Abstract
<title>Abstract</title> <p>Generative artificial intelligence (genAI) is transforming higher education, yet the mechanisms linking its acceptance to student engagement remain insufficiently understood. This study develops an integrated model to examine how AI attitudes influence student engagement, with AI self-efficacy and intention to use as mediators and supportive social norms as a moderator. Data were collected from 2,568 undergraduate students across multiple universities in China. Partial least squares structural equation modelling (PLS-SEM) was used to test direct, mediating, and moderating effects. AI attitudes positively predict student engagement, both directly and indirectly through the sequential mediation of intention to use and self-efficacy. In addition, supportive social norms strengthen the relationships among attitudes, intention, self-efficacy, and engagement, highlighting their role as critical boundary conditions. These findings provide a comprehensive explanation of how learners move from perception to action in AI-supported learning environments and underscore the importance of integrating individual, motivational, and contextual factors. The study offers both theoretical insights and practical implications for enhancing student engagement in AI-enhanced higher education.</p>