A study on AI-enabled course development, AI proficiency, and first-year students' academic and psychological...

The rapid integration of digital technologies and generative artificial intelligence (AI) into higher education has reshaped first-year students' learning experiences and self-reported psychological adjustment. Using a cross-sectional self-report design, this study examined the statistical associations among AI-enabled course development, AI proficiency, academic adjustment, and a composite indicator of self-reported psychological adjustment.

A cross-sectional self-report survey was conducted among 551 first-year university students from a single institution. Descriptive statistics, Pearson correlations, multiple and hierarchical regressions, Bootstrap mediation analysis (PROCESS), moderated mediation analysis, and structural equation modeling (AMOS) were employed to examine the hypothesized associations. Causal inference is not warranted given the cross-sectional and self-report nature of the data.

Students reported high satisfaction with AI-enabled course development and strong acceptance of AI learning tools, with over 80% recognizing AI's learning efficiency and actively using AI for study. However, self-reported indicators of psychological strain were endorsed by more than two-thirds of the sample; these items are descriptive self-report indicators and do not constitute clinical diagnoses. Regression analyses showed that course development satisfaction and AI proficiency were statistically and positively associated with emotional management, stress coping, and academic adjustment. Hierarchical regression demonstrated that these variables substantially improved model explanatory power, with the highest explained variance observed for academic adjustment (R ² = 0.240). Mediation analysis indicated that AI proficiency partially mediated the association between course development satisfaction and self-reported psychological adjustment (indirect effect = 0.124, 95% CI [0.076, 0.181]). Moderation analysis revealed that AI-related usage pressure attenuated the positive association between AI proficiency and self-reported psychological adjustment (moderation index = -0.051, 95% CI [-0.095, -0.017]). Structural equation modeling showed acceptable model fit (χ ²/df = 2.31, CFI = 0.956, TLI = 0.941, RMSEA = 0.049, SRMR = 0.041).

In this cross-sectional sample, AI-enabled course development was statistically and positively associated with first-year students' self-reported psychological adjustment, both directly and indirectly through self-reported AI proficiency. Excessive self-reported AI-related stress attenuated these positive associations. Because the design is cross-sectional and self-report-based, these findings should be interpreted as statistical associations rather than causal effects. Universities may consider strengthening AI literacy training, optimizing course design, managing technology-related workload, and providing psychological support to maximize the potential benefits of AI-integrated education; future longitudinal and multi-institution studies are needed to test these associations more rigorously.