The Dark Side of Generative AI: Sequential Psychological Mechanisms and Heterogeneous Responses Underlying Teachers' Innovative Behavior.

PURPOSE: Although generative artificial intelligence (GAI) has become increasingly integrated into higher education, its potential psychological costs for teachers remain insufficiently understood. Drawing on the Transactional Model of Stress and Coping (TMSC), this study examined how GAI use influences teachers' innovative behavior through sequential cognitive and emotional mechanisms while exploring heterogeneity in teachers' psychological responses. METHODS: A three-wave time-lagged survey was conducted among 898 university teachers from public higher education institutions in Northeast China. Structural equation modeling (SEM) was used to examine the proposed sequential mediation model, and latent profile analysis (LPA) was performed to identify distinct psychological response patterns. RESULTS: Within the present study, greater generative AI use was associated with lower levels of teachers' innovative behavior. Job insecurity and AI anxiety independently mediated this relationship and jointly formed a significant sequential mediation pathway. In addition, five distinct psychological profiles were identified. Teachers characterized by lower job insecurity and AI anxiety exhibited significantly higher levels of innovative behavior than those experiencing greater psychological risks. CONCLUSION: By integrating variable-centered and person-centered approaches within the TMSC framework, this study advances understanding of the psychological mechanisms and heterogeneous responses underlying teachers' adaptation to generative AI. The findings contribute to the emerging literature on the psychological consequences of AI adoption and provide practical implications for promoting sustainable teacher innovation and effective psychological management during AI-enabled educational transformation.