Medellín teaches AI to teenagers through urban music: literacy, not just the model, is the race that matters

By Análisis Urbano · July 20, 2026.

Medellín's Secretariat of Education opened this July 20 the call for Urban Labs, an initiative within the Experiencia Reto program aimed at eighth, ninth, tenth and eleventh grade students from priority public institutions. During the second semester, teams —of five to ten members, endorsed by their school and accompanied by a teacher— will work with two languages: artificial intelligence and urban music, to analyze their neighborhoods and propose local transformation projects. The 25 best-performing labs will present their results at the Smart Fest, a city-wide event scheduled for October. Registration is open at urbanlabs.medellin.edu.co, according to the Secretariat itself.

This is a local story of modest scope: there's no new model here, no investment figure, no data point that moves the market. But it deserves attention because it touches on an underlying issue that almost always falls off the radar when talking about artificial intelligence: early literacy.

In general, the public debate about AI is dominated by who trains the largest models and who wins the geopolitical race. Much less is said about who, at 14 or 15 years old, has the opportunity to touch these tools with pedagogical support and without fear. That's not a minor detail: the gap in access to AI education —not the computing gap between labs— is what will determine, ten years from now, who designs with these tools and who simply endures them at their job.

Our reading is that programs like this are worth more for the habit they instill than for the technical content taught in a few weeks. Combining AI with urban music —a language these teenagers already identify with— is a smart pedagogical design decision: it narrows the distance between "AI is something for distant labs" and "AI is a tool I can use to tell the story of my block." That early familiarity is, on a small scale, the same principle underlying our core thesis about the work of the future: the inequality of the transition toward AI isn't resolved solely with income or regulation, it's also resolved by who understands how the tool works before the labor market demands it.

We should be honest about the limits: a semester-long school lab within a municipal program doesn't close any structural gap, and its impact will depend on continuity —whether this repeats and scales year after year— more than on this particular call. But as a signal of where public education should be heading in the next decade, it points in the right direction: AI not as an isolated subject, but integrated into challenge-based learning and the cultural identity of those who will use it.