Roberto Serrano, a professor of advanced mathematical economics at Brown University, noticed a spike in midterm exam performance last semester that raised red flags. Suspecting students relied on artificial intelligence tools to generate answers, Serrano took action to detect and combat the problem.
The case reflects a growing challenge across higher education. Students increasingly use AI writing assistants and problem-solving tools like ChatGPT to complete assignments and exams, making it difficult for instructors to distinguish between original work and AI-generated content. This shift has forced professors to rethink assessment methods and detection strategies.
Serrano's response represents one approach professors use when they suspect widespread AI cheating. Rather than simply accepting the suspicious results, he implemented measures to identify which submissions violated academic integrity policies. His experience at an Ivy League institution highlights how even elite universities struggle with AI-enabled academic dishonesty.
The stakes are high. Accurate grades depend on genuine student work and understanding. When AI tools produce solutions, grades no longer reflect actual learning. Additionally, unfair advantages emerge for students who use AI versus those who don't, creating equity problems within classrooms.
Brown University, like most institutions, maintains academic integrity policies that prohibit submitting AI-generated work as one's own. However, enforcement remains uneven. Some professors use plagiarism detection software adapted for AI content, while others employ in-class assessments or oral exams where students must explain their reasoning in real time.
Serrano's experience underscores a broader institutional challenge. Universities face pressure to update grading practices, honor codes, and detection systems faster than AI technology evolves. Training faculty on AI literacy and detection methods remains incomplete at many schools. Students, meanwhile, receive inconsistent guidance about what constitutes acceptable AI use versus cheating.
The outcome of Serrano's investigation at Brown could influence how other institutions address similar problems. His willingness to con