LEARNING PATHWAYS IN THE DEVELOPMENT OF AI LITERACY FOR ACADEMIC STUDY: EXPERIENCES OF UNDERGRADUATE STUDENTS AT THE UNIVERSITY OF HONG KONG

The rapid diffusion of generative AI in higher education is transforming how students learn, while simultaneously raising concerns about equity in AI literacy development. This qualitative case study examines how undergraduate students at the University of Hong Kong develop AI literacy for academic purposes and how this process is shaped by multilevel ecological factors. Drawing on Ng et al.’s (2021) AI literacy framework and Bronfenbrenner’s ecological systems theory, the study conceptualizes AI literacy development as a set of learning pathways that unfold through dynamic interactions between individual agency and layered environments. Fifteen undergraduates from diverse disciplines participated in semi-structured interviews. Data were analyzed using reflexive thematic analysis, supported by NVivo, with a deductive–inductive coding strategy guided by ecological systems levels. Findings reveal that the most developmentally salient influences were located in microsystemic contexts. Peers, online communities, and self-directed experimentation with AI tools functioned as key proximal settings in which students acquired procedural know-how, refined prompts through trial-and-error, and developed more critical and reflective AI practices. In contrast, institutional resources were often experienced as misaligned with students’ situated academic needs. At macro- and exo-system levels, regionally bounded platform restrictions and institutionally mediated access regimes structured which AI tools and affordances students could realistically use. These access patterns, in turn, differentially enabled or constrained sustained, higher-order engagement with AI, contributing to stratified AI literacy trajectories and raising equity concerns. The study contributes an ecological learning pathway model of AI literacy development that foregrounds the primacy of microsystemic processes, while showing how broader sociotechnical arrangements channel opportunities for advanced AI use. It offers implications for designing AI literacy initiatives that align institutional support with students’ lived practices and that address equity not only in classrooms but also in regional and institutional governance of AI access.