Ethical concerns related to using ChatGPT in biomedical research: a scoping review

Background: Tools such as Chat Generative Pre-trained Transformer (ChatGPT), based on large language models (LLMs), are rapidly transforming the reporting of biomedical research. Objectives: To examine the ethical challenges surrounding ChatGPT highlight the risk of its misuse, and recommend ways to promote its responsible and ethical use. Methods: A scoping review was conducted using Scopus and PubMed. The inclusion criteria were open-access articles in English published between January 2023 and June 2025. Citation data were collected from the same two databases, and journal metrics were collected from Clarivate’s Journal Citation Reports . The articles were cat-egorized by the ethical issue(s) they focused on, namely bias, accountability, plagia-rism, transparency, and confidentiality. Results: Of the 568 screened articles, 48 met the inclusion criteria and appeared in 43 journals, mainly those published by Springer Nature and Elsevier. Half the articles were narrative reviews and seven were editorials. The leading contributing countries were the United States (nine articles) and India (eight articles). The articles earned 5707 citations, with a 20% trimmed mean of 36.13 ± 40.35 and a median of 19.5. The most cited article had been cited 2516 times. The average journal impact factor of the 43 journals was 4.08 ± 4.4. Among the ethical concerns, roughly three-quarters (38 articles) addressed bias and plagiarism (36 articles); only 10 addressed all five issues. Conclusions: Although ChatGPT offers clear benefits, ethical risks in its use are sig-nificant, and clear guidelines are essential to ensure that the tool is used safely and responsibly.