Unveiling the Shift: Trends in Medical Literature: A Comparative Bibliometric Analysis of ChatGPT vs. Traditional Methods.

Background: In recent years, the integration of Artificial Intelligence (AI) into various fields has led to transformative changes in research and practice. One such domain that has seen substantial impact is the medical field, where AI-powered tools like ChatGPT have emerged as potential alternatives to traditional methods for generating medical literature. This study presents a comprehensive comparative bibliometric analysis of medical literature generated by ChatGPT and traditional methods, aiming to uncover the emerging trends and potential shifts in the landscape. Materials and Methods: This study collects and analyzes a substantial corpus of articles, reviews, and papers generated by both ChatGPT and traditional methodologies. Bibliometric indicators such as publication frequency, citation counts, collaboration patterns, and keyword usage are examined to discern differences in output and impact. We extract data from Web of Science citation database and selected 18087 publications from the year 2019 to 2023 for our study. The data and descriptive analysis were categorised, collected one at a time, and imported into the Bibliometric R-package programme to produce science maps and statistical graphs. They were exported to MS-Excel for bibliometric analysis and VOSviewer software was used to analyse Co-Occurrence networks. Results and Discussion: A total of 18087 publications on ChatGPT and traditional methodologies from the year 2019 to 2023, namely 12519 (69.29%) original articles, 2836 (15%) reviews, 233 (01.2%) letters, and others. The most productive institution was found to be the Indian Institute of Technology System IIT System (n=1718, 0.09%), followed by National Institute of Technology NIT System (n=1275, 0.07%). the most productive author was found to be the Kumar, Atul, All India Institute of Medical Sciences (AIIMS) New Delhi (n=421, 2.39%), followed by Kumar, Satish, Indian Institute of Management Nagpur (n=405, 2.24%). The most productive journal was found that the IEEE Acess (n=373, 2.063%, TC=6740, ACP=18.06) followed by Multimedia Tools and Applications (n=279, 1.543%, TC=1552, ACP=5.56). The most frequent of authors keywords and occurrences was found that the 'artificial intelligence' 1517 occurrences and 1898 total link strength followed by 'deep learning' 1156 occurrences and 1764 total link strength. Conclusions: This bibliometric analysis sheds light on the evolving landscape of medical literature production, comparing the outputs of ChatGPT and traditional methods. While ChatGPT shows promise in its ability to quickly generate content on cutting-edge topics, traditional methods maintain their dominance in terms of research depth and impact. The findings have implications for researchers, clinicians, and policymakers, suggesting potential ways to leverage both approaches for a more comprehensive and impactful medical research ecosystem. Further research is warranted to monitor the trajectory of this evolving paradigm shift in medical literature and its long-term implications. [ABSTRACT FROM AUTHOR]