Artificial Intelligence use policy

1. Introduction
This document defines the principles, ethical principles and technical rules for the use of tools based on artificial intelligence (AI), in particular large language models (LLM), generative algorithms and automated translation/editing tools in the process of preparing, reviewing and publishing materials in the publication "In the World of Mathematics" (hereinafter referred to as the Journal).
The editorial board of the Journal recognizes that AI can be a useful tool for visualizing complex mathematical concepts, correcting language or structuring data. However, given the specifics of the publication, we emphasize: AI is not a subject of scientific activity and cannot be responsible for the reliability of mathematical proofs or the authorship of ideas. The use of AI must be transparent, ethical and subject to the requirements of academic integrity.
2. Scope
This policy is mandatory for all participants in the editorial and publishing cycle:
Authors: when writing articles, creating illustrations, graphs or program code to solve mathematical problems.
Reviewers: when evaluating manuscripts (with strict confidentiality).
The Editorial Board: when making decisions about publication, technical editing and layout of the Journal.
3. Use of AI tools by authors
Authors have the right to use artificial intelligence tools to improve their manuscripts, but such use is strictly regulated by the following conditions.
Personal responsibility for content. Authors bear full sole responsibility for the accuracy, originality and reliability of all content of their manuscript. This applies to mathematical statements, theorem proofs, calculations and interpretations, even if they were created, checked or edited using AI. The use of technology does not exempt the author from liability for plagiarism, fabrication or falsification of data, as well as for any other violations of academic integrity.
Transparency and disclosure. Authors are required to clearly declare the fact of using AI. The relevant information should be indicated in the "Acknowledgements" section or in a separate subsection "Declaration on the use of AI tools". It is necessary to indicate:
the name of the tool and its version (for example, ChatGPT-4, Claude 3, DeepL, etc.).
the specific purpose of use (for example, stylistic editing of text, grammar checking, writing software code for mathematical modeling, generation of illustrative drawings).
Prohibition of AI authorship. No artificial intelligence tool can be indicated as an author or co-author of a manuscript. According to international standards, authorship implies the ability to bear responsibility for work that algorithms do not own. Authorship belongs exclusively to individuals who have made a significant intellectual contribution to the research.
Verification of generated content. Mathematical text requires absolute accuracy. Authors should carefully check each step of reasoning, formula, or numerical result proposed by AI for logical errors (“hallucinations” of AI), relevance to the context, and scientific correctness.
Use for data analysis and calculations. In the case of using AI for statistical analysis, processing of large data sets, or symbolic calculations, authors should describe the methodology in detail in a separate paragraph of the article that characterizes the research tools and methods (or in the corresponding description of the algorithm for solving the problem), noting the role of AI in obtaining the final result.
Generation of visual materials. If AI was used to create graphs, geometric models, or model illustrations for mathematical plots, this should be reflected:
directly in the caption to the figure (for example: “Image generated using [Name of AI]”);
in a general paragraph about the use of AI at the end of the article.
5. Use of AI tools by the editorial board and staff
The editorial board and technical staff of the Journal have the right to use artificial intelligence tools exclusively for the optimization of technical and administrative processes such as
Plagiarism checks. Use of specialized services to detect text matches and incorrect borrowings.
Technical control. Checking the compliance of the design of references, metadata and manuscript structure with the requirements of the Journal.
Search for experts. Assistance in the selection of potential reviewers based on the analysis of their previous publications in relevant fields of mathematics.
Analytics. Monitoring of current scientific trends and the popularity of topics among the readership.
Confidentiality and Security. When using any third-party AI services, the Journal staff is required to strictly adhere to the non-disclosure policy. It is prohibited to upload full texts of unpublished manuscripts into open generative models (for example, for automatic annotation creation), as this creates a risk of leakage of authors' intellectual property.

Avoidance of bias. The editorial board is aware that AI algorithms may have built-in biases. Therefore, the results of AI tools are considered only as auxiliary data that require critical human evaluation.

Priority of human decision. The final decision to accept, send for revision or reject a manuscript is made exclusively by the members of the editorial board. No automated report (for example, the percentage of text uniqueness or automatic article rating) can be the only reason for refusing publication without a substantive analysis of the scientific value of the work by specialists.

Communication ethics. When using AI for mailings or technical responses to authors, the editorial board ensures control over the correctness and personalization of such messages.

6. Ethical aspects and responsibility

The editorial board of the Journal emphasizes that the use of artificial intelligence technologies does not change the fundamental principles of scientific ethics, but only creates new contexts for their application.

Counteraction to plagiarism and borrowing. The use of text, mathematical proofs, algorithms or images generated by AI tools without proper declaration and significant intellectual processing by the author is regarded as hidden borrowing (plagiarism). Authors should be aware that AI models are trained on existing texts, so the content they generate may contain fragments of other people's works without references to the original sources.

Critical evaluation and bias. AI tools can create content that contains logical errors, outdated data or hidden biases inherent in the data sets on which they were trained. Authors, reviewers, and editorial board members are required to critically evaluate the results of AI work, verify the mathematical correctness of conclusions, and avoid disseminating unreliable or biased content.

The principle of maximum transparency. Transparency regarding the methods of preparing the material is a key element of responsible scientific communication. Concealing the fact of using AI is unacceptable and may be grounds for refusal to publish or retraction (withdrawal) of an already published article.

Distribution of responsibility. The ultimate responsibility for compliance with ethical norms, scientific reliability, and compliance with the standards of the Journal lies with the authors, reviewers, and members of the editorial board. Moreover, the authors are responsible for the originality and accuracy of the submitted materials; the reviewers are responsible for the objectivity of the examination and maintaining confidentiality; the editorial board members are responsible for the final verification of compliance with this Policy and making unbiased decisions.

7. Review of violations and sanctions

The Editorial Board of the Journal reserves the right to conduct an investigation in case of suspicion of improper or unethical use of artificial intelligence tools. Any detected cases (for example, hidden use of AI, improper declaration or generation of erroneous mathematical data) are considered by the Editorial Board in accordance with the established procedures of the Journal and the recommendations of COPE (Committee on Publication Ethics).

Depending on the severity of the violation, the Editorial Board may take the following measures: sending a formal request to the author or reviewer for clarification; requiring the provision of source data, program code or drafts confirming the independent performance of the work; rejecting the manuscript at any stage of consideration; retracting an already published article with an official notification of the reasons on the Journal website; temporarily or permanently banning the author from submitting new materials for publication.

8. Policy Updates and Adaptability

Generative AI technologies are evolving rapidly, so this Policy is not static. This Policy may be reviewed and updated by the Journal’s editorial board to reflect developments in AI technologies and changing standards in the scientific community.

All changes to the Policy will be published on the Journal’s official resources to inform the scientific community and ensure full transparency of the editorial process.