Beyond Perceived Usefulness: A Human-Centered AI Acceptance Framework for Teacher Education

Abstract

The rapid integration of artificial intelligence (AI) into educational ecosystems necessitates a deeper understanding of how future teachers conceptualize AI-supported tools. This study examines pre-service teachers’ perceptions across four dimensions: perceived usefulness, digital self-efficacy, trust in AI, and perceived ethical risks. A quantitative descriptive-correlational design was employed with a sample of 100 pre-service teachers from a Romanian university. Data were analyzed using descriptive statistics, reliability testing, exploratory factor analysis, Pearson correlations, and multiple linear regression. Results indicate high levels of perceived usefulness (M = 3.99) and ethical risk awareness (M = 4.58), alongside moderate digital self-efficacy (M = 3.64) and trust in AI (M = 3.65). Perceived usefulness and digital self-efficacy emerged as strong, significant predictors of trust in AI, collectively explaining approximately 70% of the variance. Contrary to assumptions that ethical concerns hinder technology acceptance, perceived ethical risks did not significantly predict trust and showed a positive association with perceived usefulness. The findings suggest that ethical awareness and technological trust can coexist within a human-centered pedagogical stance. Building on these results, we propose the Teacher-AI Agency Framework (TAAF) as a preliminary model positioning professional agency as the integration of pedagogical utility, digital competence, informed trust, and ethical reflection. Implications for teacher education, instrument validation, and responsible AI integration are discussed.