Exploring psychological pathways to pedagogical AI overreliance among chinese teachers through cognitive–affective–conative framework

This sequential explanatory mixed-methods study examines how Chinese teachers’ post-adoption engagement with AI may develop into pedagogical overreliance. Drawing on a cognitive-affective-conative framework, quantitative data from 1,457 teachers and follow-up interviews with 16 participants were used to investigate how teachers’ evaluations of AI shape affective responses and, in turn, dependence on AI in instructional work. The findings showed that perceived pedagogical usefulness, reliability, ease of use, and workload reduction generally predicted trust, flow, and attachment, although ease of use did not significantly predict trust. Flow and attachment significantly predicted pedagogical AI overreliance, whereas trust had no direct effect. Qualitative findings indicated that overreliance was not simply frequent use, but a gradual reduction in pedagogical independence through smooth, routinized, and emotionally comfortable AI use.