How AI anxiety and contextual factors shape employee deviant innovation

Although employees' deviant innovative behavior lacks formal authorization, it may become a key driver of organizational innovation performance in the digital era. However, its generation mechanism-shaped by the interplay of psychological perception, organizational design, and external institutional forces-remains underexplored, particularly in the Chinese context. To investigate the motives for innovation in a turbulent era, this paper adopts the Technology-Organization-Environment (TOE) framework as the analytic system and applies fuzzy-set Qualitative Comparative Analysis (fsQCA) to carry out a configurational study of sample data from 326 corporate employees in China. Focusing on six antecedent conditions including AI anxiety, generative AI skills, gamification of platform work, autonomous decision-making, job insecurity and government support. Based on the empirical study, the following have been found: (1) No single element in the system needs to be triggered to motivate employees' deviant innovation behavior; (2) There are three equivalent system pathways for achieving high levels of employee deviant innovation behavior, categorized as the Anxiety-autonomy driven pathway, the External supported driven pathway, and the Passive triggered driven pathway; (3) A specific combination of conditions inhibits deviant innovation (Low activation environment inhibited type), demonstrating asymmetric causal complexity with the high-drive pathways. This study breaks through the limitations of traditional linear analysis, broadens the application boundaries of the TOE framework in innovation management from a complex systems perspective, and provides valuable theoretical support and practical guidance for systematically stimulating employees' innovative potential in localized digital contexts.