Artificial Intelligence in Academic Publishing: A Comparative Analysis of Author and Reviewer Guidelines Across Major Obstetrics and Gynecology Journals.

The rapid integration of Large Language Models (LLMs) into biomedical publishing necessitates clear ethical frameworks, yet current policies in obstetrics and gynecology remain heterogeneous. This study systematically evaluated the "Instructions for Authors" and ethical guidelines of ten high-impact obstetrics and gynecology journals-representing Elsevier, Wolters Kluwer, Oxford University Press, and Wiley-to identify areas of consensus and critical policy gaps. Data collected through January 4, 2026, revealed that while some journals rely on broad publisher mandates, others maintain specific, divergent guidelines. Universal consensus exists regarding the prohibition of artificial intelligence (AI) authorship, the requirement for full human accountability, and the exclusion of AI from peer review to preserve confidentiality. However, significant variation was identified in disclosure protocols, the permissibility of AI for content drafting versus linguistic editing, and the management of AI-generated visual content. Critical gaps remain regarding standardized taxonomies, bias mitigation, and post-publication accountability. To address these disparities, we propose a unified, tiered framework-categorizing AI use as prohibited, supervised, or allowed-alongside a standardized disclosure statement. Harmonizing these standards is essential to maintain research integrity and reduce author confusion, while preserving scientific rigor and trust in obstetrics and gynecology research.