Academic Pressures, Generative AI Use, and Academic Behaviors and Perceptions Among University Students: An Empirical Study

This study draws on cognitive offloading theory and self-regulated learning theory to investigate why Chinese university students use Generative AI (GenAI) and what academic correlates are observed. Using a cross-sectional survey design, data were collected from 500 university students and analyzed through structural equation modeling (SEM) and bootstrap mediation analysis. The results show that academic workload, time pressure, and sensitivity to grades are positively associated with GenAI usage, whereas sensitivity to quality is not. In addition, GenAI usage is positively associated with procrastination and self-reported memory difficulties. Regarding the competing hypotheses on creative engagement, GenAI usage was positively associated with creative engagement, which may indicate that students may use GenAI for ideation rather than answer retrieval. Mediation analyses confirm that GenAI usage significantly mediates the associations between academic workload, time pressure, sensitivity to grades, and the three outcomes. However, the indirect effects of sensitivity to quality on the three outcomes were not significant, consistent with the unsupported hypothesis. The study highlights the need to strengthen students’ AI literacy training, manage academic workload, and redesign assessment methods so as to balance responsible use with academic integrity. Future research should employ longitudinal designs, multi-method measures, and explore these relationships in diverse educational contexts to gain a more comprehensive understanding.