Scientific journal
Fundamental research
ISSN 1812-7339
"Перечень" ВАК
ИФ РИНЦ = 1,798
Generative ai use policy
The policy is effective from the moment of publication on the journal's website and is regularly reviewed as technology and industry standards evolve. Changes apply to new manuscript submissions; for already published materials, post-publication notification procedures are used.
Purpose and scope
The policy establishes rules for the use of generative AI tools (services and models that create or transform text, images, data, code) by authors, reviewers, and editors at all stages: manuscript preparation, peer review, editorial processing, and post-publication support. The policy works in conjunction with the journal's policies on authorship, research ethics, data and images, conflicts of interest, and retractions.
General principles
- AI is not an author. Decisions are made by people; all responsibility lies with the authors/editors/reviewers.
- Any use of AI must be disclosed.
- Manuscripts, reviews, non-public data/images, and personal information are not uploaded to public AI services.
- Results obtained using AI must be verifiable; a trace of use (prompts/scripts/parameters) is retained and provided to the editorial board upon request.
- Security, data protection, copyright, and related rights standards are observed.
- Automatic "AI detectors" are used only for screening; final decisions are made only on the basis of expert evaluation.
*"Public service" refers to an external online platform where the provider can store/use the data entered. "Local/corporate" solutions are managed by the authors/journal organization and exclude the transfer of data to third parties.
Policy for authors
Authors are permitted to use AI as an auxiliary tool, provided that each result is transparently disclosed and subsequently verified by the authors.
Machine translation, grammar and style editing, simplification of wording, and paraphrasing without changing the scientific meaning are permitted. Responsibility for accuracy and correctness remains with the authors; the final text is checked manually. AI can be asked to suggest an article outline, section headings, and wording for the abstract. These suggestions are considered drafts, which the authors edit and approve themselves. It is permitted to obtain code templates, data preprocessing scripts, and sample commands for analysis or visualization, provided that their correctness is verified, modifications are documented, and reproducibility is ensured (by fixing library versions, parameters, etc.). Any analysis results obtained using AI must be repeated and verified by the authors. It is permissible to use AI to summarize your own protocols, notes, and experiment logs, and to generate explanatory text for tables and graphs that have already been obtained, provided that the text is checked against the primary data. It is permitted to use AI to bring already found and verified sources into the required citation style (formatting of surnames, titles, volumes, pages). Each reference must be manually verified against the original source, DOI (and EDN, if available). Conceptual diagrams, block diagrams, and pictograms created by AI may only be used with the prior written consent of the editorial board and marked "image created by AI." Such images do not replace graphic material with factual data.
AI cannot be used to substitute scientific results, violate confidentiality, or mislead the reader. It is prohibited to generate data and text, retouch or artificially "clean" images, fill in gaps in tables and time series, or issue AI-generated calculations and conclusions. AI images cannot be used as images reflecting real results (photographs, spectra, measurement graphs, etc.). Any post-processing of images is only permitted within the limits of standard linear operations and must be described; it is not permitted to delete/add objects, change content, or scale data. It is prohibited to upload manuscripts, reviews, non-public data, images of research participants, personal or sensitive information, or trade secrets to public AI services. It is not permitted to automatically "generate" bibliographies, citations, and DOIs without subsequent manual verification of each entry against the original source. It is unacceptable to include non-existent or unread works, as well as to adjust references "for volume." It is prohibited to include in the article text, code, images, tables, and other materials generated by AI based on protected content arrays if the authors do not have the rights to use them. Any borrowings must be formatted according to the rules, with the permission of the copyright holders if necessary. It is unacceptable not to disclose the use of AI, to conceal its role in the analysis/text/code, or to indicate AI as a "co-author" or "person in charge." Code/data generated by AI cannot be included in open materials if their distribution violates the licenses of the source components or the terms of the AI service provider; authors are required to check licensing issues in advance.
Disclosure of use
Disclosure is made in both the cover letter and the manuscript (section "Methods") and includes: purpose (translation, language editing, etc.), name of the tool and version/model, type of service (local/corporate/public), brief description of human verification. Authors confirm that the rights of third parties and data processing regimes have not been violated. If AI influenced the results/conclusions, prompts/scripts, model versions, parameters, environment description, and logs sufficient for reproducibility are included in the Supplementary Materials.
Template wording: "Generative AI tools were used: [name, version/model] for [purpose]. All results have been verified by the authors; no confidential data was transferred to public services. Prompts/scripts and parameters are provided in the Supplementary Materials."
Policy for reviewers
The manuscript and all accompanying review materials are confidential. Reviewers are not permitted to transfer them to AI services or upload excerpts from the manuscript, illustrations, tables, unpublished data, or parts of the review to such services. Any drafts and prompts from which the content of the manuscript can be reconstructed are also considered confidential materials and may not be uploaded to external systems.
Preparing an expert review is human work; the use of generative AI for content analysis of the manuscript, formulation of conclusions, and recommendations is not permitted. The reviewer is fully responsible for the accuracy, validity, and originality of the review.
Policy for editors
The editorial office may use AI tools only for technical tasks in a controlled environment: automated verification of references and DOIs, extraction and normalization of metadata, format compliance checks, and identification of obvious duplications of text fragments or inconsistencies in formatting. Such operations are carried out on internal or contractually protected systems that exclude the use of uploaded materials by the provider for its own purposes. Before launching any tool, the editorial board assesses its legal and technical risks (license and data processing terms, server location, storage policy, output formats, and logs) and identifies those responsible for its use.
AI is not used for scientific evaluation of manuscript content and does not influence final editorial decisions. Automatic "AI text/image detectors" are only permitted as preliminary screening tools. Any results are considered grounds for expert review, but not as independent evidence. Decisions on manuscripts are made only after human evaluation of all materials: authors' explanations, source data/images, reviews, and, if necessary, additional expertise.
Confidential materials (manuscripts, reviews, unpublished data and images, personal and sensitive information) are not transferred to external public services. Editorial staff undergo training on confidentiality and the limited use of AI tools; access to such tools is distributed according to roles and the principle of minimum necessity.