Drivers, pedagogical uses, and barriers of teachers’ AI adoption: Cross-national evidence from TALIS 2024

Drawing on TALIS 2024 data from four countries, this study examines AI adoption patterns among 8026 teachers in Singapore, the United Arab Emirates, France, and Japan. Using binary logistic regression, it identifies key correlates across high- and low-adoption contexts. The results indicate that perceived usefulness and self-efficacy are significantly associated with AI adoption, whereas perceived risk is not. Self-efficacy shows a stronger association in low-adoption contexts. Regarding applications, teachers in high-adoption countries are more likely to adopt AI for lesson preparation, content generation, and communication-related tasks, while no clear or consistent pattern of adoption is observed in low-adoption contexts. For non-users, multiple barriers including resource constraints, limited skills, institutional restrictions, value-based reservations, and technology-related anxiety are more prevalent in low-adoption countries. This cross-national comparison highlights multi-level patterns associated with teachers’ AI adoption and suggests that policy should address resource availability, teacher capacity and value-related perceptions to support more effective integration. • Large-scale TALIS2024 data identify key correlates of teachers’ AI adoption. • National context moderates the link between teacher perceptions and AI use. • AI supports planning in high-adoption countries, but shows no stable pattern elsewhere. • Systemic barriers explain non-adoption patterns in low-adoption countries.