Mapping artificial intelligence research in primary education: a bibliometric study

Abstract Artificial intelligence (AI) has increasingly gained attention in primary education due to its potential to support personalized learning, adaptive assessment, and instructional decision-making. Despite the rapid growth of publications, research in this field remains fragmented, with limited understanding of its intellectual structure and thematic evolution. This study aims to systematically map the development, research trends, and thematic trajectories of AI research in primary education through a bibliometric approach. A total of 547 peer-reviewed journal articles indexed in the Scopus database from 2018 to 2025 were analyzed using performance analysis and science mapping techniques supported by VOSviewer. The findings reveal a significant increase in publication output, indicating the consolidation of AI in primary education as a distinct research domain. Keyword co-occurrence and cluster analysis identify three major thematic clusters: (1) educational technology and teacher capacity building, (2) AI-driven instructional innovation and immersive learning, and (3) humanistic and developmental perspectives in primary education. Thematic evolution analysis demonstrates a clear shift from general digital technology integration (2018–2020), to institutional adaptation and teacher readiness (2021–2022), and finally toward an AI-centered and application-oriented research focus emphasizing personalized learning and assessment (2023–2025). However, ethical considerations, child-centered impacts, and long-term developmental outcomes remain underrepresented. This study provides a comprehensive overview of the intellectual landscape of AI research in primary education and offers strategic insights for future research agendas, emphasizing the need for ethically grounded, pedagogically informed, and developmentally appropriate AI implementation.