The mediating role of engagement in the relationship between performance expectancy, effort expectancy, and students' behavioral intention to use ChatGPT

This study investigates the factors influencing university students' behavioral intention (BI) to use ChatGPT for academic learning, emphasizing the mediating role of learning engagement (LE) between performance expectancy (PE), effort expectancy (EE), and BI. Grounded in the Unified Theory of Acceptance and Use of Technology (UTAUT), the study employed a quantitative, cross-sectional design involving 215 students from the University of Jeddah. The data were analyzed using structural equation modeling (SEM) via AMOS. The results revealed that PE (β = 0.666, p < 0.001) and EE (β = 0.398, p = 0.012) significantly influenced students' engagement with ChatGPT, with PE emerging as the stronger predictor. By contrast, PE and EE had non-significant direct effects on BI (p > 0.05). LE exhibited a strong and significant direct effect on BI (β = 0.537, p < 0.001). Mediation analysis confirmed that LE fully mediated the relationships between PE and BI (β = 0.358, p = 0.003) and between EE and BI (β = 0.213, p = 0.035). These findings highlight the pivotal role of engagement in translating expectancy beliefs into students' BI to utilize ChatGPT. This study extends traditional technology acceptance models by positioning engagement as a key psychological mechanism linking cognitive appraisals to technology adoption. Practically, the results suggest that enhancing student engagement with artificial intelligence (AI) tools can strengthen motivation, confidence, and responsible generative AI use in higher education.