Research Trends and Knowledge Structure of Large Language Models in Medical Ethics: A Bibliometric Analysis

Abstract

Background: Large language models (LLMs) are increasingly being applied in healthcare and medical education, raising growing ethical concerns regarding reliability, accountability, bias, and governance. However, the global research landscape and thematic structure of studies addressing LLM-related medical ethics remain insufficiently characterized.

Methods: A bibliometric analysis was conducted using publications indexed in the WoSCC. Relevant studies published up to December 31, 2025 were retrieved and screened according to predefined eligibility criteria. Bibliometric mapping was performed using VOSviewer and the Bibliometrix R package to analyze publication trends, collaboration networks, author keyword co-occurrence, and interdisciplinary citation structures.

Results: A total of 1,604 publications were included in the final analysis. The annual number of publications increased markedly after 2022, coinciding with the rapid expansion of generative AI technologies. Sixty-five countries or regions contributed to this research field, with the United States producing the largest number of publications and occupying the most central position in the collaboration network. Keyword co-occurrence analysis identified five major thematic clusters, including clinical and educational applications of LLMs, ethical governance and regulation, methodological foundations of AI in healthcare, digital mental health applications, and core bioethical principles. Citation flow analysis revealed strong interdisciplinary knowledge exchange between healthcare sciences and computer science, highlighting the hybrid nature of this emerging research domain.

Conclusions: Research on LLMs and medical ethics is expanding rapidly and demonstrates increasing interdisciplinary integration. Future studies should focus on developing robust ethical frameworks, governance strategies, and empirical evaluations to support the responsible implementation of LLM technologies in healthcare.