For AI practitioners and educators, the Brown incident highlights how accessible generative models can undermine take-home assessments and push institutions toward supervised or redesigned evaluation strategies. According to reporting by Inside Higher Ed, Fortune, Ars Technica and other outlets, Brown economics professor Roberto Serrano gave a take-home midterm to his ECON 1170 class after a December campus shooting; the midterm produced an average score of 96 percent, with 40 of 86 students earning perfect or nearly perfect marks (Inside Higher Ed, Fortune). Inside Higher Ed and Ars Technica reported that Serrano and his graders compared student answers with outputs from ChatGPT, finding similar wording and reasoning. After Serrano announced an in-person final, several students dropped the course, and reporting shows the final average fell to roughly 48 percent with 19 students failing (Ars Technica, Inside Higher Ed). Multiple outlets characterized the case as a large-scale AI-assisted cheating incident in the Ivy League (Fortune, Futurism).
Mass AI Cheating at Brown Raises Shocker Query: Is an Ivy League Degree Worth $90,000 a Year?