Use of Generative Artificial Intelligence

In light of the development of generative artificial intelligence (AI) and assistive AI technologies, which are being increasingly used by authors in the creation of scientific manuscripts, the journal establishes requirements governing their use. The editors will closely monitor developments in this field and revise this policy as necessary.

1. Use of AI by Authors

General Principles

Author's Responsibility. Authors always bear full responsibility for the content of their manuscript, including any information generated or processed by AI. The use of any AI tools (including generative language models, image processing models, analytical and optimisation algorithms) does not relieve authors of their responsibility for the accuracy, scientific validity, and ethicality of the data, methods, and conclusions presented. AI tools cannot be listed as authors or co-authors of a manuscript.

Transparency. The use of AI tools must be disclosed in the manuscript, specifying the tool type and purpose in the Methods section (e.g., for data analysis, text writing, image generation, etc.). For each mentioned use, the following must be provided: the name of the AI ​​tool(s) and its version, the tool manufacturer (if applicable), the specific purpose and method of using the AI ​​(e.g., "AI tool X was used for stylistic editing of the text," "Program Y was used for initial categorisation of the data," "ChatGPT-4 was used to generate introduction sentences, which were then refined by the Authors"). Failure to comply with disclosure requirements may be considered a violation of publication ethics.

Verification of AI-generated content is mandatory. It is necessary to ensure that the AI ​​does not generate false or inaccurate information; it is important to check all facts and references.

Data Confidentiality. It is prohibited to upload confidential information, subject data, or any other protected information to publicly available AI tools.

Permitted Uses of AI by Authors

1. Language and Style Enhancement: paraphrasing, readability enhancement, grammar and spell checking (but not generating the main text). Authors may use generative AI and AI tools solely to enhance the readability of their manuscripts. This work must be performed under human supervision, and the resulting AI-generated text must be carefully reviewed and edited by the Authors. It is important to recognise that AI can generate text that appears authoritative but may contain inaccuracies, incomplete information, or biases.
   
2. Information Synthesis and Generalisation: assistance with searching, literature review, and data systematisation.
3. Data Analysis: application of AI algorithms to process and analyse big data, pattern recognition, if the methodology for such analysis is clearly described.
4. Visual Creation: assistance in creating graphs, charts, and images (with mandatory indication that AI was used and confirmation of the originality of the data).

Prohibited Uses of AI by Authors

1. Passive or covert generation of scientific content. It is prohibited to present text, data, analysis, or interpretations generated in whole or in part by AI as original work without explicitly and fully disclosing the role and contributions of the AI.
   
2. Fabrication and falsification of data using AI. Using AI to create fictitious research data, falsify research results, or generate "fake" datasets that are presented as real is strictly prohibited. Any synthetic or simulated data must be clearly labelled, described in detail in the methods, and justified with limitations. If synthetic data is used to train models, this must be disclosed and supported by the availability of source codes/settings for verification.
3. Violation of the confidentiality and security of personal data. Uploading personally identifiable data or other sensitive data to cloud or third-party AI services without the explicit consent of subjects and without ensuring compliance with data protection laws is prohibited. Using AI services with unknown data retention/use policies to process sensitive information is prohibited. Authors should select tools with documented security guarantees and/or use anonymisation supported by a methodology.
4. Copyright infringement and plagiarism using AI. Using AI to automatically compile copyrighted text, graphics, or images without appropriate rights or citation is prohibited. Generating text clearly based on the work of others without citing sources is considered plagiarism. When using AI to paraphrase, organise, or translate the text of other publications, care must be taken to provide correct citation and adherence to fair use rules.
5. Unsubstantiated conclusions and practical recommendations. It is prohibited to publish practical recommendations, diagnostic conclusions, or therapeutic decisions generated entirely by AI without review by qualified specialists and without specifying the limitations.

Violation of these limitations may result in a request for corrections, expression of doubt regarding publication, retraction of the article, prohibition of further manuscript submission, and notification of the Author's institution for investigation. In cases of data falsification, serious breaches of confidentiality, or fraud, the journal reserves the right to refer the materials to the appropriate regulatory authorities.

2. Use of AI by Reviewers

Reviews must remain professional, independent, and based on human expertise. AI tools may be used only as an aid; final judgments and recommendations must be made by a human Reviewer. The Reviewer is responsible for the completeness, accuracy, and ethicality of the review.

Permitted Uses of AI by Reviewers

1. Assistance in refining the wording, style, and structure of the review text (formatting, grammar), provided that the content of the assessment and scientific criticism are formulated by the Reviewer.
2. Rapid search for and summarising of publicly available information (e.g., extracting key facts from non-specialised sources), provided this does not violate the confidentiality of the manuscript.
3. Automatic check of references, citations, and basic formal aspects (e.g., compliance with journal requirements), provided that the results are verified by a human.
4. Use of specialised AI tools developed to assist in methodological evaluation (e.g., statistical verification), only if the tool is certified/clearly described and the results are confirmed by the Reviewer's independent expertise.

Prohibited Uses of AI by Reviewers

1. Manuscripts under review are confidential documents. Uploading these or any parts thereof (text, data, figures, tables, unsolicited supplementary materials) to AI tools is prohibited, as this may violate the Authors' confidentiality and intellectual property rights. This rule also applies to reviews, which may contain confidential information about the manuscript and its Authors.
2. Using generative AI to assist in scientific peer review is not permitted, as the peer review process requires critical thinking and independent assessment, which is beyond the capabilities of AI. Reviewers are fully responsible for the content of their reviews.
3. Using AI tools that store or index uploaded materials and may use them to train models, if this violates confidentiality or copyright.
4. Sharing confidential information about an unpublished manuscript with third parties through AI services.

Violation of this policy (e.g., disclosure of a manuscript through an unverified AI service) is considered a serious breach of confidentiality and may result in disciplinary action by the Editors, including revocation of review privileges and notification of relevant institutions in cases of serious violations.