Learning in AI-Augmented Environments Through Dual Pathways of AI Literacy

Generative artificial intelligence (AI) tools are becoming embedded in higher education, raising questions about how learning unfolds in AI-augmented environments. This convergent mixed-methods study examines how three dimensions of AI literacy - technical, cognitive, and social-ethical - relate to learning processes and outcomes. Qualitative analysis of ten interviews with college students shows that learners engage with AI pragmatically but cautiously, using it to support efficiency, comprehension, and academic work while actively regulating use to preserve learning quality and integrity. Quantitative survey results show that technical AI literacy is associated with greater use of AI-assisted learning behaviors, which are in turn associated with higher perceived learning gains. Cognitive AI literacy is directly associated with perceived learning gains and stronger social-ethical awareness, whereas social-ethical AI literacy shows no direct association with learning behaviors or outcomes. Together, the findings indicate that learning in AI-augmented environments is shaped by two complementary pathways: 1) a technical–behavioral pathway that supports effective learning engagement with AI tools, and 2) a cognitive–ethical pathway that supports evaluative judgment and responsible use. These results highlight the importance of designing learning environments and assessment practices that support both procedural fluency and reflective judgment in AI-integrated higher education.