Updated
Updated · Ars Technica · Jul 8
Brown Professor Flags 40 Perfect Midterms as AI Cheating Suspicions Sweep 86-Student Economics Class
Updated
Updated · Ars Technica · Jul 8

Brown Professor Flags 40 Perfect Midterms as AI Cheating Suspicions Sweep 86-Student Economics Class

3 articles · Updated · Ars Technica · Jul 8

Summary

  • Forty students scored 100 and the class average hit 96 on a March 5 take-home midterm in Brown economist Roberto Serrano’s ECON 1170, far above the course’s usual 65-80 range.
  • Serrano said the exam was harder than past midterms, making the results especially suspicious and fueling his belief that generative AI was widely used to cheat.
  • Enrollment in the class jumped to 86 from a typical maximum of 30 after Serrano switched to take-home midterm and final exams for spring 2026.
  • That change followed Brown’s December 2025 campus shooting, which killed two people and left Serrano shaken after one victim had recently introduced herself to him.
  • The case adds a concrete example to broader concern over AI cheating at elite schools; a recent Princeton survey found 29.9% of students admitted using AI to cheat on at least one exam or assignment.

Insights

With AI acing exams, can universities still guarantee the value of a degree?
When AI detectors are biased, how can schools fairly punish student cheating?

AI Cheating at Brown: 22 of 27 No-Shows Scored 100—A Wake-Up Call for Universities

Overview

The Brown University AI cheating scandal began in Professor Roberto Serrano’s ECON 1170 course, where suspicions arose after a take-home midterm. Serrano warned students that the midterm would only count if the final exam’s grade distribution matched the midterm’s, expecting a clear difference if cheating had occurred. This conditional approach acted as an early detection method. When the in-person final revealed a dramatic drop in scores and attendance, it confirmed widespread AI-assisted cheating. The incident exposed major challenges for academic integrity, prompting Brown and other universities to rethink assessment methods and policies in the age of artificial intelligence.

...