UC Berkeley law tightens Artificial Intelligence rules without banning it

UC Berkeley Law is adopting a stricter Artificial Intelligence policy that limits student use in core legal work while preserving space for specialized courses and instructor discretion. Faculty behind the change say the aim is to protect foundational lawyering skills as generative tools become more capable and more common in practice.

UC Berkeley Law is introducing a stricter Artificial Intelligence policy that bars students from using the technology for conceptualizing, outlining, drafting, revising, editing, translating, or for any purpose in an exam situation. The school is not trying to ban the technology outright. The policy is designed to ensure students develop the core skills of legal analysis and writing before relying on automated tools. Chris Hoofnagle, a professor involved in writing the policy, said the focus is preserving the professional value of lawyers’ own judgment and making sure students can assess Artificial Intelligence output rather than depend on it.

Faculty concluded that the 2023 Artificial Intelligence use policy was too permissive as generative models improved. The earlier rules allowed uses such as brainstorming and conceptualization, including asking a chatbot to suggest a paper topic. Hoofnagle said newer large language models can now produce a research paper from start to finish, forcing the school to reconsider how much reliance should be allowed. UC Berkeley Law’s new policy, which will go into effect this summer, was approved by a faculty vote, though instructors can deviate from it, and some Artificial Intelligence-focused courses will operate under different standards.

The tighter restrictions come even as legal employers increasingly expect graduates to be comfortable with Artificial Intelligence tools. Hoofnagle said students want these courses and are learning during summer work that law firms already use the technology extensively. Startups like Harvey and Legora are competing for the estimated ? trillion global legal market, and Harvey has offered free access to law schools as part of its expansion. Hoofnagle noted that Stanford Law School, which had a stricter policy when Berkeley adopted its initial rules in 2023, is part of Harvey’s law school alliance program.

Enforcement remains difficult because Artificial Intelligence features are spreading across standard research and search products. Hoofnagle acknowledged loopholes and said policing misuse is increasingly hard when search engines and legal databases add large language model summaries. He said Lexis and Westlaw now include generated summaries, making a clean prohibition unrealistic because the school obviously cannot ban search. Berkeley Law has seen an uptick in misconduct cases and has shifted more take-home exams to in-person testing. Princeton recently announced the most significant change to its honor code in 133 years, and as of July 1, all in-person examinations will be proctored, with the rise of Artificial Intelligence cited as one reason.

52

Impact Score

EU eases parts of its Artificial Intelligence Act

The EU has agreed targeted changes to its landmark Artificial Intelligence Act, delaying some deadlines, narrowing parts of the high-risk category, and cutting overlapping compliance requirements. The package also adds a ban on tools that generate non-consensual sexually explicit images.

Artificial Intelligence food leaders see disruption moving at uneven speeds

Food technology leaders say Artificial Intelligence is reshaping the food system quickly, but adoption remains uneven across functions such as strategy, compliance, and research and development. Companies already see immediate returns in some areas, while others expect broader impact to take longer.

Lisa su pitches AMD as China’s alternative to NVIDIA

AMD used its Shanghai developer event to position China as central to its roadmap and to court developers looking for an alternative to NVIDIA’s CUDA ecosystem. The strategy focuses less on headline chip specs and more on migration support, open-source tools, and long-term bets on the next wave of Artificial Intelligence applications.

DeepWeb-Bench tests limits of deep research models

DeepWeb-Bench is positioned as a tougher benchmark for evaluating whether frontier language models can handle real deep research tasks beyond existing tests. Results point to derivation and calibration, rather than retrieval, as the main weaknesses in current Artificial Intelligence systems.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.