Mapping the evolving landscape of AI literacy research: an integrative review combining bibliometric analysis and thematic synthesis

Although AI literacy research has proliferated in recent years, the field remains fragmented across disciplines, methodologies, and educational contexts, limiting coherent knowledge accumulation. This study addresses this fragmentation through an integrative review combining bibliometric analysis with thematic synthesis, examining 838 peer-reviewed studies from Scopus and Web of Science. Bibliometric analysis using CiteSpace and VOSviewer reveals three developmental phases: formative, acceleration, and exponential expansion, with the 2024-2026 period alone accounting for over 90% of the corpus, reflecting the post-ChatGPT diffusion of GenAI technologies. Research production is geographically concentrated, with the United States, mainland China, Hong Kong, and Germany collectively contributing 51.4% of global output, while Africa, Latin America, and South Asia together remain markedly underrepresented. Thematic synthesis identifies ‘educational strategies and curriculum design’ (31.6%), ‘relational studies of AI literacy with adjacent constructs’ (30.5%), and ‘theoretical framework development’ (15.5%) as top three research foci. The field is predominantly empirical (80.9%), with quantitative (55.6%), mixed-methods (26.4%), and qualitative (18.0%) approaches represented among empirical studies. Critical gaps including framework proliferation without theoretical consolidation, the unresolved theoretical relationship between AI literacy and GenAI literacy, near-universal reliance on cross-sectional self-report designs, and underrepresentation of early childhood, vocational, and Global South contexts, require significant future research efforts.