Attendance testing reveals AI fraud at elite college

A recent incident at a prestigious university in the United States has exposed widespread academic misconduct involving artificial intelligence tools. The scandal came to light after a professor noticed a dramatic drop in grades during an in-person exam compared to earlier assessments conducted remotely. This case highlights growing concerns about how AI is reshaping traditional educational practices and challenging long-standing academic integrity standards. The controversy began in a course on Mathematical Economics taught by Professor Roberto Serrano. For this semester, he decided to conduct both the mid-term and final exams as "take-home, closed-book" assessments—meaning students could complete them at home but without using external resources. This format was chosen because it allows students more time to work through problems, according to Serrano. However, the results of the mid-term exam were unusually high, with an average score of 96 out of 100 points. Remarkably, 40 students achieved perfect scores, which raised red flags among the grading team. Professor Serrano became suspicious when he noticed several unusual patterns in the responses. Some answers contained exact phrases that matched outputs generated by chatbot systems such as ChatGPT. These observations prompted him to investigate further. When the final exam required in-person attendance, the average score plummeted to 48 points, and 27 students who had taken the mid-term did not even show up for the final. This stark contrast between the two sets of results confirmed his suspicions of academic dishonesty. The issue escalated when Serrano brought his concerns to the administration of Brown University, where he has been teaching for over three decades. According to reports, the initial response from the university leadership was dismissive. It wasn't until Serrano escalated the matter to the Academic Integrity Committee that the institution took notice of the situation. He emphasized that this was the first time he had encountered such a significant breach of academic ethics in his career. In response to the scandal, Serrano has decided to make substantial changes to his assessment methods for future semesters. He will no longer include weekly exercises in the final grade, as these can be easily completed using AI tools. Additionally, he plans to eliminate take-home exams entirely, despite their pedagogical benefits, due to the increased risk of cheating facilitated by AI technologies. This incident is part of a broader trend affecting universities worldwide. At Princeton University, for instance, there have been discussions about reinstating in-person supervision for examinations—a practice that had not been used since 1893. Previously, students followed an honor code that required them to self-regulate and report any instances of cheating. With the rise of AI, however, the ability to detect and prevent academic fraud has become increasingly complex. The impact of AI on education extends beyond just examinations. Universities are now grappling with the need to adapt their policies and teaching strategies to address the challenges posed by rapidly evolving technology. While some institutions are exploring new ways to integrate AI into their curricula, others are focusing on reinforcing existing measures to ensure academic honesty. As the debate continues, educators and administrators are left with difficult decisions regarding the balance between technological advancement and maintaining ethical standards in academia. The incident at Brown University serves as a cautionary tale about the potential pitfalls of relying too heavily on remote learning formats and the importance of continuously evaluating and updating educational practices in the face of emerging technologies.