Generative artificial intelligence in the preparation of scientific medical articles on surgery: principles of responsible use and current requirements of editorial organisations
- 1O. O. Shalimov National Scientific Centre for Surgery and Transplantology of the National Academy of Medical Sciences of Ukraine, Kyiv
- 2Kharkiv National Medical University
Abstract
Background. The rapid adoption of large language models between 2022 and 2026 has transformed the routine preparation of scientific manuscripts. In surgery, where the author is simultaneously a practising clinician with limited time for analytical work and is often compelled to publish results in English, the temptation to delegate text preparation to generative tools is particularly high. At the same time, leading editorial organisations have restricted the permissible scenarios for the use of artificial intelligence and tightened the requirements for disclosing such use.
Objective. To systematise the current positions of editorial organisations regarding the use of generative artificial intelligence in scientific manuscripts, to outline the functional classes of tools relevant to surgical writing, and to formulate a practical framework for the responsible use of artificial intelligence, suitable for implementation in a surgeon’s daily scientific work.
Materials and methods. A narrative literature review involving a targeted search of the PubMed/MEDLINE, Scopus, Web of Science and Google Scholar databases for the period from January 2022 to July 2026, with restrictions on language (English, Ukrainian) and source type (peer-reviewed articles, official position papers from editorial organisations, and guidance documents on disclosure). The formal PRISMA–ScR protocol was not applied.
Results. The ICMJE, COPE and WAME have agreed on three key positions regarding artificial intelligence: artificial intelligence cannot be an author; human authors bear full responsibility for the accuracy, integrity and originality of scientific manuscripts; and the use of artificial intelligence must be disclosed. The main clinically significant risks are the fabrication of references, factual inaccuracies in a medical context, bias, model drift, lack of reproducibility and breaches of patient data confidentiality. Automated detectors and expert assessment of content generated using artificial intelligence have limited reliability, which highlights the importance of voluntary disclosure of the use of artificial intelligence.
Conclusions. Generative artificial intelligence should be regarded as an editorial and structural tool rather than a source of evidence. Responsible use of artificial intelligence requires mandatory verification of every statement and reference, de-identification of data, compliance with applicable guidelines on reporting, and transparent disclosure in the ‘Methods’ and ‘Acknowledgements’ sections.
Related
- Edible Mushroom Polysaccharides as Functional Food Ingredients for Metabolic Health: Preparation, Techno-Functional Behavior, and Requirements for Translation.
- Balancing innovation and ethics: responsible governance of artificial intelligence hallucinations in healthcare.
- Responsible Use of Artificial Intelligence Tools in Imaging Reporting: Practical Guidelines for Sonographers-A Narrative Review.
- Ensuring Safe and Responsible Use of Artificial Intelligence in Medical Imaging: Pitfalls and Practical Solutions through the Lens of Obstetric Ultrasound.
- Stakeholder-Reported Requirements for Future Artificial Intelligence Decision Support in Periodontal Care: A Cross-Sectional Survey of Dentists, Physicians and Patients in Kazakhstan.
BibTeX
@article{usenko2026generative,
title = {{Generative artificial intelligence in the preparation of scientific medical articles on surgery: principles of responsible use and current requirements of editorial organisations}},
author = {O. Yu. Usenko and I. A. Kryvoruchko and S. A. Andreieshchev and M. V. Kostylev},
journal = {Український журнал клінічної хірургії},
year = {2026},
month = oct,
doi = {10.26779/2786-832x.2026.5.65},
url = {https://doi.org/10.26779/2786-832x.2026.5.65},
note = {CC BY},
}
RIS
TY - JOUR
TI - Generative artificial intelligence in the preparation of scientific medical articles on surgery: principles of responsible use and current requirements of editorial organisations
AU - O. Yu. Usenko
AU - I. A. Kryvoruchko
AU - S. A. Andreieshchev
AU - M. V. Kostylev
PY - 2026
DA - 2026/10/23
JO - Український журнал клінічної хірургії
DO - 10.26779/2786-832x.2026.5.65
UR - https://doi.org/10.26779/2786-832x.2026.5.65
AB - Background. The rapid adoption of large language models between 2022 and 2026 has transformed the routine preparation of scientific manuscripts. In surgery, where the author is simultaneously a practising clinician with limited time for analytical work and is often compelled to publish results in English, the temptation to delegate text preparation to generative tools is particularly high. At the same time, leading editorial organisations have restricted the permissible scenarios for the use of artificial intelligence and tightened the requirements for disclosing such use. Objective. To systematise the current positions of editorial organisations regarding the use of generative artificial intelligence in scientific manuscripts, to outline the functional classes of tools relevant to surgical writing, and to formulate a practical framework for the responsible use of artificial intelligence, suitable for implementation in a surgeon’s daily scientific work. Materials and methods. A narrative literature review involving a targeted search of the PubMed/MEDLINE, Scopus, Web of Science and Google Scholar databases for the period from January 2022 to July 2026, with restrictions on language (English, Ukrainian) and source type (peer-reviewed articles, official position papers from editorial organisations, and guidance documents on disclosure). The formal PRISMA–ScR protocol was not applied. Results. The ICMJE, COPE and WAME have agreed on three key positions regarding artificial intelligence: artificial intelligence cannot be an author; human authors bear full responsibility for the accuracy, integrity and originality of scientific manuscripts; and the use of artificial intelligence must be disclosed. The main clinically significant risks are the fabrication of references, factual inaccuracies in a medical context, bias, model drift, lack of reproducibility and breaches of patient data confidentiality. Automated detectors and expert assessment of content generated using artificial intelligence have limited reliability, which highlights the importance of voluntary disclosure of the use of artificial intelligence. Conclusions. Generative artificial intelligence should be regarded as an editorial and structural tool rather than a source of evidence. Responsible use of artificial intelligence requires mandatory verification of every statement and reference, de-identification of data, compliance with applicable guidelines on reporting, and transparent disclosure in the ‘Methods’ and ‘Acknowledgements’ sections.
SN - 2786-832X
SN - 2786-8311
ER -