<![CDATA[Higher education is rapidly adopting AI-supported learning systems, yet the effectiveness of these tools depends on how students engage with them psychologically, not merely on their availability. However, mere access to AI tools does not automatically translate into meaningful student engagement, indicating a psychological “adoption gap” between technology availability and learners’ active involvement. This study aims to test how key AI features AI usage, personalization/adaptivity, and feedback/analytics relate to student engagement, while examining technology engagement as a mediating mechanism that explains how AI features become educationally effective. Using a quantitative, non-experimental cross-sectional survey of 71 undergraduate students in Eastern Indonesia, the proposed model was analyzed using PLS-SEM (SmartPLS 4) to estimate direct and indirect effects. The model demonstrated strong predictive power, explaining 74.4% of the variance in technology engagement (R² = 0.744) and 66.4% in student engagement (R² = 0.664). AI personalization/adaptivity emerged as the strongest driver, significantly predicting technology engagement (β = 0.516, p < 0.001) and also exerting a significant direct effect on student engagement (β = 0.310, p = 0.010), whereas AI usage and feedback did not show significant direct effects on student engagement but exhibited significant indirect effects through full mediation by technology engagement. These findings imply that technology engagement functions as a “gatekeeper”: institutions should prioritize adaptive personalization and deliberately cultivate students’ sense of control, competence, and psychological involvement with AI systems, rather than relying on high usage intensity or automated feedback alone to drive engagement.]]>
Copyrights © 2026
