Bulletin of National Defence University of Ukraine

The Editorial Office recognises that generative artificial intelligence (AI), large language models (LLMs), and other AI-based tools may serve as useful supporting tools in the preparation, editing, and processing of scholarly materials. At the same time, their use must comply with the principles of transparency, academic integrity, confidentiality, authorial responsibility, and the protection of intellectual property.

This policy applies to authors, reviewers, editors, and other individuals involved in the editorial and publishing process.

Use of AI by Authors

Generative AI may be used by authors only as a supporting tool and must not replace the author’s original intellectual and scholarly contribution.

The use of AI is permitted for:

• language editing, grammar and spelling checks, and stylistic improvement of text written by the authors;
• improving the clarity, readability, and structure of text written by the authors;
• translation, provided that the result is checked by the author;
• structuring and summarising drafts prepared by the authors;
• technical formatting of materials;
• assistance with formatting bibliographic references, provided that every source is verified against the original source;
• assistance with writing computer code, processing, or analysing data, provided that such use is appropriately described in the research methodology and allows the relevant procedure to be assessed or reproduced.

The use of AI must not:

• replace the author’s original scholarly contribution;
• independently generate scientific hypotheses, interpretations of results, arguments, or conclusions that are presented as the product of the authors’ intellectual work;
• generate all or a substantial part of a scholarly text on behalf of an author without appropriate disclosure of such use;
• generate, alter, or substitute actual research results;
• create fabricated data, statistical indicators, sources, quotations, DOIs, bibliographic references, or other unverified claims;
• be used to conceal plagiarism, fabrication, falsification of data, or other breaches of academic integrity.

Authors are required to independently verify the accuracy, reliability, originality, and consistency with the cited sources of all material prepared or processed with the assistance of AI.

AI systems, chatbots, and other automated tools must not be listed as authors or co-authors of a manuscript because they cannot meet authorship criteria, assume responsibility for the content of the work, declare conflicts of interest, or provide consent for publication.

Disclosure of AI Use

Authors must transparently disclose the use of generative AI during the preparation of a manuscript.

Relevant information should be provided in the Acknowledgements, Materials and Methods, or another appropriate section, depending on the nature of the AI use.

The disclosure must specify:

• the name of the tool used;
• the developer or provider of the tool;
• the model or version, where this information is available;
• the purpose of its use;
• the nature and extent of its use;
• the part of the manuscript or research process in which AI was used.

If generative AI was used to create or substantially transform text, computer code, tables, figures, or other materials, the authors must also provide the prompt used or a sufficiently detailed description of the prompts and procedures applied.

If AI was used directly as part of the research methodology, for data analysis, generation of results, creation of computer code, or other procedures that affect the research findings, sufficient information must be provided in the Materials and Methods section to enable scientific scrutiny and, as far as possible, reproduction of the procedure. This information should include the name and version of the system, the date on which it was used, the principal parameters, and the full prompts that directly influenced the results obtained.

Disclosure of AI use does not relieve authors of responsibility for the content of the manuscript.

Use of AI for Images, Figures, Graphs, and Tables

The use of generative AI to create or edit visual materials must be transparent and must not distort the research findings.

Generative AI must not be used to create, add, remove, move, conceal, or alter elements of primary research images in a way that may change their scientific meaning. This applies, in particular, to micrographs, medical and laboratory images, photographs of experimental objects, images of gels, and other materials constituting primary research data.

An exception may be made where the use of generative AI is itself the subject or method of the research. In such cases, the procedure must be described in detail in the Materials and Methods section.

AI may be used to create diagrams, conceptual illustrations, and other supporting visual materials that are not presented as primary research data, provided that:

• their content is verified by the authors;
• no copyright is infringed;
• the use of AI is clearly disclosed;
• the tool used is identified in the caption to the relevant figure or in another location specified by the journal.

Graphs, charts, and tables created or formatted with the assistance of AI must be based exclusively on genuine and verified data. The generation of missing values or their arbitrary ‘completion’ by AI is not permitted.

Authors’ Responsibilities

Regardless of how AI is used, authors bear full responsibility for the entire content of the submitted manuscript.

