Columbia Law School sets its default rule: AI allowed to learn, forbidden to reason in the student's place

The law school's policy, which came into force on August 1, 2026, distinguishes between support for study and substitution for legal reasoning, and requires the declaration of all use.

Applicable since August 1, 2026 for the academic year 2026-2027, Columbia Law School's default policy is based on a simple principle: the student assumes full intellectual responsibility for any work performed, and AI cannot substitute for its legal reasoning.

Unless otherwise indicated by the teacher, the tool is allowed to understand concepts, conduct research and identify sources, test arguments, clarify writing or correct grammar, and to revise and prepare exams. However, the original legal analysis of a duty, any use in examination without prior written permission, the delivery of a generated text presented as personnel, the use during courses without the teacher's consent and the seizure of confidential information from clients are excluded.

Declare, verify, defend

Three obligations apply to students: to declare any use of AI in the work performed, to independently verify the sources produced before quoting them, and to be able to explain and defend each argument. Legal clinics and internships are subject to an enhanced regime, as confidential client information cannot be paid to an AI system without express authorization. Detail revealing actual uses: translating a source into a foreign language into English is allowed, but translating one's own work into another language is not.