Integrating artificial intelligence into scientific writing: a narrative review for clinical and surgical researchers

Abstract

Artificial intelligence (AI), particularly large language models, is reshaping academic and clinical research by supporting key stages of the scientific writing process, including idea generation, literature review, drafting, and revision. Tools such as ChatGPT enhance efficiency, clarity, and accessibility while assisting with hypothesis development, evidence synthesis, data analysis, and figure generation. This narrative review examines the benefits, challenges, and ethical considerations of AI-assisted scientific writing, with emphasis on medical and surgical research. While AI tools streamline manuscript preparation when effectively prompted, they also pose risks such as hallucinations, biased outputs, plagiarism, and reduced critical engagement. Concerns have emerged regarding overreliance and diminished scholarly autonomy. Editorial policies increasingly require transparency and human oversight. To ensure responsible use, structured training for researchers and reviewers is essential. AI should be seen as an augmentative tool that complements—not replaces—human expertise, offering opportunities to broaden access to publishing and promote equitable participation in scientific discourse.