Creative self-efficacy and innovative teaching behaviour in AI-supported teaching contexts: the serial mediating roles of teacher resilience, job autonomy, and perceived organizational support

Background As artificial intelligence (AI) teaching has escalated, it is critically important to explore the effectiveness of coordinating teachers’ psychological complexity with respect to AI practical acceptance within this emerging context. Drawing on Conservation of Resources (COR) theory, this study examines the serial mediation mechanisms through which creative self-efficacy is associated with innovative teaching behaviours in AI-supported teaching contexts. Methods A total of 371 teachers within AI-supported teaching contexts were invited via a cross-sectional design. Questionnaires were distributed using the Creative Self-Efficacy Scale, Teachers’ Innovation Behaviour Scale, Resilience Scale, Job Autonomy Scale, and Perceived Organizational Support Scale. Results The findings demonstrated significant positive interrelations among the variables and supported a statistically significant serial mediation model with teacher resilience, job autonomy, and perceived organizational support as mediators in explaining the association between CSE in innovative behaviour within the present AI-supported teaching context. The direct effect represented 60.49% of the total effects (0.74), while the indirect effect (of the 7 total indirect effects) accounted for 39.51% of the total effect (0.74). Notably, teacher resilience was associated with the strongest mediator and exerted relatively stronger effects than job autonomy and perceived organizational support. Conclusion This study extends the literature by integrating psychological resources and institutional enablers into a multilevel pathway of innovation in the current AI educational initiative. By comparatively testing internal psychological and external organizational mediators, this study illustrates how these factors play distinct roles in relation to teaching in AI-supported contexts where innovation is essential. It also helps bridge the gap between psychological theory and technology-supported innovative pedagogy and offers directions for future research in AI-mediated teaching environments.