Ethical and responsible use of artificial intelligence in medical education

Artificial intelligence (AI) has rapidly transformed medical education through virtual patient simulations, adaptive learning systems, and large language models like ChatGPT, while raising concerns about bias, transparency, reproducibility, and ethical integration into training and assessment. Despite its growing relevance, the ethical and responsible use of AI in medical education remains understudied. This bibliometric analysis aims to characterize the intellectual landscape of this field.

Bibliographic data were obtained from the Web of Science Core Collection using a search strategy combining medical education and AI-related terms, limited to publications from 1995 to 2026. A total of 1403 articles were retrieved. Microsoft Excel was used to generate descriptive figures, and VOSviewer (version 1.6.20) was used to construct bibliometric networks of country, institutional, and keyword co-occurrence relationships.

Publication output rose sharply after 2023, from 113 publications to 500 by 2025, coinciding with the rise of ChatGPT and other large language models. The United States led in publication volume (533) and citation influence, followed by China (189) and England (95). Harvard University and its affiliated entities led institutional output, while Harvard Medical School, Stanford, and University of California San Francisco led citation strength. The most-cited publications centered on ChatGPT's performance on medical licensing examinations.

This field has grown rapidly but remains concentrated among few countries and institutions, with citation impact dominated by capability-testing studies rather than governance-focused scholarship. Future research should prioritize empirical evaluation of AI governance frameworks and broader institutional and geographic representation.