Authors must ensure that:

• all factual claims are accurate and verified;
• all cited sources genuinely exist and accurately support the information for which they are cited;
• the use of AI has not resulted in the fabrication or falsification of data;
• the material does not contain plagiarism;
• the use of AI does not infringe copyright, intellectual property rights, confidentiality, or the rights of third parties;
• personal, confidential, or otherwise protected data have not been submitted to external AI systems in breach of applicable law, ethical requirements, or the consent obtained.

AI-generated material must not be used as a primary source of scientific evidence in place of verified and appropriately cited original sources.

Use of AI by Reviewers

Manuscripts submitted to the journal are confidential materials. Reviewers are required to maintain their confidentiality at all stages of the peer-review process.

Reviewers must not upload the full text of a manuscript or any part of it, unpublished data, figures, tables, supplementary materials, or other confidential information to publicly accessible or external generative AI systems unless the journal has expressly authorised the use of a specific secure tool and an appropriate level of confidentiality and data protection is ensured.

If a reviewer intends to use AI and is uncertain whether a particular tool complies with the journal’s policy, they must consult the Editorial Office in advance.

AI must not replace the reviewer’s expert judgement. Reviewers remain personally responsible for the content, reasoning, accuracy, and integrity of their reviews.

Generative AI must not be tasked with independently assessing the scientific novelty, reliability of results, or suitability of a manuscript for publication in place of the reviewer.

If an AI tool authorised by the journal has been used substantially in preparing a review, the reviewer must inform the Editorial Office of the name of the tool and the nature of its use.

Use of AI by Editors and Editorial Staff

Editors and editorial staff are likewise required to maintain the confidentiality of unpublished manuscripts.

Editors must not upload submitted manuscripts, parts thereof, figures, tables, personal data relating to authors or reviewers, or other confidential information to external generative AI systems unless appropriate confidentiality, data protection, and intellectual property safeguards can be guaranteed.

The journal may use automated and AI-based tools approved by the Editorial Office for supporting technical tasks, including:

• similarity checking;
• technical and formal checks of manuscripts;
• language and stylistic checks;
• manuscript classification;
• metadata verification;
• other supporting editorial and publishing operations.

Such tools may be used only where confidential information is appropriately protected and where submitted materials are not used for unauthorised model training, reuse, or disclosure to third parties.

AI must not independently make editorial decisions or serve as the sole basis for rejecting, accepting, or requesting revision of a manuscript. Final decisions regarding manuscripts are made exclusively by editors on the basis of expert assessment and, where required by the journal’s procedures, the results of independent peer review.

The results of automated systems designed to detect text that may have been generated by AI are not regarded as sufficient evidence of a breach of academic integrity and must not serve as the sole basis for editorial sanctions.

Any substantial use of generative AI by an editor or member of the editorial staff in connection with a specific manuscript must be documented by the Editorial Office. Where such use goes beyond standard technical processing of the material, the authors must be appropriately informed.

Confidentiality and Data Protection

All participants in the editorial and publishing process should be aware that information entered into external generative AI systems may, depending on the terms and conditions of the relevant service, be stored, processed, or used for other purposes.

Accordingly, unpublished manuscripts, confidential research findings, personal data, peer-review reports, and other protected information must not be submitted to external AI systems without appropriate safeguards for confidentiality, data protection, and intellectual property rights.

This requirement applies equally to authors, reviewers, editors, and any other individuals who have access to unpublished materials.

Undisclosed or Inappropriate Use of AI

The concealment of substantial use of generative AI, the fabrication or falsification of materials with its assistance, the creation of fabricated sources or data, breaches of confidentiality or copyright, or other forms of improper use of AI will be regarded by the Editorial Office as a potential breach of publication ethics.

Depending on the nature and seriousness of the breach, the Editorial Office may:

• request explanations or additional information from the authors;
• require correction or resubmission of the manuscript;
• reject the manuscript;
• following publication, issue a correction, an editorial notice, or retract the article, as appropriate to the nature of the established breach;
• where appropriate, contact the author’s institution or another competent authority.

Each case will be considered individually, taking into account the nature of the AI use, its impact on the scientific reliability of the material, and the principles of due process.

International Standards and Recommendations

This policy has been developed with reference to current international recommendations on the use of artificial intelligence in scholarly publishing and the editorial and publishing process, including those of the Committee on Publication Ethics (COPE) and the World Association of Medical Editors (WAME).