Teachers’ attitudes towards artificial intelligence: a cross-national latent profile analysis

Abstract

Artificial intelligence (AI) is increasingly positioned as transformative in education, yet teachers’ attitudes remain unevenly understood across national contexts. Guided by a conceptual framework informed by the Technology Acceptance Model (TAM), this study used a person-centred approach and data from the 2024 OECD Teaching and Learning International Survey to examine teachers in the United States, Finland, and Japan. Latent profile analysis identified three recurrent attitudinal groupings: Concern-Dominant, Moderate-Benefit/Cautious, and Benefit-Dominant. Although comparable groupings emerged, their prevalence and internal configurations varied across countries. Regression analyses showed that AI-related professional development was consistently associated with more favourable profiles, whereas educational attainment and teaching experience had limited, context-specific effects. Profiles were also closely associated with reported AI use and specific instructional applications, with benefit-dominant teachers reporting greater pedagogical integration. These findings underscore the importance of addressing teachers’ beliefs and concerns to support meaningful AI integration in education.