An unprecedented technological milestone has resulted in an institutional crisis for higher education in Mexico. For the 2026-2027 bachelor's degree selection process, the National Autonomous University of Mexico (UNAM) implemented for the first time a completely remote admission exam for more than 158,000 applicants. To guarantee the integrity of the process, the institution implemented an intelligent monitoring system based on automated proctoring software, artificial intelligence (AI), and biometric recognition. However, the effectiveness of the supervision system was severely questioned after a wave of complaints and viral confessions on social media.
Throughout the process, platforms such as TikTok were filled with methods devised by the applicants themselves to circumvent the exam's detection algorithms, ranging from the use of blind spots to hide second devices to the use of tools such as ChatGPT. After publishing the results where more than 3,100 exams were canceled due to inconsistencies and alleged fraud (equivalent to 2% of the total), the controversy escalated, leading UNAM to make the unprecedented decision to temporarily suspend the delivery of documents and registrations for the academic year. The Maximum House of Studies also filed complaints with the competent authorities for fraudulent networks that sold AI assistance during the test.
The decision to halt the process will remain until a technical commission of experts appointed by the rector thoroughly reviews the development of the competition and determines the failures or inconsistencies of the model. The case has opened an intense international debate on whether current Artificial Intelligence supervision systems are truly capable of monitoring massive high-impact evaluations, or whether the rapid evolution of generative tools has surpassed the defensive capacity of traditional proctoring software.
From the perspective of the next+ team, the controversy surrounding the UNAM exam reflects a decisive turning point in the adoption of Artificial Intelligence for the automation of critical processes. Entrusting the control and surveillance of human behavior exclusively to algorithmic models generates a false sense of security when evasion technology evolves at a much faster pace. For organizations, academic institutions, and companies, the lesson is clear: AI should not operate as a sole and isolated judge, but as a support tool within a hybrid supervision framework. Digital governance, transparency in audit criteria, and constant updating of security protocols are essential to prevent automation from compromising legal certainty and institutional reputation.