An AI-supervised remote exam went so badly that 58,000 students must retake it
Nearly 160,000 applicants sat Mexico's UNAM university entrance exam remotely for the first time this year, using a lockdown browser and AI-powered webcam proctoring. The results were so anomalous compared with previous years that a specially convened expert commission has recommended around 58,000 people retake the test in person, after the remote format appears to have enabled widespread cheating despite the AI supervision meant to prevent it.
The proportion of students scoring 100 or more out of 120 jumped from an average of 3.5 per cent between 2021 and 2025 to 16.3 per cent this year, while those scoring 110 or more rose from 0.9 per cent to 5.5 per cent. The new "control exam" will apply to everyone who secured a place based on this year's results or on past minimum qualifying scores since 2021, with details expected imminently as classes are due to start on 10 August. UNAM's rector has apologised to honest candidates forced to retest, while reports suggest students may have exploited the remote setup by hiding phones, using off-screen devices to access AI chatbots, or even having someone else sit the exam for them.
- UNAM found 58,000 students must retake its entrance exam.
- Remote AI-proctored exam saw scores spike suspiciously versus past years.
- Cheating tips reportedly circulated online before the test.