This study reviews AI-related publication policies from the top 10 academic publishers and identifies major inconsistencies in how generative AI and AI-assisted tools are regulated. Publishers generally accept AI-assisted tools for grammar, editing, and formatting, but impose strict limitations on generative AI due to risks such as bias, hallucinations, copyright concerns, and research integrity issues. All publishers agree that AI cannot be listed as an author, and human researchers remain fully responsible for the accuracy, originality, and ethical compliance of manuscripts. The paper proposes consolidated guidelines distinguishing recognized versus unrecognized uses of AI in scholarly publishing. It concludes that unified ethical standards are urgently needed as AI tools continue to evolve rapidly.

Link to the article: https://escienceediting.org/journal/view.php?number=358