<b>ChatGPT-assisted mathematics learning in the era of generative AI: A </b><b>bibliometric</b><b> analysis of research trends, collaboration networks, and emerging pedagogical themes</b>

Source abstract

Purpose – This study aims to map the development of research on ChatGPT-assisted mathematics learning during the 2023–2025 period. Methodology – This study uses bibliometric analysis to examine the intellectual, social, and conceptual structures of research on ChatGPT-assisted mathematics learning. Data were obtained from the Scopus database and analyzed using Bibliometrix Biblioshiny. Document selection followed the PRISMA 2020 reporting guidelines as a checklist for the identification, screening, eligibility, and inclusion stages. Of the 1,227 documents identified, 143 met the inclusion criteria and were analyzed further. Findings – The number of publications increased from 12 in 2023 to 36 in 2024 and 95 in 2025, indicating an annual growth rate of 181.37% and an average of 18.63 citations per document. International co-authorship accounted for only 16.78% of these documents, indicating limited cross-national collaboration. ChatGPT, mathematics education, and artificial intelligence were the most dominant keywords, while geometry, prospective mathematics teachers, and personalized learning were relatively underrepresented. Novelty – Unlike previous studies that generally treat ChatGPT as part of generative artificial intelligence, this research analytically positions ChatGPT-assisted mathematics learning as a separate bibliometric unit and proposes AI-mediated mathematics learning as a conceptual proposition that emphasizes critical validation of AI-generated results. Significance – These findings provide an evidence-based research agenda to strengthen international collaboration and develop mathematics instruction that emphasizes reasoning, justification, and critical evaluation of AI-generated solutions, while identifying priorities for future empirical research.