Professor Criticizes University Response to AI Cheating

Brown University economics professor Roberto Serrano criticized the university’s handling of an AI cheating incident involving his students. Earlier this year, he observed suspicious behavior during his course exams.

Sudden Shift to Online Exams and Its Impact

Serrano shared in The Free Press that he decided to move his midterm and final exams online after a campus shooting in December. This change increased his course enrollment to 86, far surpassing the typical maximum of 30 students.

The average midterm grade rose to 96 as compared to the previous average of 65 to 80. Forty students earned perfect scores, and some answers resembled those generated by AI tools like ChatGPT.

Revised Exam Format and Results

Due to the suspicious scores, Serrano reverted to in-person exams for the final. He informed students that midterm scores might be voided if they didn’t align with final scores. Ultimately, only 59 students completed the final exam, with an average grade of 48.6. The scores ranged from a minimum of zero to a high of 95.

Serrano pointed out that 22 of the 27 students who dropped out after the midterm had secured perfect scores earlier.

University’s Response

Serrano presented his findings to the university’s Committee on the Academic Code and The Brown Daily Herald. Both sources did not respond until his story gained widespread attention.

Brown University stated they were engaged with Serrano since May 2026. They outlined how the claims could be formally reviewed. After receiving the needed details on July 8, they began processing the claims according to established procedures.

Serrano’s Perspective on AI and Cheating

Serrano alleges the university’s lack of action prompted him to make his case public. He believes addressing his publicized story was crucial for the university to investigate these claims seriously.

He argued that AI has made cheating easier, dismissing claims that students use AI to manage stress. He maintained that there’s no excuse for deliberate cheating, which was evident when students had ample time to complete exams.

Serrano concluded that effective deterrence needs to be in place to curb AI-assisted cheating. He emphasized that without accountability, cheating becomes a likely choice.