AI and education: Scaling teacher support with AI: Evidence from Chile and Costa Rica

preparation

This TC aims to generate rigorous evidence on how artificial intelligence can be leveraged to deliver high‑quality, relevant, and cost‑effective support to teachers and tutors at scale, ultimately improving educational outcomes in Latin America. Teachers across the region face significant challenges—insufficient preparation, heavy workloads, large class sizes, limited institutional support, and difficult working conditions—that contribute to widespread burnout, declining mental health, and high attrition. While mentoring programs, psychoeducational support, and high‑dosage tutoring have shown strong potential to improve teacher wellbeing, instructional quality, and student learning, these approaches remain difficult to scale due to their dependence on trained human providers. AI offers a promising way to augment these interventions without reducing the personalization that drives their effectiveness. This TC adopts a phased strategy: Phase 1 (2026) will design, adapt, and pilot AI‑enabled delivery models; Phase 2 (2027) will conduct randomized controlled trials (RCTs) to assess effectiveness, cost‑efficiency, and scalability across the region.