Princeton University is ending its 133-year [1] tradition of unproctored exams and will now require monitoring for all in-person tests [1, 2].

The policy shift marks a significant departure from a long-standing culture of trust at the Ivy League institution. It reflects the growing difficulty universities face in maintaining academic integrity as generative artificial intelligence becomes more accessible on mobile devices.

University officials said the decision was driven by concerns that AI-enabled cheating on smartphones makes it difficult to detect academic dishonesty [2, 3]. The administration also said that modern social-media dynamics have reduced the likelihood of students reporting peers who cheat [2, 3].

For more than a century, Princeton students took exams without a supervisor present, relying instead on a strict honor code. This system was designed to foster a community of integrity, a cornerstone of the university's academic environment.

Under the new rules, all in-person examinations must be proctored to ensure students are not using unauthorized digital tools. The university announced that the new proctored-exam policy takes effect on July 1, 2026 [1].

This change follows a broader trend in higher education where institutions are redesigning assessments to mitigate the impact of large language models. By requiring physical supervision, Princeton aims to close the gap created by discreet, handheld AI tools that can generate complex answers in seconds.

The move signals that even the most prestigious honor codes may be insufficient against the speed and invisibility of modern technology. The university will now prioritize surveillance over the traditional trust-based model to protect the value of its degrees [2, 3].

“Princeton University is ending its 133-year tradition of unproctored exams.”

The collapse of Princeton's unproctored exam system suggests that generative AI has reached a tipping point where traditional honor codes are no longer viewed as viable deterrents. As AI tools become smaller and more integrated into personal devices, the 'trust-based' model of education is being replaced by a 'verification-based' model, potentially altering the psychological relationship between students and institutions.