AI Use and Learning Outcomes Among Undergraduate Students: The Mediating Role of Learning Motivation, Locus of Control, and Learning Preferences

Abstract

Background. The growing use of artificial intelligence (AI) in higher education has created new opportunities for personalized learning, although its contribution to student achievement remains theoretically and empirically contested. Purpose. This study examined the direct and indirect effects of AI use on conceptual understanding and problem-solving ability through learning motivation, locus of control, and learning preferences. Method. A quantitative explanatory design was employed with 88 pre-service teachers enrolled in an Instructional Media course at a teacher education institution in Central Java, Indonesia. Data were collected using validated questionnaires and cognitive assessments. Preliminary analyses were conducted using SPSS, while the measurement and structural models were evaluated through PLS-SEM with 5,000 bootstrap subsamples. Novelty. This study simultaneously tested three psychological and learning-related mediators within a single structural model, thereby extending previous research that primarily examined the direct relationship between AI use and academic performance. Results. AI use significantly predicted learning motivation and locus of control but did not significantly influence learning preferences. Learning motivation significantly predicted learning outcomes and mediated the relationship between AI use and learning outcomes. In contrast, locus of control and learning preferences showed no significant mediating effects. The direct effect of AI use on learning outcomes was also nonsignificant. Conclusion. AI contributes to student learning primarily by strengthening learning motivation rather than through direct technological exposure. Effective AI integration should therefore emphasize reflective, autonomous, and cognitively engaging learning activities.