At least 50 students at Brown University used ChatGPT to cheat on a take-home economics midterm, marking what experts say is the largest known AI-facilitated cheating scandal in Ivy League history, according to a detailed report published by El País on June 28, 2026.
The case was uncovered by Roberto Serrano, a Kravis University Professor of Economics who has taught at Brown for 34 years. After administering a closed-book, take-home exam for his advanced mathematical economics course ECON 1170 in March 2026, Serrano discovered that students had submitted answers containing passages that matched ChatGPT outputs almost word for word. The findings have sent shockwaves through higher education and underscore a growing crisis of academic integrity that extends far beyond a single campus. For ongoing reporting on how AI is reshaping education and society, this case offers a rare inside look at the scale of the problem.
How the Cheating Was Discovered
The exam itself was designed to be unusually rigorous. Serrano had modified some of the model assumptions students had studied in class and asked them to derive new results from scratch. Because it was a take-home exam with virtually unlimited time, students could take as long as they needed. But the design also meant there was no proctoring.
The red flags came quickly. The average score on the midterm was 96 out of 100 — a remarkably high figure for a notoriously difficult course that typically enrolls fewer than 30 students. This semester, however, 86 students had signed up, nearly triple the usual number. Serrano suspects the take-home format itself attracted students looking for an easier grading path.
The graders flagged unusual passages in multiple exams. Some answers contained identical phrasing that matched the output of ChatGPT when the same questions were fed into the model. After cross-referencing, Serrano concluded that at least 50 students — more than half the class — had used AI to complete the exam.
A Professor Left Without Support
What happened next frustrated Serrano even more than the cheating itself. When he reported the scale of the fraud to senior university officials, the response was muted. Brown's president offered what Serrano described as "absolute silence." The dean did not respond until Serrano escalated the matter to the Academic Code Committee, at which point he received a brief note calling the incident "a wake-up call."
"That cannot be the university's position before an incident of this magnitude," Serrano told El País. "The faculty cannot be left on its own in a battle that is decisive if we want to preserve the future of higher education."
Serrano is calling for universities to publicly acknowledge the severity of AI-assisted cheating rather than treating each case as an isolated disciplinary matter. He argues that without a broader, institution-wide conversation, the problem will only worsen as AI tools become more powerful and accessible.
The Blind Economist Who Caught the Cheaters
Serrano's personal story adds a striking dimension to the case. A Madrid-born economist who earned his PhD at Harvard and received the King of Spain Prize for Economics in 2024, Serrano lost his sight entirely in recent years due to a progressive retinal dystrophy. He relies on a teaching assistant for whiteboard work and slide management but handles all other tasks — from preparing exercises to writing research papers — himself.
He credits technology, including AI tools, with making many of his daily tasks more manageable. But he draws a sharp line between using AI as an aid and using it to replace independent thought. "We economists understand reality as a set of people responding to optimization problems with restrictions," he said. "I view my disease simply as one more restriction that I have to deal with, and I optimize based on that."
A Crisis Extending Beyond Brown
The Brown scandal is not an isolated incident. Universities across the United States and beyond are grappling with a surge in AI-assisted cheating that threatens to undermine the credibility of take-home assessments, a staple of Ivy League pedagogy. A recent study found that the temptation to use generative AI to cut corners is shaking up even the most elite institutions, where honor codes have long been considered sufficient deterrents.
The problem is compounded by the speed at which AI models are improving. When Serrano's students took their exam in March 2026, ChatGPT was already capable of producing graduate-level mathematical derivations that closely resembled correct human answers. Detecting such cheating requires faculty to actively cross-reference student submissions against AI outputs — a labor-intensive process that many professors are not trained or equipped to perform.
Some universities have responded by returning to in-person, proctored exams. Others are experimenting with AI-resistant assessment designs, such as oral examinations or project-based evaluations. But these solutions come with trade-offs: oral exams are impractical in large courses, and project-based assessments can be even harder to police for AI use.
What Comes Next
Serrano's case has sparked a broader conversation about whether the traditional take-home exam can survive the AI era. For now, Brown has not publicly commented on the specific incident, and it remains unclear what disciplinary actions, if any, will be taken against the students involved.
What is clear is that the status quo is unsustainable. As one of the nation's most respected economists, Serrano's warning carries weight: if universities fail to address AI-assisted cheating head-on, they risk presiding over the erosion of the very academic standards that define their value.
---
Stay Ahead of the AI ConversationThe intersection of artificial intelligence and society is evolving daily. From campus cheating scandals to breakthroughs in model safety, staying informed has never been more critical. Bookmark AI Buzz Wire for comprehensive coverage of the stories shaping our AI-driven future.
Read more AI news →
