Brown University ECON 1170 sees AI cheating by at least 50 students on March 2026 midterm

Serrano, Harrison S. Kravis Professor of Economics, submitted evidence of coordinated answer patterns to Brown's Academic Code Committee after the president and dean initially offered no response. The exam required students to evaluate statements under modified assumptions not covered in lectures, producing identical solution structures inconsistent with individual derivation. Committee acknowledged the incident as a wake-up call but issued no sanctions or procedural changes.

Take-home exams remain standard across Ivy League economics departments, with Brown recording zero prior large-scale integrity cases in the prior decade per internal records. Serrano's detection relied on solution clustering rather than external tools, mirroring patterns reported in 2024 Stanford CS courses where 18% of submissions matched GPT-4 output signatures. No public dataset tracks Ivy League AI violations, leaving scale unknown.

Current Brown policy leaves detection and adjudication to individual faculty without centralized verification protocols or updated honor code language on generative models. This case exceeds documented incidents at peer institutions by an order of magnitude and exposes absence of audit trails for remote assessments.

Operational response will require either in-person proctoring or randomized oral defenses within one semester if similar clusters appear in final exams, per committee minutes.