Incorporating Artificial Intelligence in Medical Research Publications: Boon or Bane?

Abstract

A BSTRACT Background: Large language models and other artificial intelligence (AI) tools like ChatGPT are increasingly used by researchers to prepare research publications. This narrative review synthesizes recent (2023–2025) evidence and commentary across medical specialities on the benefits and risks of AI use in manuscript preparation and publication. Methods: A focused search of PubMed/PMC for editorials, reviews, surveys, and observational analyses published between January 2023 and September 2025 was conducted. Selected articles addressing either supportive roles or concerns were tabulated and synthesized. Results: Seventeen key publications (editorials, reviews, surveys, and specialty analyses) were selected. Benefits of AI from data synthesis include efficiency, improved clarity for non-native English speakers, and assistance with literature synthesis. Concerns include hallucinations, fabricated references, authorship ambiguity, and risks to research integrity. Conclusions: AI is a potent augmentation tool but requires transparent disclosure, human oversight, and explicit journal policies to mitigate risks. The most effective workflow is the human-AI-human sandwich, where humans define the hypothesis, select the data, and outline the argument; AI then expands the outline, checks the grammar, and suggests flow improvements; humans again verify every claim, check every reference, and infuse the final text with nuance and clinical judgment.

Ähnliche Arbeiten