Trends in Artificial Intelligence and Education Research: A Preliminary Scientometric Analysis in Scopus

Abstract

This study presents a preliminary scientometric analysis of research on Artificial Intelligence (AI) and education, aiming to characterise the global scientific output in this emerging field. After removing duplicates and erroneous records, a total of 8,755 documents (articles, reviews, book chapters, and conference proceedings) indexed in Scopus were analysed. The research was structured in five phases: data collection, extraction, analysis, visualisation, and interpretation. Variables examined included temporal evolution, document typology, geographical distribution, language, open access modality (OA), funding, institutional affiliation, publishers, citation by categories, and compliance with the laws of Price (1963), Bradford (1985), and Lotka (1926). Results indicate that over 75% of the production occurred between 2020 and 2024, positioning the field in a precursor phase (Price, 1963). Scientific articles predominate (87%), followed by reviews (4.93%) and conference proceedings (2.90%). The United States, China, and the United Kingdom lead research output, with English remaining the dominant language. More than 50% of records correspond to Open Access Gold (OAG) publications, concentrated in ten publishers. Most studies do not declare funding sources, and research is dispersed across non-specialised journals. Overall, the data reveal a rapidly expanding field with limited editorial specialisation and low transparency in funding practices, highlighting the need for stronger disciplinary consolidation and greater scientific openness at the intersection of AI and education.