Costa Rica's National AI Guide for Teachers — 'Inteligencia Artificial en Clase'

Costa Rica's Ninth State of Education Report (2023) found only three in ten students reach expected learning competencies, with many secondary students reading and calculating at levels typical of much younger children. Generative AI was already reaching classrooms informally and unevenly when the Ministry of Public Education (MEP) began building a national response.

In February-March 2026, MEP — working with the UN System in Costa Rica, UNESCO and the private Universidad Latinoamericana de Ciencia y Tecnología (ULACIT) — launched 'Inteligencia Artificial en Clase: guía para docentes', a seven-chapter national guide aimed at all 65,000 public-school teachers and around one million students, intended to keep teachers' professional judgement central rather than delegate decisions to AI.

The guide covers AI fundamentals, prompt design and classroom project design, and sets out explicit ethical and data criteria for tool selection: bias reduction, discrimination prevention, filtering of inappropriate automated content, privacy protection and data-accuracy checks.

The guide was only just being rolled out as the 2026 school year began; the Ministry's headline claim of a projected 30% improvement in maths and science outcomes is a projection, not a measured result, and no independent measurement of learning impact, teacher uptake or actual classroom data-handling practice yet exists.

This is a governance and capacity-building intervention rather than a proven-impact one at this stage: its value lies in setting a national ethical and pedagogical baseline for AI use ahead of adoption, with outcome evidence still to come.