Brown professor exposes 50+ students in Ivy League's biggest AI cheating scandal

Roberto Serrano, a tenured economics professor at Brown University, says he has conclusive evidence that at least 50 students used AI to cheat on the March midterm of his advanced mathematical economics course. The signs were hard to miss: enrollment had tripled to 86 students under a new take-home, closed-book exam format, the midterm average hit 96 out of 100, and 40 students scored a perfect 100 — with some answers containing passages that matched ChatGPT output. Serrano calls it the largest known academic integrity breach at Brown and across the Ivy League.

Rather than void the midterm, Serrano made the final worth 50% of the grade and held it in person, warning that grades would be recalculated if the distributions diverged sharply. They did. The in-person average collapsed to 48, and of the 27 students who skipped the final entirely, 22 had previously scored a perfect 100. Going forward, he is scrapping take-home exams and dropping graded weekly problem sets, since both are trivially gameable with AI.

The more pointed grievance is institutional. Serrano says Brown’s leadership met his report with near silence — no response from the president, and only a belated acknowledgment from the dean that the episode was ‘a wake-up call.’ His argument is that faculty cannot police generative AI alone, and that universities risk hollowing out the value of a degree unless they openly confront how widespread the problem has become.