Mexico’s National Autonomous University of Mexico (UNAM) has mandated that 58,000 students retake their entrance examinations following a widespread failure of a remote proctoring system. Earlier this summer, approximately 160,000 candidates sat for the exam, which utilized a lockdown browser paired with AI-driven webcam monitoring software. The attempt to modernize the testing process, however, resulted in significant data inconsistencies that compromised the integrity of the admissions cycle.

According to Ars Technica, the primary indicator of the system's failure was a drastic surge in high-scoring candidates compared to historical data. While the percentage of students achieving a score of 100 or higher on the 120-question test typically hovered around 3.5 percent between 2021 and 2025, this year saw that figure jump to 16.3 percent. These statistical anomalies suggested that the automated surveillance technology failed to prevent cheating or experienced a major technical malfunction that invalidated the results.

University officials now face the logistical burden of re-testing a substantial portion of the applicant pool to ensure academic fairness. The incident highlights the growing risks associated with the rapid adoption of AI-supervised testing environments in large-scale academic institutions. UNAM has not yet released a detailed technical post-mortem regarding whether the software was compromised or if the AI generated false positive validations during the remote sessions.