Mapping the thematic evolution of AI literacy research: a text-mining analysis of Scopus-indexed articles

Abstract

AI literacy has evolved into a multifaceted socio-technical phenomenon that extends far beyond a specialized skill. Research on AI literacy is inherently interdisciplinary, covering fields such as curriculum studies, educational ethics, and policy development. This paper examines the thematic evolution of AI literacy research through a detailed textual analysis of 881 publications indexed in the Scopus database between 2016 and 2025. Using natural language processing techniques, co-occurrence analysis, and topic modeling, we identified dominant themes and disciplinary shifts related in AI literacy research. The results showed a significant increase in AI literacy-related publication from a single publication in 2016 to 714 publications in the combined years 2024 and 2025. This growth is accompanied by a clear expansion in disciplinary scope, with a rising concentration of studies in the social sciences and a psychology. The results identified eight key topics in AI literacy studies during the formative phase, early education, human-AI interaction, AI-supported teaching and learning, design and ethical considerations, and the intersection of technical, pedagogical, and socio-cultural perspectives. There is a rapidly growing global contribution to AI literacy research, with East Asian countries, especially Hong Kong, joining leading nations such as the United States, China, Germany, and the United Kingdom.