The relationship between teachers’ GAI use and students’ social-emotional competence: the chain-mediating roles of teachers’ AI literacy and AI self-efficacy

BACKGROUND: The emergence of Generative Artificial Intelligence (GAI) presents both opportunities and challenges for fostering students’ social–emotional competence (SSEC). However, how teachers’ GAI use (TGAIU) influences SSEC through teachers’ AI literacy (TAIL) and teachers’ AI self-efficacy (TAISE) remains underexplored. METHODS: An online survey was conducted among 550 teachers from primary and secondary schools in Central China piloting AI education. Structural equation modeling examined relationships among TGAIU, TAIL, TAISE, and SSEC. Bootstrapping with 5,000 resamples tested mediation effects. Model fit was evaluated using multiple indices. RESULTS: The model showed good fit (χ²/df = 1.545, RMSEA = 0.031, GFI = 0.952, SRMR = 0.032, NFI = 0.957, TLI = 0.982, CFI = 0.985). TGAIU positively predicted SSEC, TAIL, and TAISE. TAIL and TAISE positively predicted SSEC, and TAIL also positively predicted TAISE. Mediation analysis confirmed significant indirect effects: TAIL accounted for 54.89% of the total effect of TGAIU on SSEC, TAISE accounted for 14.93%, and a combined pathway through TAIL and TAISE was significant. CONCLUSIONS: TGAIU is positively associated with SSEC, with TAIL and TAISE playing important chain-mediating roles. Incorporating these teacher-related factors clarifies mechanisms linking technology use and social–emotional learning, emphasizing the importance of targeted teacher training, alignment of technology and pedagogy, and collaborative professional communities in AI-supported education.