AI-TPACK and AI use intention: the mediating role of self-efficacy and the moderating role of AI dependence among university teachers

Although Artificial Intelligence (AI) is reshaping higher education, the psychological mechanisms linking university teachers’ AI-related professional knowledge to AI use intention remain underexplored. Grounded in Social Cognitive Theory and the AI-TPACK framework, this study examined a moderated mediation model among 407 teachers at a normal university in southwestern China via stratified random sampling. PROCESS macro analyses indicated that AI-TPACK significantly predicted AI use intention (β = .39, p < .001), and that AI self-efficacy partially mediated this relationship (indirect effect = 0.087, 95% CI [0.054, 0.126]), accounting for 29.60% of the total effect. AI dependence significantly moderated both the first-stage path to AI self-efficacy (b = −0.10, p = .027) and the direct path to AI use intention (b = −0.08, p = .033), with the index of moderated mediation reaching −0.027 (95% CI [−0.054, −0.003]). One-way ANOVA further revealed significant disciplinary differences in AI-TPACK (ηp2 = .162), AI self-efficacy (ηp2 = .101), and AI use intention (ηp2 = .045), with Natural Sciences teachers scoring highest, whereas AI dependence remained stable across disciplines. These findings clarify how AI-TPACK shapes AI use intention through confidence and reflective engagement, and inform discipline-sensitive faculty development for AI-enhanced learning environments.