Bengaluru: Brown University economics professor Roberto Serrano has taught at the university for 34 years. This year he encountered a novel problem. He’s accusing dozens of students in his 86-person Welfare Economics and Social Choice Theory class of using AI to cheat.
Serrano had allowed the midterm to be taken at home, departing from his usual practice of in-person, closed-book exams. This came after several students said they were anxious about sitting exams in a classroom following a December mass shooting at Brown that killed two students and injured nine, Fortune reported.
The class averaged 96 per cent in the midterm, against a historical range of 65 to 80 per cent. Forty of the 86 students scored a perfect 100, according to Futurism‘s report on the story. Serrano figured something was off.
To test his suspicion, he changed the format of the final exam from take-home to in-person and closed-book. He told the class that if the grade distribution on this in-person final closely matched the distribution on the take-home midterm, he would let the midterm score stand as valid. If it diverged sharply, he would treat that as confirmation of cheating.
“I am going to give the class a chance to prove me wrong,” he wrote in a message to the students.
The outcome confirmed his suspicion. Eighteen students dropped the course before the final. Among those who sat it, the average score fell to 48.6 per cent, with three students scoring zero. Nineteen students ultimately failed the course. Of the 27 students who skipped the in-person final altogether, 22 had scored a perfect 100 on the take-home midterm, per Futurism.
“If workers are just going to press a button to ask an AI agent to do the work for them, that’s inscribing a world in which humanity has chosen to become idiots. We stop thinking,” he said to Fortune.
Brown’s response
According to Inside Higher Ed, Serrano submitted evidence of what he described as cheating to Brown’s Standing Committee on the Academic Code in May and received no response. He then went public with the story in June.
Brown then asked him to file individual, evidence-backed complaints against each suspected student—a process Serrano considers impractical given the limitations of AI-detection tools.
Speaking to Fortune, Serrano said: “I’m very frustrated. I believe the arrival of AI has been like a tsunami for all of us. It’s caught everybody unprepared.”
The case is not isolated. A Cornell-led study of more than 95,000 students across 20 US public research universities found that roughly a third used generative AI regularly for coursework, and 9 per cent admitted using it to cheat outright. Princeton University has ended a 133-year tradition of unproctored, honour-code exams in favour of mandatory in-person invigilation, according to Fortune.
Serrano also pointed to a recent New York Times essay that described a pervasive culture of AI cheating in Stanford—elite university students are there for a line in their CV, not for learning. “What they miss in that very naive analysis is that the Brown label is Brown for a while. But if Brown continues to produce mediocre students who refuse to learn, sooner or later the market is going to find out that the Brown label is not what it used to be,” he told Fortune.
He has already changed his syllabus for the coming academic year: Take-home exams will no longer be offered, and ungraded homework assignments will carry no weight toward final grades.
“It’s too easy for students to succumb to temptation,” he said.