Artificial Intelligence in Scientific Manuscript Preparation in Medical Science: A Scoping Review of Applications, Ethical Considerations, and Emerging Practices

A

BSTRACT

Artificial intelligence (AI) has rapidly entered scientific publishing workflows, including biomedical manuscript preparation. In Medical Science , where published evidence may directly influence clinical decision-making for both mother and fetus, the implications of AI-assisted writing warrant careful scrutiny. Although AI tools can support literature exploration, drafting, and language refinement, concerns persist regarding fabricated references, incomplete disclosure, and potential shifts in authorship responsibility. Existing evidence remains heterogeneous and dispersed across empirical and nonempirical sources. This study was conducted as a scoping review guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 recommendations to map the range and characteristics of the available literature. Searches of PubMed/MEDLINE, Embase, Scopus, Web of Science, the Cochrane Library, and PubMed Central identified relevant studies. Following screening of 850 records, 43 studies were included in a qualitative synthesis. Given the diversity of study designs, a structured qualitative methodological appraisal was applied rather than formal risk-of-bias tools. AI applications were most frequently reported in literature discovery, drafting support, and language editing. However, recurrent concerns included hallucinated citations, variable accuracy, and inconsistent transparency in AI use. AI may enhance efficiency in manuscript preparation, but current evidence, largely nonempirical, remains limited. Findings should be interpreted cautiously and are intended to inform, rather than prescribe, responsible integration of AI in medical scientific writing.