Breaking

UC Berkeley Law Bans AI Use in Graded Coursework and Exams

July 5, 2026 at 17:24 EDT

Policy & Regulation
Education

UC Berkeley School of Law has adopted a policy that broadly bans the use of AI in graded coursework and exams starting in the summer of 2026. As "hallucinations"—in which generative AI fabricates nonexistent case law—have repeatedly caused problems in the legal profession, one of the nation's top schools has moved toward a strict stance.

Summer 2026 · UC Berkeley School of Law

A Top Law School Bans Generative AI on Graded Work

After a wave of fabricated case citations in court filings, Berkeley Law is adopting the most far-reaching AI restriction among the nation's T14 law schools — with prohibition as the default rule.

AI tasks banned on graded work: conceptualizing, outlining, drafting, revising, translating, editing

600+

Cases nationwide reportedly involving AI-fabricated citations

Sole exception — identifying sources during paper research

Monetary sanctions by case

Fabricated AI citations are already costing lawyers

$5K

Wyoming

$10K

California
21 of 23 citations fake

$110K

Oregon

Bars scaled to sanction amount — Oregon's penalty is 22× Wyoming's.

The core rule

Students bear full responsibility for accuracy — citing a non-existent source creates a presumption of prohibited AI use.

No AI in exams; no uploading assignments, course materials or recordings to AI tools.

Exceptions only for AI-fluency courses or written permission from an instructor.

Supporters say

Protects the core of legal education — the capacity to reason and write independently, without leaning on AI.

Critics warn

It may diverge from real practice, and its nature as a relaxable "default" rule is not widely understood.

The shared lesson: AI can be useful, but its output must always be verified by a human.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.