Brown University professor Roberto Serrano altered his exam format after suspecting AI-powered cheating during a take-home midterm. He shifted the final exam to an in-person format, citing concerns over academic integrity. Only two students scored within 10% of their midterm grades, raising questions about the role of AI in student performance.

Serrano's decision to make the final exam in-person followed a noticeable drop in student scores. The professor observed that most students performed significantly worse on the final compared to their midterm results. This discrepancy led him to suspect that AI tools were being used to complete the take-home exam, undermining the assessment's fairness.

According to the data, only two students scored within 10% of their midterm grades. This stark contrast with the rest of the class suggests a possible reliance on AI-generated content during the take-home exam. The professor noted that the majority of students performed between 65% and 80% of their midterm scores, which he attributed to the use of AI tools.

The shift to in-person exams highlights broader concerns about academic integrity in the age of AI. Institutions are grappling with how to prevent AI misuse in assessments while maintaining fairness. The move may influence future exam policies, prompting universities to reconsider the use of take-home exams and explore stricter monitoring methods.

The situation remains under review, with the university’s Standing Committee examining the implications of AI in academic settings. The outcome of this review could shape future policies on exam formats and AI usage. As the debate over AI's role in education continues, institutions will need to balance innovation with academic standards.