TO SUMMARIZE:
UNAM, Mexico’s most prestigious university, just canceled thousands of exam scores over suspected AI cheating, a scandal that has spread all the way to the country’s president.
UNAM put its entrance exam online for the first time this year, hoping to make it easier for applicants outside Mexico City. More than 158,000 people took it in May and June. The scores that came back didn’t add up.
Here’s what happened:
Between 2021 and 2025, about 3.5% of applicants historically scored 100 or more correct answers.
This year, that rate spiked high enough to trigger a formal review.
UNAM annulled roughly 3,000 of the ~150,000 tests over suspiciously perfect scores.
The university suspended new-student enrollment (the semester starts August 10) while it investigates.
Officials suspect a mix of AI tools like ChatGPT and old-fashioned tricks, like leaked questions.
The exam software, built by Mexican company Territorium Life, was supposed to catch exactly this: it blocked test-takers from opening extra browser tabs and used AI to flag suspicious behavior. Students say the system leaned too hard on AI and not enough on humans watching. Territorium Life maintains its software worked as intended, and argues stopping cheating also requires the university to enforce honesty on its end.
High-stakes testing is moving online everywhere, and AI is increasingly the referee. When that referee can’t reliably tell a cheater from a hard studier, everyone pays for it, including the students who did nothing wrong.
Nobody built a system that can tell the difference between a cheater and someone who just studied hard, and that gap is the actual scandal here. Turns out “trust but verify” gets messy fast when the verifier is an algorithm nobody fully understands.
When I sat for the Security+ test, it was during a time when you were required to wear a stupid "COVID" mask. It limited my oxygen and steamed my glasses. I was beyond pissed off, but endured it. I needed the cert for a new contract.