How self-efficacy influences Chinese university students' GenAI dependency: A study of the mediating roles of social anxiety and rumination and the moderating role of mindfulness

With the recent, rapid advancement of artificial intelligence (AI) technology, an increasing number of university students are relying on generative artificial intelligence (GenAI) in their daily academic and personal lives. Although self-efficacy is known to influence dependency on GenAI, the mechanism through which this transpires remains insufficiently explored. Working from an I-PACE model, this study collected 758 valid survey responses from Chinese university students and subjected that data to quantitative analyses. The objective was to assess self-efficacy, social anxiety, rumination, mindfulness, and GenAI dependency. The findings indicate that students' dependency on GenAI is positively associated with their self-efficacy, rather than negatively, as predicted. This unexpected relationship is attributed to the distinctive context of higher education in China, which is characterized by high academic pressure, collectivist cultural norms (including tendencies toward conformity), and the illusion of self-efficacy produced by GenAI use. Further analyses reveal that social anxiety and rumination jointly mediate this association through a significant sequential mediation pathway. Specifically, diminished self-efficacy intensifies social anxiety, which subsequently triggers rumination, ultimately increasing GenAI dependency. This is the opposite of the main path result, in which mindfulness exerted a significant moderating effect on this mediation process. Together, these results elucidate a systematic pathway through which self-efficacy shapes university students' GenAI dependency within a specific cultural context. Therefore, the results offer theoretical foundations and practical implications for timely interventions by higher education practitioners from the perspectives of psychological traits and psychological processes.