Free AI courses in Spanish: access is no longer the excuse, habit is still the real barrier

By Zendoric · September 9, 2026.

Digital Learning Week 2026 is taking place from September 8 to 11 at UNESCO headquarters in Paris, under the theme "Education in the AI era: Facts | Frictions | Frontiers," bringing together education policymakers, teachers and researchers to discuss how artificial intelligence is transforming the way we teach and learn. Coinciding with that event, Mexico's catalog of free artificial intelligence (AI) training in Spanish, aimed at people with no prior technical background, continues to grow.

The facts are concrete and verifiable on the platforms themselves. The Secretaría de Educación Pública (SEP) offers on Cursos @prende.mx the 20-hour program "Inteligencia Artificial aplicada al aprendizaje," designed for teachers but open to the general public, alongside "Herramientas de IA para la práctica docente," developed with Microsoft. IBM SkillsBuild offers "Fundamentos de la IA: bases para comprender la IA" (nine hours) and a 90-minute introduction that can be sampled without registering. Microsoft Learn adds paths such as "Trabaje de forma más inteligente con IA," lasting 42 minutes and aimed at professional users with no programming background. All these courses are in Spanish and free, and some award a certificate or digital credential on completion.

This offering is not an isolated philanthropic gesture: it is market strategy. IBM and Microsoft have spent years giving away basic training in their own tools because every person trained in Copilot or in the fundamentals of machine learning is, potentially, a future customer or an employee who already knows how to operate their ecosystem. It is the same calculation that has led the frontier of open models —those that can be run without paying a license— to advance at a pace comparable to that of closed models: democratizing access to AI, whether in the form of a model or a course, benefits first whoever offers it for free and only afterwards, if at all, the user. That the effect is positive for the user does not make it pure altruism, and that is worth bearing in mind when assessing these initiatives.

Our reading is that the real bottleneck in AI literacy was never the price of the course, but the friction involved in finishing it: time, persistence and the lack of an immediate professional incentive. We have pointed out in earlier analyses that AI's impact on employment is uneven across sectors, and that administrative and back-office work is the most exposed in the short term. Faced with that, the most sensible individual response is not to resist or wait for things to "settle down," but to invest a few hours —sometimes 90 minutes is enough to understand what a language model can and cannot do— in no longer treating AI as a black box. It is a cheap defense, though insufficient on its own, against a transition that in the short term will displace tasks and entire roles.

In the long term, the very exercise of "learning AI for free" should become less and less necessary in its current form: if the thesis of a technology that generates abundance and frees up time holds, the tools will be so intuitive that the friction of using them will disappear, just as no one today needs a 20-hour course to use a search engine. What will endure as a permanent skill will not be handling a specific product —Copilot, Gemini or whatever comes along in 2030— but the judgment to verify, question and responsibly use what AI proposes. That is exactly the ground these initiatives are beginning, modestly, to cover.