Roles and research trends of ChatGPT-based learning: A bibliometric analysis and systematic review

Abstract

This study analyzed the trends in ChatGPT-based learning research through a systematic review with bibliometric analysis. Based on the PRISMA guidelines, relevant articles were identified by searching several reputable sources (i.e., Embase, Scopus, PubMed, CINAHL, and WoS), and a total of 50 empirical articles that met the selection criteria were included. The results are twofold: (1) Bibliometric attributes of the research were identified. First, we found that authors from China, the United States, Australia, and Saudi Arabia are the most productive contributors to the field. Second, over 64% of the analyzed research was published in educational technology journals (e.g., EIT, AJET, ILE). (2) The systematic review identified key learning features of the research. Most research focused on non-specific learning, and subjects related to English learning in ChatGPT use were the second most frequently researched topic. The third most frequently researched subject is related to STEAM courses, in which most research employed ChatGPT-based assistant learning as an instructional design in the course. The role of ChatGPT-based learning in education involves domain experts, teachers/tutors, administrators, and learning tools. Regarding the educational context of the research, the participants were mainly in higher education settings, including undergraduate and graduate students. Based on the results, we concluded that research on ChatGPT-based learning in education is still in its early stages, with limited empirical studies addressing effective instructional design and learning strategies.

Keywords

• artificial intelligence
• chatgpt-based learning
• global trends
• systematic review