UC Berkeley Law Implements Strict AI Policy

The University of California, Berkeley School of Law will enforce a new policy starting in summer 2026 that prohibits students from using artificial intelligence (AI) for drafting or editing coursework and exams. This announcement was shared on X and has garnered significant attention.

Richard Albert, a law professor at the University of Texas at Austin, highlighted this initiative, noting that while students can use AI for research purposes—like identifying statutes and case law for dissertations—they cannot use it to brainstorm topics, create exam summaries, or translate their work into English. Other restrictions include asking AI to generate summaries of legal rules, identify duplicates in papers, or correct grammatical errors, all of which are considered prohibited editing tasks. Moreover, students are also barred from uploading any course materials, such as assignments or class notes, to AI systems.

The policy does allow for some flexibility; during courses specifically designed to foster AI fluency, instructors might set different rules. They are required to inform students in writing with adequate notice about any deviations. Students must also disclose when they use AI to their instructors and get written permission if their use conflicts with course policies.

This new directive comes on the heels of a previous AI policy introduced in 2023, which allowed limited use of AI tools for research and grammar correction, but prohibited their use for writing assignments or exams. The former policy was crafted by Professor Chris Hofnagle and a few colleagues.

In response to a rise in academic dishonesty linked to AI misuse, Hofnagle played a significant role in drafting the new policy. He expressed that the intention behind these changes is to safeguard foundational legal skills, emphasizing that students should first develop their analytical abilities before incorporating AI into their work. Interestingly, he mentioned utilizing Onetutor.ai in his classes.

Overall, this policy aims to strengthen the cognitive skills essential for legal education, making it clear that AI tools cannot be used to create, draft, revise, or translate work that students submit for credit. Comments from a UC Berkeley Law spokesperson directed inquiries to the new AI policy document and Hofnagle’s post on X for further clarification.