Artificial Intelligence in Education: A Topic Analysis of Large Language Model Applications

The fast development of artificial intelligence (AI) has led to the widespread adoption of large language models (LLMs) in many fields. Over time, education has become one of the areas where LLMs have seen the most rapid growth and experimentation. Even so, the literature on the use of LLMs in education is characterized by a fragmented thematic structure, thus requiring the creation of a clear synthesis of dominant and emerging research directions. The purpose of this study is to systematically analyze the thematic structure of the existing literature, identify the main subfields investigated, and highlight emerging directions and future research opportunities. To achieve this objective, this paper conducted a systematic review of the literature, using articles indexed in the Web of Science database and taking into account PRISMA 2020. The final dataset consisted of 2,610 scientific articles published between 2021 and 2025. A combined methodological approach was applied in the analysis, integrating the thematic map of the authors’ keywords, the Latent Dirichlet Allocation (LDA) method, and BERTopic. The results showed an unbalanced thematic distribution of the literature, with an emphasis on the educational applications of LLMs and medical and clinical education, which serves as the main field of validation for these technologies. By using several methods of analysis, this study provides a coherent and in-depth perspective on the current state of research on LLMs in education, highlighting both established directions and existing gaps, thus providing a solid basis for future research.