UC Berkeley School of Law’s stricter AI policy prohibits students from using AI for things like brainstorming, but a professor who helped write the new rules said the school has no intention of banning the technology entirely.
“Our policy is to develop students with the foundational skills needed to be AI lawyers,” Chris Hoefnagle, a professor at the University of California, Berkeley School of Law, told Business Insider.
Hofnagl said the school believes the 2023 AI Usage Policy is “too permissive” in allowing students to use AI, especially given advances in generative AI models since then.
“You can practically write your research paper soup on nuts,” Hofnagl said. “So as LLM capabilities have grown, we have had to rethink our students’ reliance on LLMs.”
Chris Hoofnagle, a law professor at the University of California, Berkeley, helped develop the new policy. Provided by University of California, Berkeley Law
A new UC Berkeley Law School policy scheduled to go into effect this summer will not allow students to use AI for any purpose, including conceptualizing, outlining, drafting, revising, editing, or translating exams. The 2023 policy allows for the use of AI for brainstorming and conceptualization, such as having chatbots help come up with topics for a paper.
Hofnagle said the new policy was approved by a vote of the faculty, but it is possible for faculty to deviate from the policy. He also said that certain courses focused on AI will follow different standards.
The goal, Hofnagle said, is to ensure that first-year law students learn the fundamentals of being a lawyer, including “how to read, analyze, and write about cases persuasively.”
“Of course, the question becomes, what is the added value of a lawyer?” he said. “And if that lawyer cannot use their own analytical judgment to evaluate the output of the AI, that lawyer has little value. That is the purpose of our policy.”
Overall, the expectation and demand from law firms is for students to graduate proficient in using AI, Hoofnagle said.
“Students are asking for these courses and are learning over the summer that law firms are already making extensive use of AI,” he said.
Startups like Harvey and Legora are fiercely competing for the estimated $1 trillion global legal market. Harvey has sought to expand its presence by offering free access to law schools. Stanford Law School, which Hoofnagle said had a stricter AI policy when Berkeley introduced its first AI policy in 2023, is part of Harvey’s Law School Alliance program.
Hofnagle acknowledged that there are loopholes in the policy. After all, policing AI is becoming a very difficult task in a world where search engines are adding AI overlays to ensure continued relevance. Standard searches in Lexis and Westlaw now also display summaries generated by LLM, he said.
“There is no clear answer to this,” he said, adding that schools “clearly cannot prohibit searches.”
The school is trying to keep it that way.
Law schools and faculty alike are trying to stay ahead of the wave of advances in AI and the challenges that come with it.
Princeton University recently announced the most significant changes to its honor code in 133 years, the Daily Princetonian reported. Starting July 1, all in-person exams will be proctored. The emergence of AI is cited as one of the reasons for this change.
Hofnagle said Berkeley’s law has led to an “increase” in cheating cases, resulting in more take-home exams being converted to in-person exams. In-person exams are administered on computers with special software that prevents you from accessing the Internet or using copy-and-paste functions. Despite the precautions, fraud is still possible, he said.
“You can’t protect yourself from everything in the world, but there are very strong indications that if you cut corners, students will ultimately pay the price when they have to take in-person exams,” he says.