Examining the Impact of Generative AI on Student Motivation and Engagement: The Mediating Role of Autonomy-Support and Autonomous Motivation in Education

Synthesis: Ahmed and Sultan (2026) investigated how perceived autonomy, competence, relatedness, expectancy, and value influence autonomy support for AI use, autonomous motivation, and ultimately student motivation and engagement in GenAI-supported learning. Integrating Self-Determination Theory, Expectancy-Value Theory, and the Technology Acceptance Model, and analyzing data from 297 undergraduates and postgraduates at King Saud University (Saudi Arabia) with PLS-SEM, they found that autonomy support and autonomous motivation significantly increased student motivation, which emerged as the strongest predictor of student engagement. Perceived expectancy showed no significant influence, and perceived competence did not significantly affect autonomy support.

Key Findings

• Perceived autonomy, perceived relatedness, and perceived value significantly enhanced autonomy support for AI use, while perceived autonomy, competence, and relatedness positively influenced autonomous motivation for AI use.
• Autonomy support and autonomous motivation significantly increased student motivation, which subsequently emerged as the strongest predictor of student engagement.
• Perceived expectancy showed no significant influence on either autonomy support or autonomous motivation, and perceived competence did not significantly affect autonomy support.
• The study integrates SDT, EVT, and TAM within a single framework to explain student engagement in generative AI-supported learning environments, addressing the gap of single-theory models.
• It provides context-specific evidence from Saudi higher education, extending AI-in-education research beyond Western and technologically advanced contexts.

Study Design & Method

A quantitative research design was used with data from 297 undergraduate and postgraduate students at King Saud University. The proposed model was analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). Constructs spanned perceived autonomy, competence, relatedness, expectancy, and value (from SDT and EVT); autonomy support for AI use and autonomous motivation for AI use (contextualized mediators); and student motivation and student engagement (outcomes). The integrated framework allowed the authors to trace how psychological need satisfaction and technology acceptance perceptions translate into motivation and engagement in AI-supported learning.

What this means for practice

• Instructors. Design for Motivation rather than for tool access: student motivation was the strongest predictor of engagement (β = 0.74), while autonomy support and autonomous motivation for AI use had no direct effect on engagement (β = 0.070 each).
• Instructors. Give students real control over how they use Generative AI, since perceived autonomy raised both autonomy support for AI use (β = 0.290) and autonomous motivation for AI use (β = 0.170), and build competence through feedback and skill development rather than expecting it to create autonomy support — perceived competence predicted autonomous motivation (β = 0.130) but not autonomy support (β = 0.010).
• Faculty developers. Build relatedness into AI-supported tasks — collaborative and inclusive uses, not only individual tool time — because perceived relatedness predicted autonomy support (β = 0.200) and autonomous motivation (β = 0.120).
• Instructors. Spend instructional time on the educational value of the tools rather than on performance expectations: perceived value predicted autonomy support (β = 0.350), while perceived expectancy predicted neither autonomy support (β = 0.060) nor autonomous motivation (β = 0.050).
• Administrators. Treat autonomy-supportive teaching as the integration strategy: autonomy support drove autonomous motivation (β = 0.370) and student motivation (β = 0.280), and both reached engagement only through motivation.

Limitations

• The cross-sectional design limits causal inference, and the cross-sectional PLS-SEM analysis cannot establish temporal ordering among the mediators and outcomes.
• The sample is drawn from a single Saudi university, bounding generalizability to other national and institutional contexts.
• All measures are self-report.
• The integration of three theories, while comprehensive, relies on the specific operationalization of autonomy support and autonomous motivation for AI use developed for this context.

Citation

Ahmed, A., & Sultan, A. (2026). Examining the impact of generative AI on student motivation and engagement: The mediating role of autonomy-support and autonomous motivation in education.