Scientific Journal ISSN 1991-0533
Use of Artificial Intelligence
The Editorial Board considers a scholarly publication to be an independent intellectual result produced by its authors. This policy has been developed with reference to the recommendations and editorial practices of the International Committee of Medical Journal Editors (ICMJE), World Association of Medical Editors (WAME), Elsevier, JAMA Network, and Nature Portfolio. The journal, however, applies stricter restrictions on the use of generative artificial intelligence in the creation of author-generated scholarly text.
The use of generative artificial intelligence and large language models to create the scholarly content of a manuscript is not permitted.
Artificial intelligence systems may not be listed as authors or co-authors. Authorship requires personal intellectual contribution and responsibility for the formulation of the research problem, selection and justification of the methodology, reliability of data and results, interpretation of findings, and the conclusions presented in the manuscript. The principle of human accountability and the ineligibility of AI systems for authorship are reflected, in particular, in the ICMJE Recommendations.
Prohibited Uses of Artificial Intelligence
Generative artificial intelligence must not be used for:
- writing a complete manuscript or any part of it;
- generating, substantively paraphrasing, shortening, expanding, or stylistically rewriting scholarly text;
- preparing abstracts, introductions, discussions, conclusions, or other substantive sections of the manuscript;
- formulating the research problem, aims, research questions, or hypotheses;
- selecting or justifying the research methodology on behalf of the authors;
- preparing literature reviews, conducting substantive analysis of scholarly publications, or identifying research gaps;
- selecting scholarly sources or generating and formatting bibliographic references;
- interpreting research results or formulating scientific conclusions;
- creating data, facts, research results, or other information presented as genuinely obtained;
- creating or modifying scientific images, graphs, diagrams, or other materials in a manner that may alter, supplement, or misrepresent actual research results;
- preparing responses to reviewers or substantive justifications for revisions to a manuscript;
- translating a manuscript or its substantive sections using generative artificial intelligence.
Standard software used solely for spelling, punctuation, and technical formatting checks, provided that it does not generate new content or alter the meaning of the text, is not covered by this restriction.
Bibliographic References
Generative artificial intelligence must not be used to search for, create, generate, or format bibliographic references. All cited sources must exist and must be personally verified by the authors against the original sources. Authors are responsible for the accuracy of bibliographic descriptions and for ensuring that cited statements correspond to the actual content of the source.
This approach takes account of the current guidance of JAMA Network, which specifically highlights the risk of inaccurate and nonexistent references generated by AI tools and advises against using AI to generate or format bibliographic references.
Use of Artificial Intelligence in Research and Computational Procedures
Artificial intelligence technologies may be used only as a research or computational tool when their use is justified by the objectives of the study and does not replace the authors’ independent intellectual contribution.
Permissible uses may include:
- technical processing, structuring, and aggregation of large datasets;
- automation of large-scale or repetitive computational procedures;
- computational operations involving substantial volumes of information;
- creation or adaptation of program code used exclusively for data processing and calculations, provided that the code is fully reviewed and verified by the authors;
- use of artificial intelligence as the direct object or research instrument of a study when justified by its methodology.
Such use is permitted only when all of the following conditions are met:
- the use of the tool is methodologically justified;
- the source data are lawful and verifiable;
- artificial intelligence does not replace the authors’ substantive interpretation of the results;
- the authors control the rules applied to data cleaning, merging, classification, and transformation;
- AI-assisted results are checked by the authors and, where possible, confirmed by independent calculation or another verification procedure;
- the procedure is described in sufficient detail to allow its reproducibility to be assessed;
- the authors accept full responsibility for the accuracy of source data, algorithms, program code, calculations, and results.
Automatic generation of missing observations, alteration of source data, substantive classification of research objects, or other operations capable of affecting research results are not regarded as purely technical data processing. Such procedures require separate methodological justification and full disclosure in the manuscript.
The distinction between the use of artificial intelligence in manuscript preparation and its use as part of a research procedure is also reflected in the policies of Elsevier and JAMA Network. The requirements of this journal regarding author-generated scholarly text are, however, more restrictive.
Disclosure of Artificial Intelligence Use
Any use of artificial intelligence directly related to the research procedure, data processing or analysis, programming, or computations must be disclosed by the authors. The principle of transparency in the use of AI is also reflected in the recommendations of ICMJE and WAME.
The Materials and Methods section must specify:
- the name of the tool or system used;
- the version of the tool, where available;
- the purpose for which it was used;
- the nature of the operations performed;
- the nature and origin of the data processed;
- material parameters, algorithms, or instructions capable of affecting the results;
- the procedure used to verify and validate the results obtained.
Information about the use of artificial intelligence must also be disclosed in the accompanying consent letter submitted with the manuscript.
Where artificial intelligence technologies constitute the object of the study itself, their use must be reported in accordance with general requirements for the reproducibility of research methodology.
Author Responsibility
Authors bear full responsibility for the manuscript, source data, calculations, program code, tables, figures, bibliographic references, research results, and conclusions, irrespective of the technical tools used.
Failure to disclose the use of artificial intelligence in research or computational procedures, presentation of AI-generated text as independent author-generated scholarly content, or use of fabricated references, data, or results may constitute grounds for rejection of the manuscript.
If such a violation is identified after publication, the Editorial Board will consider correction or retraction of the article in accordance with the journal’s Corrections and Retractions Policy.
Use of Artificial Intelligence in Peer Review and Editorial Evaluation
A manuscript submitted to the journal is confidential. Reviewers and members of the Editorial Board must not upload the full manuscript, any part of it, unpublished data, peer-review reports, or editorial correspondence to external generative artificial intelligence systems.
This restriction is intended to protect the confidentiality of unpublished materials and is consistent with the principles reflected in the policies of ICMJE, Elsevier, and JAMA Network.
Generative artificial intelligence must not be used to prepare peer-review reports, perform the scholarly evaluation of a manuscript, make editorial decisions, or formulate substantive recommendations to authors. Scientific evaluation and editorial decisions must be made by humans.
The use of specialised software for technical manuscript checks, similarity screening, and other supporting editorial procedures is not considered a delegation of scholarly peer review to artificial intelligence.