The Future of Education
Large Language Models in Education: A Thematic and Topic Analysis of Recent Research
Corina Ioanăș, Bucharest University of Economic Studies, Romania (Romania)
Bianca Cibu, Bucharest University of Economic Studies (Romania)
Camelia Delcea, Bucharest University of Economic Studies (Romania)
Liviu-Adrian Cotfas, Bucharest University of Economic Studies (Romania)
Abstract
The rapid expansion of research on large language models (LLMs) in education has resulted in a increasing body of literature, dedicated to educational applications and contexts, in a broader environment marked by digital transformation and changing skill requirements in the labour market. While prior studies have focused on various aspects related to LLMs into several research areas, the present study addresses the research themes and topics uncovered in the research papers related to the use of LLMs in education. The corpus of the papers has been extracted from Clarivate Analytics Web of Science Core Collection database and have been analyzed through two complementary approaches, namely thematic maps and topic analysis. The thematic map approach has been used to identify and classify research themes according to their centrality and level of development, allowing the distinction between core, niche, and emerging themes. On the other hand, topic analysis, conducted through the use of BERTopic and Latent Dirichlet Application (LDA) and applied to papers’ abstracts, has provided information related to the latent thematic patterns and dominant research directions within the literature. Thematic analysis underpins core themes in the area of artificial intelligence (AI), ChatGPT, and educational applications of LLMs, while the emerging themes are focusing on ethical considerations, responsible use of AI, and concerns related to academic integrity. Topic analysis further confirms the prominence of student-centered applications and assessment-related discussions. The conclusions can serve as a roadmap for interested parties.
|
Keywords |
artificial intelligence; LLMs; thematic analysis; topic analysis; LDA; BERTopic |
|
REFERENCES |
[1] Bates,T.; Cobo,C.; Mariño,O.; Wheeler, S. Can Artificial Intelligence Transform Higher Education? Int. J. Educ. Technol. High.Educ. 2020, 17, 42. [2] Kieser,F.; Tschisgale,P.; Rauh,S.; Bai,X.; Maus,H.; Petersen,S.; Stede,M.; Neumann,K.; Wulff,P. David vs. Goliath: Comparing Conventional Machine Learning and a Large Language Model for Assessing Students’ Concept Use in a Physics Problem. Front. Artif. Intell. 2024, 7, 1408817. [3] Morris,W.; Holmes,L.; Choi,J.S.; Crossley, S. Automated Scoring of Constructed Response Items in Math Assessment Using Large Language Models. Int. J. Artif. Intell. Educ. 2025, 35, 559–586. [4] Rahman,M.M.;Watanobe,Y. ChatGPT for Education and Research: Opportunities, Threats, and Strategies. Appl. Sci. 2023, 13, 5783. |
Acknowledgement: This work was funded by the EU’s NextGenerationEU instrument through the National Recovery and Resilience Plan of Romania - Pillar III-C9-I8, managed by the Ministry of Research, Innovation and Digitization, within the project entitled „JobKG—A Knowledge Graph of the Romanian Job Market based on Natural Language Processing”, contract no. 760274/26.03.2024, code CF 178/31.07.2023
