Artificial Intelligence Adoption and Academic Performance in Higher Education: Testing the Mediating Roles of Digital Competence and Learning Engagement Using Structural Equation Modeling

Abstract

Background of study: Artificial intelligence (AI) tools are rapidly entering higher education, yet adoption alone does not explain whether students achieve better academic outcomes. Aims and scope of paper: This study develops and tests a mediation model in which digital competence and learning engagement transmit the effect of AI adoption on academic performance. Methods: A cross-sectional quantitative design is specified for university students, with multi-item Likert measures and covariance-based structural equation modeling (CB-SEM). The measurement model is evaluated through confirmatory factor analysis, reliability, convergent validity, and discriminant validity, followed by direct and bootstrapped indirect-effect testing. Result: The illustrative model indicates that AI adoption is positively associated with digital competence and learning engagement, while both mediators are positively related to academic performance. The direct AI-adoption effect remains positive but smaller after the mediators are included, indicating complementary partial mediation. Conclusion: The model suggests that universities should treat AI adoption as an educational capability-building process rather than mere tool access, pairing AI use with digital competence development and engagement-oriented pedagogy