A Brown University professor changed an ECON 1170 midterm format to take-home after a mass shooting, citing student trauma and mental health concerns. The exam, however, produced an unusually high score distribution and was later linked to the largest known AI-assisted cheating scandal in the Ivy League. Professor Roberto Serrano reported that the exam design—harder than usual with an atypical score spread—was meant to reduce campus stress but became a warning signal for misconduct. After graders ran the questions through ChatGPT, they found AI-generated reasoning patterns appearing across dozens of student submissions. The incident highlights the tension between student well-being and assessment security when exams move online or outside proctored settings. It also underscores how quickly AI tools can be used to mass-produce plausible answers even on complex, mathematical prompts. Brown’s experience is likely to reverberate in faculty governance and course-design discussions across higher education as instructors weigh when take-home formats are appropriate and what safeguards are needed.