AI Cheating in College Exams: Professor's In-Person Test Exposes Half Grades 2026

AI cheating in college exams has become a pressing concern, and a Brown University economics professor's experiment offers striking evidence of its impact. When Roberto Serrano suspected his students were using AI to ace take-home tests, he switched to in-person exams. The result? Grades dropped by half.

Why a Professor Suspected AI Cheating

In December 2025, a shooting on Brown's campus killed two people, including someone who had just met with Serrano. The tragedy prompted Serrano to change his evaluation method for his Spring 2026 course, ECON 1170. He decided to allow take-home exams for both midterms and the final.

The course, known for its rigor, typically attracted few students—sometimes just eight. But this spring, enrollment surged to 86, likely because students learned about the take-home format. Serrano grew suspicious when the quality of work seemed unusually high.

The In-Person Retest That Changed Everything

To verify his suspicions, Serrano administered an in-person exam. The results were dramatic: scores plummeted by about 50% compared to the take-home exams. This stark contrast suggested that many students had used AI tools to complete their take-home work.

The experiment provided concrete data on the prevalence of AI cheating. It also raised questions about how universities should adapt their assessment methods in an era of powerful generative AI. The background to this story is set out in Cambridge Professor Arday Resigns Amid Plagiarism Accusations. A related thread runs through Cambridge Professor Arday Faces Plagiarism Claims, Academic. Earlier coverage in Jason Arday Death: Cambridge Professor Found Dead shows how the argument developed.

Key Takeaways for Educators and Students

• Take-home exams may be highly vulnerable to AI cheating, inflating grades significantly.
• In-person assessments can reveal the true understanding of students.
• Course design and enrollment incentives can inadvertently encourage academic dishonesty.
• Universities need clear policies and innovative assessment strategies to maintain integrity.
• Students should focus on genuine learning rather than short-term shortcuts.

Comparing Exam Formats: Take-Home vs. In-Person

What This Means for Academic Integrity

The Brown University case highlights the challenges institutions face as AI tools become more accessible. While take-home exams offer flexibility, they also open the door to misuse. In-person exams, though less convenient, provide a more accurate measure of student knowledge. A closer look at the same subject appears in Israeli Professor Wins $250,000 Settlement After University. Anyone following the wider picture can start with Academic Suspended Over Plagiarism Allegations Against Professor. The same theme surfaces in Cambridge University Investigates Professor Accused of Plagiarism, which takes a different angle.

Educators must weigh these trade-offs carefully. Some may opt for hybrid models, such as in-class writing assignments or oral defenses, to mitigate AI cheating while preserving some flexibility. Context on how this played out before sits in AI Cheating and Exam Scandals: Global Student Protests Rise. A companion piece, Denmark Fights AI Cheating with Oral Essay Defenses, approaches the same subject from another angle.

FAQ