Abstract
Abstract
With the growing integration of AI in higher education, it remains unclear how AI literacy relates to academic performance among non-STEM students. This study examined the sequential mediating roles of learning management system (LMS) quality and blended-learning satisfaction (BLS) in the relationship between AI literacy and academic performance. 432 students from the non-STEM at Kwame Nkrumah University of Science and Technology in Ghana participated in a quantitative cross-sectional survey. Partial least squares structural modelling (PLS-SEM) in SmartPLS 4 was used for the data analyses. AI literacy strongly predicted LMS quality (β = .814, p < .001) and showed a notably significant effect on academic performance (β = .222, p = .001). The LMS quality significantly predicted BLS (β = .604, p < .001), which then predicted academic performance (β = .535, p < .001). LMS quality did not significantly influence academic performance (β = .074, p = .233). It did not independently mediate the correlation between AI literacy and academic performance (β = .061, p = .236). In contrast, in a direct sequential pathway, LMS quality proved significant (β = .263, p < .001). University policymakers should support students with blended learning support and organise AI literacy training in addition to funding digital infrastructure. AI literacy strongly correlates with academic performance when digital systems are built to support rewarding blended learning experiences.