Princeton University’s 133-year tradition of unproctored exams ended due to AI, with nearly 30% of students admitting to cheating.

On May 11, 2026, Princeton University’s Academic Senate passed a resolution to enforce proctoring during all on-campus exams with only one vote against, which took effect on July 1, 2026. This ended a 133-year-old tradition of the honor system that explicitly prohibited proctoring since 1893.
(Background: Columbia students create AI online interview tool, scam internships at Amazon and other companies! After suspension, start a business earning $170k per month.)
(Additional context: The emergence of ChatGPT learning models: twilight of tutoring, or dawn of a golden era of education?)

In 1893, Princeton students exchanged a petition for a change: no longer would there be someone watching over them during exams. This mutual trust between teachers and students was then written into the school rules, continuing for 133 years.

But on May 11, 2026, the Academic Senate ended this agreement with only one vote against. The reason for overturning it was not a sudden increase in cheating rates, but a more fundamental issue: after the proliferation of AI tools, the traditional system relying on “peer supervision” had become ineffective.

133 Years of Autonomous Experiment, and a Termination Order

Before 1893, Princeton, like most American universities, had teachers present during exams to supervise. After the students’ petition, the school based its rules on the principle of “mutual trust” between teachers and students, explicitly banning proctoring, and established a student-led Honor Committee to handle academic integrity cases. Since then, the “Teacher Rules and Procedures” and “Rights, Rules, and Responsibilities” both explicitly included this prohibition.

This 133-year autonomous experiment came to an end this May. The proposal underwent three rounds of review: the Exam and Enrollment Committee, the Faculty Policy Advisory Committee, and the Faculty Assembly, all approved it, passing overwhelmingly with only one vote against.

The new system takes effect from July 1, 2026: during in-person exams, teachers or proctors must be present “as witnesses to what happens on site,” but not intervene proactively; if suspicious behavior is detected, it is reported to the student-led Honor Committee. The Honor Code itself remains unchanged; only the two enforcement regulations behind it are modified.

AI Does Not Change the Will to Cheat, but the “Observability” of Cheating

Why now?

The proposal states that the accessibility of AI tools on small personal devices “has already changed the appearance of misconduct during exams.” In plain terms: in the past, cheating was visible because classmates could see what you were doing; now, with a phone on the desk, AI can generate answers in real-time, and those around cannot tell whether you’re checking the time or copying answers.

A 2025 survey of recent graduates (more than 500 respondents) revealed an unsettling figure: 29.9% admitted to cheating during their studies; 44.6% knew classmates cheated but chose to remain silent; only 0.4% had ever reported a peer.

This means that out of every 100 informed individuals, fewer than 1 chooses to activate the reporting system.

The proposal also explains this phenomenon: students are reluctant to report peers partly because they fear “online doxxing or bullying within peer circles.” The emergence of AI has made this already fragile reporting mechanism even harder to operate.