A Brown University economics professor reported what he says was a large AI-assisted cheating scandal after switching an exam format to a take-home assessment aimed at reducing stress for students after the December 13 campus shooting. Professor Roberto Serrano redesigned the March midterm to be take-home and closed-book, citing ongoing trauma impacts on campus participation. According to Serrano, the exam results were unusually high: 40 of 86 students scored 100, and graders reported that many answers contained patterns consistent with ChatGPT output. He said the course team used the same prompts to test for overlap and found the same convoluted reasoning appearing across dozens of student exams. The incident highlights how student mental health goals around assessment access can collide with rapidly evolving academic-integrity risks, particularly when generative AI can generate plausible responses. It is likely to intensify institutional scrutiny of take-home, open-ended assessments across the sector.