A student can encounter an AI tool long before encountering a clear explanation of what a university permits. A UNESCO IESALC report launched September 9, 2026 gives that mismatch a regional scale. Its evidence comes from 200 higher education institutions in 19 Latin American and Caribbean countries. Among those institutions, 87 percent reported using AI somewhere in their activities, while 26 percent had a formal AI strategy.
These are findings about the institutions studied, not a census of every university in the region or a measure of individual student behavior. They also do not establish that having a strategy improves learning. Their immediate significance is organizational: widespread activity and formal institutional direction are different things, and a university needs to know which it is counting.
The study’s recommendations include student participation in AI governance and clearer assessment guidance. A separate institutional initiative helps make that discussion less abstract. On March 17, 2026, UNAM announced the creation of its Consejo Coordinador de Inteligencia Artificial. Its responsibilities include university-wide strategic policies, ethical standards and personal-data protection. The announcement also calls for projects responsive to the linguistic, cultural and socioeconomic conditions of Mexico and Latin America.
That last responsibility deserves attention. A campus deciding how to use AI is also deciding whose questions its systems should be able to handle. Local languages and research priorities should shape the work from its beginning. Whether UNAM’s council achieves those ambitions will require evidence about implementation; the creation of a body establishes responsibility without demonstrating its eventual results.
At Tecnológico de Monterrey, the academic-integrity website provides separate routes to AI guidance for students and teachers and directs employees to an internal institutional policy. The public page shows how guidance can be organized for different users. It does not establish how consistently those rules are understood or enforced. A student reading the page still needs course-level clarity about a particular assignment.
A useful campus policy should make ordinary decisions easier to explain. Can a student use a tool to translate a draft? What must be disclosed? Who can challenge an automated judgment? Where should a researcher take a data-privacy concern? These are practical questions for evaluating a policy, rather than allegations about any institution named here.
For Latin American universities, the next meaningful announcement may be less dramatic than a product launch: an understandable rule, a named person responsible for it, and a way to revise it when experience exposes a problem. The educational value of AI will depend in part on whether students can participate in those decisions before the decisions become routine.