Pathways linking AI dependence to psychological distress: the mediating role of self-esteem and the role of academic achievement
TL;DR
These findings identify SE as a plausible statistical pathway, but they do not establish a causal mechanism; restricted CGPA variability, gender imbalance, purposive sampling, self-report measurement, and incomplete cultural validation of the Chinese AI-dependence scale warrant cautious interpretation.
Abstract
The growing integration of artificial intelligence (AI) into educational contexts has raised questions about whether reliance on AI tools is associated with students’ psychological well-being. This three-wave time-lagged survey examined temporally ordered associations between AI dependence (AD), self-esteem (SE), and psychological distress (PD), and evaluated academic achievement, indexed by cumulative grade point average (CGPA), as an additional predictor and potential moderator. Purposive sampling recruited undergraduate students from two universities in China. Of 947 baseline participants, 865 provided complete and valid data across all three waves. AD, SE, and PD were assessed using self-report measures, and CGPA was self-reported. PLS-SEM in SmartPLS 4.0 and supplementary analyses in SPSS 26.0 tested direct, indirect, and interaction associations. Greater AD was associated with higher PD and lower SE. SE was negatively associated with PD and showed a significant statistical indirect association between AD and PD. CGPA was positively associated with SE but was not directly associated with PD; its indirect association through SE was significant. The hypothesized moderating effects of CGPA were not supported. These findings identify SE as a plausible statistical pathway, but they do not establish a causal mechanism; restricted CGPA variability, gender imbalance, purposive sampling, self-report measurement, and incomplete cultural validation of the Chinese AI-dependence scale warrant cautious interpretation.