Berkeley Law sets Summer 2026 AI policy banning use in exams and credited coursework
UC Berkeley School of Law will allow AI only for limited source identification in papers, while prohibiting use for drafting, editing, translation, and exam work
UC Berkeley School of Law’s Summer 2026 AI policy prohibits AI use for exams and most credited coursework tasks.
UC Berkeley School of Law has introduced an Artificial Intelligence Policy, effective Summer 2026, that prohibits students from using AI to conceptualize, outline, draft, revise, translate, or edit work submitted for credit.
The policy also bans AI use for any purpose in any exam situation and prevents students from uploading course materials, including assignments, readings, slides, class recordings, or other class content, into generative AI systems.
Berkeley Law frames the policy around legal education, ethical obligations, and the need for students to develop the cognitive skills used in lawyering. The policy says future lawyers may need to use AI fluently, but that AI use must be paired with the ability to deploy the technology strategically, assess its output, and uphold obligations to clients and the legal system.
The default rule applies to work submitted for credit, but instructors can set a different rule in writing where they decide another approach is appropriate. That includes courses intentionally designed to teach AI fluency.
Students can use AI for research on papers only for the limited purpose of identifying sources, such as cases, statutes, or secondary sources. Berkeley Law states that students remain responsible for research accuracy and all other aspects of submitted work.
Policy blocks AI use in core writing tasks
The policy gives specific examples of prohibited AI use, including asking an AI tool to brainstorm a paper topic or thesis, propose an organizational structure, draft a paragraph summarizing a legal rule, identify repetitive passages, correct grammar, generate an exam outline, or translate a paper into English.
The school says the purpose is to keep courses focused on the skills students need to conceptualize, outline, draft, revise, and edit their own work.
The policy states: "In short, thinking remains the sine qua non of good lawyering (and of a quality legal education)."
The rule also covers translation, with Berkeley Law saying the restriction gives students the opportunity to develop and exercise their own fluency with legal English.
Exams face full AI ban
The exam rule is more restrictive than the paper research rule. AI use is prohibited for any purpose in any exam situation.
The policy includes AI-generated exam outlines in its list of prohibited conduct where elements of the outline are then used on the exam.
The school also warns that citations to sources that do not exist will raise a presumption of prohibited AI use. That gives the policy a sharper enforcement edge around hallucinated legal authorities and fabricated references.
Instructors can set written exceptions
Berkeley Law gives instructors discretion to deviate from the default policy, provided the alternative rule is given in writing, students receive appropriate notice, and any authorized AI use is disclosed.
The policy says students who are unsure whether a proposed AI use is allowed must ask their instructor and receive written clarification before using the technology.
The Summer 2026 policy leaves room for AI fluency courses, but sets a restrictive default for credited work, exams, course materials, and legal writing tasks. Its narrow permitted use is limited to identifying sources for papers, with responsibility for accuracy remaining with the student.
