Multidimensional perceived school support and AI learning anxiety among university students: the mediating roles of...

As artificial intelligence (AI) becomes embedded in university learning, students may experience anxiety when they doubt their capacity to learn with it or question its academic value. This study examines the direct association between multidimensional perceived school support and AI learning anxiety and the indirect associations involving control-value appraisals, represented by AI learning self-efficacy and perceived usefulness.

Data from 547 Chinese university students were analyzed using structural equation modeling (SEM) to test direct, parallel indirect, and sequential indirect paths. Specific and sequential indirect paths were evaluated using 5,000 bootstrap resamples and 95% bias-corrected confidence intervals. Supplementary artificial neural network (ANN) analyses assessed predictive performance and relative predictor importance.

Greater perceived school support is associated with higher AI learning self-efficacy (β = 0.469, p < 0.01) and perceived usefulness (β = 0.223, p < 0.01), and with lower AI learning anxiety (β = -0.160, p < 0.01). Self-efficacy is positively associated with perceived usefulness (β = 0.587, p < 0.01). Both self-efficacy (β = -0.408, p < 0.01) and perceived usefulness (β = -0.345, p < 0.01) are associated with lower AI learning anxiety. Together, the indirect paths account for 69.5% of the total standardized association. ANN analyses identify self-efficacy and perceived usefulness as the strongest and most stable predictors of anxiety.

The findings characterize AI learning anxiety as an achievement-related emotion associated with students' perceived support and their control-value appraisals. Universities may therefore consider combining encouragement, clear guidance, and practical resources to support students' confidence and recognition of AI's academic value.