Physical Activity in Relation to University Students’ Large Language Model Dependence: Moderated Mediation Effect Between Self-Control and Exploratory Learning Ability

Nowadays, since the rapid development of artificial intelligence on these days, concerns about severe dependence on large language models (LLM) are increasing in the group of university students. This cross-sectional study aims to examine the underlying association between physical activity and LLM dependence, especially the mediating role of self-control and the moderating role of exploratory learning ability between them. Using the stratified cluster sampling, 602 university students in China were reasonably selected and then completed several standardized measurements including Physical Activity Rating Scale-3, Self-Control Scale, Large Language Model Dependence Scale, and Exploratory Learning Scale. Relevant data were analyzed using SPSS 27.0 and PROCESS Model 59. The present work disclose that physical activity was positively correlated with self-control, yet negatively correlated with LLM dependence, and its direct conditional association with LLM dependence was non-significant, whereas the indirect association through self-control was significant. Exploratory learning ability could influence both the positive association between physical activity and self-control and the negative association between self-control and LLM dependence, and the simple slope analyses showed physical activity was associated with greater self-control, and self-control with lower LLM dependence, only at lower exploratory learning ability. These findings clearly identify self-control as a potential influencing pathway and exploratory learning ability as a boundary condition underlying physical activity and LLM dependence. The models only explained 8.8% of the variance in self-control and 19.6% of the variance in LLM dependence, respectively, indicating limited explanatory scope, due to the research design was cross-sectional and the sample selection was regional, so the discovery should be interpreted as preliminary associations of them rather than causal evidence.