Investigating the relationship between AI literacy, Generative AI, SRL, Self-efficacy, Critical Thinking, and Academic Performance in Higher Education

Abstract

This study examines the relationships among AI literacy, generative AI (GenAI) integration, SRL, self-efficacy, critical thinking, and academic performance in higher education. Despite the rapid adoption of GenAI tools, existing research has largely focused on usage rather than explaining how AI-related competencies influence learning processes and outcomes through an integrated, theory-driven framework. A cross-sectional, non-experimental survey design was employed. Data were collected from 238 undergraduate and graduate students at a large public university in the United States. All constructs were measured using validated instruments, and structural equation modeling (SEM) with robust maximum likelihood estimation was used to test the hypothesized model. The results showed that AI literacy was positively associated with GenAI integration in learning activities, indicating that students with higher levels of AI literacy were more likely to engage in GenAI-supported learning. GenAI integration, in turn, significantly predicted both self-efficacy and critical thinking. Among all variables, self-efficacy emerged as a central factor, significantly predicting SRL, critical thinking, and academic performance. Notably, GenAI integration did not have a direct effect on academic performance. The structural model also revealed indirect relationships among variables. GenAI integration functioned as a pathway linking AI literacy to self-efficacy, while self-efficacy connected GenAI integration to key learning processes and academic performance. In addition, a sequential pathway from AI literacy to academic performance through GenAI integration and self-efficacy was observed. These findings highlight the foundational role of AI literacy in shaping students’ engagement with GenAI tools and underscore self-efficacy as a key mechanism linking technology use to learning and performance outcomes. This study contributes to the literature by proposing and empirically validating an integrated model that explains how GenAI integration influences student learning in higher education.