Generative AI use and future-ready learning in higher education: Engagement, AI literacy and responsible learner agency

Abstract

Generative artificial intelligence (AI) is widely used by university students, but its educational value depends on how learners engage with AI outputs, understand system limitations, and adopt responsible practices. This study develops and tests a dual-pathway socio-technical model of future-ready AI-mediated learning. A cross-sectional survey yielded 363 valid responses from students across three university contexts. The measurement model was evaluated through item-level confirmatory factor analysis, reliability and validity testing, and covariance-based structural equation modeling. Generative AI use was positively associated with student engagement, which in turn was positively associated with perceived future-ready learning outcomes. Once engagement was included, the direct association between AI use and perceived outcomes was not statistically significant, whereas bootstrap analysis supported the indirect pathway through engagement. AI literacy was positively associated with responsible AI practice, and institutional AI governance support was positively associated with ethical AI participation awareness. Governance support did not significantly moderate the relationship between AI use and perceived outcomes, suggesting a primarily normative and developmental role rather than a performance-amplifying one. These findings indicate that universities should move beyond simple permission-or-prohibition policies by providing course guidance, strengthening AI literacy, and redesigning assessment to make student judgment, verification, attribution, and learning processes visible.