A Top US Law School Just Banned AI from Almost Every Corner of Student Work
Starting summer 2026, students at one law school can't use AI to brainstorm, draft, revise, or even run a spellcheck on graded work. Here's what the policy says and why it matters beyond the classroom.
Key points
- A US law school policy taking effect summer 2026 bans AI tools from brainstorming, outlining, drafting, revising, editing, or translating any work submitted for a grade.
- AI is banned completely during exams, with no exceptions for any purpose.
- Students can't upload course materials, including readings or class recordings, into AI systems.
- The one permitted use: finding sources such as cases or statutes, not summarising or explaining them.
- Citing a source that doesn't exist will be treated as evidence of prohibited AI use.
Law school teaches you to think. One institution has decided that AI is now threatening that core goal directly enough to warrant a near-total ban.
The policy, first flagged on Hacker News, takes effect in summer 2026 and covers every graded piece of work a student submits.
What exactly is banned?
Pretty much everything you'd reach for a chatbot to do. The school lists concrete examples so there's no grey area.
Sketching a paper topic with an AI tool? Banned. Drafting a single paragraph summarising a legal rule? Banned. Spotting repetitive sentences? Banned. Fixing grammar? Banned too.
Exams are a complete lockout. No AI, for any reason, at any point.
The translation ban is pointed. A student who writes in their first language and then asks AI to render the work in legal English is, the policy argues, bypassing the development of their own fluency with the language of the law.
| Permitted AI use | Banned AI use |
|---|---|
| Finding cases, statutes, or secondary sources | Brainstorming or framing a topic |
| (Nothing else) | Outlining or structuring a paper |
| Drafting any passage | |
| Revising or editing submitted work | |
| Translating work into English | |
| Any use during an exam |
Why so strict?
The school is direct about its reasoning. Good lawyering is built on cognitive skills: reading primary texts closely and constructing original arguments under pressure. If AI does those steps, students never build the mental muscle. A competent-sounding document isn't the same as a lawyer who can think.
Fairness matters too. If some students use AI and others don't, grades stop measuring what they're supposed to measure.
The enforcement problem is real. We reported on 9 August that AI writing detectors have suspended students and derailed careers while their own makers admit the tools are frequently wrong. A ban this broad has to rest on something sturdier than detection software.
Does any AI use stay allowed?
One narrow opening remains. Students can use AI to identify sources, such as a relevant court case or statute, but must then read and evaluate those sources themselves. The school makes clear that students are responsible for the accuracy of everything they submit.
There's a catch built into that permission. Cite a source that doesn't exist, which AI tools sometimes invent through a process called hallucination (where the software confidently states something that's simply false), and the school will presume you broke the rules. The burden then falls on the student.
Instructors can opt out of the default ban for courses specifically designed to teach AI fluency. Any such exception must be given in writing.
What does this mean for future lawyers?
The policy captures a tension the legal profession is quietly debating. Fluency with AI is only useful if the person using it can catch its mistakes and take ethical responsibility for the output. Building those underlying skills first, without the shortcut, is the bet this school is making.
Whether other institutions follow is a question the next few admissions cycles will answer.
Common questions
Can a student ever get permission to use AI?
Yes, but only if their instructor explicitly grants it in writing before the work begins. The student must also disclose what AI they used. When in doubt, students must ask their instructor and get a written answer before using any AI tool.
What counts as evidence of AI use?
Citing a legal source that doesn't actually exist raises a presumption of prohibited AI use. AI tools sometimes generate fake but convincing-sounding citations, a problem called hallucination, where the software confidently states something that is simply not true.
