The relationship between academic self-efficacy and GenAI dependence among college students: a chain mediation model of academic stress and perceived usefulness of GenAI

BackgroundAs generative artificial intelligence (GenAI) becomes increasingly integrated into higher education, concerns have emerged regarding students' growing reliance on GenAI for academic tasks. Although prior studies have examined the educational benefits of GenAI, limited research has explored the psychological mechanisms underlying students' dependence on these technologies. This study investigated whether academic self-efficacy predicts GenAI dependence and examined the mediating roles of academic stress and perceived usefulness of GenAI.MethodsUsing a time-lagged survey design, data were collected from 428 college students from five universities in eastern China. Participants completed validated measures of academic self-efficacy, academic stress, perceived usefulness of GenAI, and GenAI dependence. Structural equation modeling and bootstrap analyses were conducted to examine the hypothesized mediation pathways.ResultsAcademic self-efficacy was significantly and negatively associated with GenAI dependence. Academic stress and perceived usefulness of GenAI each served as significant mediators between academic self-efficacy and GenAI dependence. Furthermore, academic stress was positively associated with perceived usefulness of GenAI, and the sequential pathway from academic self-efficacy to academic stress, perceived usefulness of GenAI, and GenAI dependence was significant. Specifically, students with lower academic self-efficacy experienced greater academic stress, which increased their perceptions of the usefulness of GenAI and subsequently strengthened their dependence on GenAI.ConclusionAcademic self-efficacy affects college students' GenAI dependence both directly and indirectly through academic stress and perceived usefulness of GenAI. The findings extend the application of the I-PACE framework to the context of generative artificial intelligence and provide practical insights for universities seeking to promote responsible GenAI use, enhance students' independent learning capabilities, and mitigate excessive reliance on AI-assisted learning tools.