from classrooms while also requiring students to complete at least one course dealing with AI or technology.
Dean Usha Rodrigues said the classroom restrictions were influenced in part by concerns that students could turn to AI during traditional Socratic-style teaching, where professors call on students without advance notice.
The objective, she explained, is not simply to produce the correct answer. Students are expected to reason under pressure, consider competing arguments and critically assess what their classmates say.
That concern carries particular weight in legal education. Lawyers deal with confidential information and have professional responsibilities that require them to understand the work they produce, whether they are handling major corporate disputes or cases involving severe criminal penalties.
Berkeley Law has taken an especially restrictive approach. Its policy limits AI use in several aspects of student writing, including brainstorming paper topics, summarizing legal rules for papers and even using AI to correct grammar or identify repetitive language.
The policy has also sparked debate among students. AI at Berkeley Law, a student-run organization, argued that there is an important distinction between allowing AI to replace legal reasoning and teaching students how to incorporate the technology responsibly into legal workflows.
Faculty members, however, retain some flexibility. Berkeley professor Colleen Chien said individual instructors can depart from the default policy when doing so supports their teaching objectives.
The disagreement is playing out across campuses as students themselves increasingly see AI knowledge as a career necessity. Similar AI-focused student organizations have emerged at law schools around the country, reflecting expectations that future lawyers will encounter the technology routinely in professional practice.
From restrictions to required AI training
The shift is not limited to classroom bans.
Suffolk University Law School dean Andrew Perlman has compiled AI policies from 180 law schools into an online database. According to the database, at least 36 schools now require some form of AI instruction, whether through orientation programs, first-year courses or dedicated classes.
Perlman said the legal education landscape has increasingly moved toward accepting that AI is not a temporary development.
Some schools are attempting to introduce students to the technology only after establishing a strong foundation in traditional legal reasoning.
The University of Chicago Law School adopted a policy barring laptops and phones from required first-year classes. The stated aim is to ensure that students develop critical, strategic and independent thinking before being taught how to use AI tools.
Other institutions are taking a more permissive approach.
Columbia Law School allows students to use AI to test arguments, seek criticism, explore different ways of expressing ideas and correct spelling and grammar. The school nevertheless requires the underlying work to remain the student’s own, with the ideas, analysis and expression reflecting their individual judgment and creativity.
The University of Michigan has adopted another variation. Students can use AI when brainstorming or researching papers but cannot use it to draft, edit or revise the final work.
That middle-ground approach has drawn support from some students who see both sides of the debate. Excessive reliance on AI could interfere with the development of legal skills, while overly restrictive policies could prevent students from exploring productive applications of the technology.
Examinations remain one area where restrictions are widespread.
University of Houston law professor Seth Chandler said nearly all law schools prohibit AI use during exams. Some schools have moved away from traditional take-home assessments and instead conduct examinations in classrooms using software that blocks internet access.
The reasoning is straightforward: when students are being assessed, faculty want the relevant legal knowledge and analytical ability to come from the student rather than an AI system.
The result is an emerging split in legal education. Some schools are drawing a line between learning law and learning to use AI, while others are trying to teach both at the same time.
What appears increasingly clear is that US law schools are no longer deciding whether AI belongs in legal education. They are now working out where it belongs, when students should use it and how much of the traditional classroom should remain untouched.
