Brown University professor Roberto Serrano alleges that at least 50 students used ChatGPT to cheat on an online midterm exam, leading to a dramatic shift in grading outcomes when the final exam returned to an in-person format. Serrano claims the university only began investigating the evidence after his account went viral in a series of reports.
The Gap Between AI-Assisted and In-Person Performance
According to Serrano, who wrote an op-ed for The Free Press, he moved his midterm and final exams to a take-home format following a deadly campus shooting in December. This change saw course enrollment spike to 86 students, beating the course’s previous record of 30.
The results of the online midterm were unprecedented. Serrano reported an average grade of 96, compared to previous years’ averages of 65 to 80. He noted that 40 students achieved perfect scores, with responses that mirrored ChatGPT output.
Did you know? The “AI effect” on enrollment was stark: Serrano’s class grew from a previous record of 30 students to 86 once the exams moved online.
When Serrano shifted the final exam back to an in-person format, the numbers plummeted. Only 59 students took the final. The average grade dropped to 48.6, with the highest score reaching 95 and the lowest hitting zero. Of the 27 students who dropped the course, 22 had previously earned a perfect score on the AI-suspected midterm.
Brown University’s Response to Cheating Allegations
Serrano alleges that the university ignored his findings until they became public. He shared his evidence with El Pais on June 28 and later with Inside Higher Ed. He claims the school did not acknowledge his evidence until the story went viral last month.
Brown University disputed this timeline in a statement to Fox News Digital. The university maintained it has been “consistently responsive,” stating that academic leaders were in touch with Serrano in May 2026 to discuss formal adjudication. The school noted that on July 8, the professor provided the necessary details to the Standing Committee on the Academic Code to move forward with procedures.
Comparing the Data: Online vs. In-Person
| Metric | Online Midterm (AI Era) | In-Person Final |
|---|---|---|
| Average Grade | 96 | 48.6 |
| Perfect Scores | 40 students | – |
| Enrollment/Participation | 86 students | 59 students |
The Future of Academic Integrity in the AI Era
Serrano rejects the idea that students use AI to cope with academic pressure. In his Free Press op-ed, he argued that because the midterm provided virtually unlimited time, the use of AI was a “deliberate act of cheating” rather than a last resort. He stated that AI has dramatically reduced the difficulty of cheating, making it easier for students to succumb to temptation.
Serrano argues that academia must “expose and punish cheating” or risk encouraging a decline in moral values among students.
Frequently Asked Questions
How did Professor Serrano detect AI cheating?
Serrano identified answers that closely resembled responses generated by ChatGPT and noted a statistical anomaly where the average grade jumped from a historical 65-80 to 96.
What happened to the students who cheated?
Serrano declared the midterm scores void if they did not align with final exam performance. Brown University’s Standing Committee on the Academic Code is currently processing the details provided by Serrano for formal adjudication.
Is Brown University ignoring the issue?
Serrano claims the school only acted after public pressure. Brown University denies this, stating they were in contact with the professor as early as May 2026.
What do you think? Should universities return to strictly in-person testing to combat AI, or is there a better way to integrate these tools? Share your thoughts in the comments below or subscribe to our newsletter for more updates on the intersection of technology and education.
Related reading
