Current Guidelines on the Use of Generative Artificial Intelligence in Peer-Reviewed Scholarly Publications

Abstract

Generative artificial intelligence (AI) tools, especially large language models, such as Chat Generative Pre-Trained Transformer (GPT), Gemini, and Copilot, have been quickly reshaping many aspects of scholarly publishing. Generative AI tools hold great potential in enhancing the ability of authors and medical writers as well as editors in tasks such as gathering insight, summarizing information, generating drafts, and refining text. However, AI technology also carries significant ethical concerns regarding the accountability of authors, biases, and inaccuracies introduced in research and writing, as well as unintentional plagiarism. To understand the landscape of generative AI use and evolving guidelines in medical publications, we reviewed published literature and summarized recent guidelines, statements, or recommendations from select groups of medical journal editors in scholarly publications of medical sciences. Common themes, differences, current limitations, and areas for future development are discussed.