Policy on the Use of Generative Artificial Intelligence

Safety Engineering of Anthropogenic Objects permits the use of generative artificial intelligence (GenAI) tools in scientific research and in the preparation of publications, subject to the principles set out in this policy. The use of such tools does not relieve authors of responsibility for the content of their work and must not replace their own analysis, interpretation of results or scientific reasoning.

Rules for Authors

Authors may use GenAI tools to assist in preparing a manuscript, including organising content, searching for literature, and improving language and readability.

Content produced with the assistance of AI must be verified by the author. This applies in particular to factual information, data, citations and references. Sources suggested by an AI tool must be checked against the original publications. When using AI tools, authors should also take into account data protection, confidentiality, intellectual property rights and the terms of use of the tool concerned.

AI must not replace the author's scientific contribution and cannot be listed as an author or co-author.

The use of generative AI to prepare or substantially modify manuscript content should be disclosed by the author. This does not apply to tools used solely for basic spelling, grammar or punctuation checks. The disclosure should identify the tool used and briefly describe the purpose and extent of its use.

If AI is used as part of the research method, for example for data analysis, creating or modifying code, or as the subject of the research itself, its use should be described in the methodology section.

Authors are responsible for images, figures, data and other materials prepared or modified using AI. AI must not be used to fabricate or alter research data or results, or to create or modify images that constitute primary research material. If AI is used to prepare diagrams or explanatory illustrations, its use should be disclosed, including in the caption of the relevant illustration. The use of AI to create data visualisations should be described in the methodology section. Such visualisations must be derived directly from actual data and produced using a reproducible method.

Rules for Reviewers

A manuscript submitted for review is a confidential document. Reviewers must not upload the manuscript or any part of it, including tables, illustrations, data or other unpublished materials, to external GenAI tools.

AI tools may be used for limited assistance, for example to improve the language of the reviewer's own report or to search for general information and literature, provided that this does not involve disclosing the content of the manuscript under review. If a reviewer uses AI tools in preparing a review, the tool and the purpose of its use should be disclosed. This does not apply to tools used solely for basic spelling, grammar or punctuation checks.

AI must not prepare a review on behalf of the reviewer or replace the reviewer's scientific judgement. The reviewer remains responsible for the content of the review and the recommendation made.

Rules for the Editorial Team

Until publication, the manuscript and materials related to its evaluation are treated as confidential. Editors must not upload or otherwise provide them to external GenAI tools.

AI may assist with technical and editorial tasks provided that this does not require disclosure of confidential manuscript content. It must not replace editorial judgement, reviewer selection or decisions on whether a manuscript should be accepted or rejected.

Concerns about an Author's Use of AI

If reasonable concerns arise during the evaluation or preparation of a manuscript regarding the author's use of GenAI, the Editorial Team may ask the author to explain whether such tools were used and, if so, to what extent.

Depending on the nature of the concerns, the Editorial Team may also request materials needed to verify the integrity of the work, such as source data, calculations, information about the methods used or cited sources. The publication process may be suspended until the matter has been clarified.

Tools designed to detect AI-generated text may be used only where their use does not compromise manuscript confidentiality, data protection or the author's rights. The result of such screening is not proof that AI was used and must not, on its own, form the basis for rejecting a manuscript or finding a breach of publication ethics. It may, however, provide grounds for further verification and for requesting an explanation from the author.

If an author has failed to disclose substantial use of GenAI or has used it in a way that undermines the integrity or originality of the work, the Editorial Team may take appropriate action in accordance with the journal's Code of Publication Ethics.

This policy may be updated as technology and standards concerning the use of artificial intelligence in scholarly publishing evolve.