Early Research Trends on ChatGPT: Insights from Altmetrics and Science Mapping Analysis

Authors

DOI:

https://doi.org/10.3991/ijet.v18i19.41793

Keywords:

large language model, cluster analysis, bibliometrics, social media, fields of research

Abstract


In the four months following its launch in December 2022, ChatGPT, the LLM bot employing deep learning algorithms to generate human-like responses, has been the subject of numerous research articles. Identifying early attention to this research is highly intriguing. As citations for these publications may take time to accumulate, our study focused on examining the early attention of ChatGPT research using the Altmetric Attention Score (AAS), a composite attention score developed by Digital Science. Our findings from the total set of publications and the top publications according to the highest AAS scores reveal the following trends: (i) The United States, Japan, and the United Kingdom are the top countries that published most of the top research articles related to ChatGPT. (ii) The most frequently mentioned source titles include journals like Nature, Science and preprint sources like medRxiv and arXiv. (iii) Among the fields of research (FoR) to which ChatGPT publications align, ‘information and computing sciences’ and ‘biomedical and clinical sciences’ received the highest mentions. (iv) Five major clusters were identified in the network formed by the interlinkage of FoRs, and the most prominent themes discussed in top articles within these five clusters include ChatGPT usage in medical writing and determining ChatGPT’s role in scientific publishing. (v) Scientists are found to be the major user category demonstrating the highest level of interest in ChatGPT research. By capturing these early trends in ChatGPT research and the early attention to this research, our work offers valuable insights for ChatGPT enthusiasts, researchers, and policymakers in fields such as education, information sciences, biomedical sciences, scientific publishing, and many others.

Downloads

Published

2023-10-04

How to Cite

Raman, R., Lathabhai, H., Diwakar, S., & Nedungadi, P. (2023). Early Research Trends on ChatGPT: Insights from Altmetrics and Science Mapping Analysis. International Journal of Emerging Technologies in Learning (iJET), 18(19), pp. 13–31. https://doi.org/10.3991/ijet.v18i19.41793

Issue

Section

Papers