Seven classrooms, one lesson, a continent: Latin America trains its own AI biologists

When Jose Arturo Molina-Mora and colleagues at the University of Costa Rica looked at artificial intelligence in the life sciences, the talent was already there — what Latin America lacked was training built from its own data, languages and questions. Their answer, BiotrAIn, run with EMBL's European Bioinformatics Institute and the CABANAnet network and funded by the Chan Zuckerberg Initiative, is now documented in Frontiers in Artificial Intelligence. A 25-person pilot in Costa Rica grew into seven simultaneous classrooms in April 2026 — 169 participants from 12 countries, taught by 57 trainers — and then an online course that drew 260 learners from 17 countries. The datasets come from Latin American projects, the local trainers stay behind when the course ends, and the whole curriculum sits openly on GitHub for anyone to reuse.

SDG 4 · Quality education — Mechanism: institutional program