Princeton Ends 133-Year Honor Code Precedent: AI Cheating Forces First Exam Proctoring Since 1893
Princeton University faculty voted on May 11 to end unproctored exams, a practice students had held since 1893 when they successfully petitioned to remove academic supervisors from examination rooms. The vote was nearly unanimous — one faculty member opposed. The policy takes effect July 1.
The move is the most significant change to Princeton’s Honor Code since its establishment 133 years ago.
What Changed
Under the Honor Code as it existed until this vote, students pledged to refrain from academic dishonesty and to report any violations they witnessed. Professors left the room. The obligation to report rested entirely on students observing one another.
Generative AI broke both sides of that system simultaneously. Cheating became easier — a phone or small device is sufficient — while the observable surface for other students to notice and report shrunk. Catching someone using ChatGPT on their phone is materially harder than catching someone passing notes.
The policy proposal came from Dean of the College Michael Gordin, citing requests from students and faculty over “the past few years and with increasing frequency over the last six months” that reflected “the perception that cheating on in-class exams has become widespread.”
How Proctoring Works
Proctors will not interfere with students. Their role is observation and documentation. If a proctor suspects an Honor Code violation, they submit a report to the student-run Honor Committee. The text of the Honor Code itself — governed by students, not administration — is not changing. Students retain their dual obligation to act with integrity and to report violations.
Specific implementation details (proctor-to-student ratios, monitoring guidelines) will be finalised before July 1.
The Broader Signal
Princeton is a useful case study because its honor system was unusually pure — no cameras, no supervision, built on 133 years of institutional trust. That it required explicit dismantling says something about where generative AI has taken academic integrity norms across higher education.
The pattern appearing at universities: AI makes violations both easier (faster, harder to detect with the naked eye) and harder to address within existing social-accountability frameworks. Peer-reporting cultures assume violations are visible to peers. AI usage in an exam room is not.
Princeton is not the first institution to respond. But it is among the most symbolically significant, given the length and intentionality of its self-governance tradition.