Statement on the use of AI
Marine and Fishery Sciences (MAFIS) Policy on the Use of Artificial Intelligence and Automated Tools
Marine and Fishery Sciences (MAFIS) recognizes that artificial intelligence (AI) tools, including generative language models, writing-assistance systems, automated detection tools, image-analysis tools, similarity-checking software, bibliographic support systems, and other automated technologies, may contribute to improving certain aspects of manuscript preparation and editorial processing. However, their use must be governed by transparency, confidentiality, human responsibility, scientific integrity, and publication ethics.
AI tools do not replace the scientific judgment of authors, reviewers, or editors. No automated tool can assume responsibility for the content of a manuscript, a peer-review report, or an editorial decision. Final responsibility always rests with the individuals involved at each stage of the editorial process.
1. Instructions for Authors
Authors who use artificial intelligence tools or automated technologies during manuscript preparation, image creation or modification, literature searching or synthesis, text generation or editing, translation, data collection, data analysis, programming, preparation of tables, figures, or supplementary material must explicitly disclose such use in the manuscript.
It is not necessary to disclose the use of tools employed solely for spelling correction, basic grammar correction, formatting, minor stylistic adjustment, or elementary language improvement, provided that such tools have not generated scientific content, interpretations, data, analyses, citations, images, tables, conclusions, or new arguments.
When the use of AI goes beyond language correction or formatting, authors must include a statement titled “Declaration on the use of AI and automated tools”, preferably placed before the Reference list or in the Material and Methods section when appropriate. The statement must indicate:
- the name of the tool used;
- the version, model, or provider, if known;
- the purpose for which the tool was used;
- the sections of the manuscript affected;
- whether the tool was used for text, images, analysis, code, translation, literature review, or another function;
- confirmation that all authors reviewed, verified, and approved the final content.
Example of an acceptable statement:
Declaration on the use of AI and automated tools. The authors declare that they used [name of tool, version if applicable] for [specific purpose: for example, improving the linguistic clarity of the Abstract and Discussion / assisting with statistical code debugging / generating an initial translation of the text]. The content produced or modified with the assistance of this tool was reviewed, corrected, and approved by the authors, who assume full responsibility for the accuracy, originality, scientific integrity, and ethical compliance of the manuscript.
AI tools cannot be listed as authors or co-authors, as they cannot assume responsibility for the integrity, originality, accuracy, or final approval of the work. Authorship is reserved exclusively for individuals who meet the criteria for intellectual contribution, critical review, approval of the final version, and public responsibility for the content.
Authors must not cite generative AI tools as bibliographic sources of scientific information. References must correspond to verifiable, retrievable, and relevant sources. Authors are responsible for checking the existence, accuracy, and relevance of all citations and references included in the manuscript.
The use of AI does not exempt authors from complying with requirements concerning originality, research ethics, permits, informed consent, ethics committee approval, data integrity, data availability, conflicts of interest, and the policies of MAFIS. The generation, alteration, or manipulation of data, images, results, citations, authorship, or literature reviews through AI without adequate disclosure will be considered a potential breach of publication ethics.
MAFIS may request additional clarification about the use of AI, including prompts, files, code, editing records, original data, raw images, or any other information necessary to assess the scientific integrity of the manuscript. Failure to disclose relevant AI use may result in manuscript return, request for correction, editorial investigation, or the application of measures established in the journal’s publication ethics policies.
2. Instructions for Reviewers
MAFIS reviewers must preserve the strict confidentiality of manuscripts received for evaluation. The content of an unpublished manuscript, including its data, figures, tables, supplementary material, hypotheses, results, interpretations, and any related editorial communication, must not be uploaded, copied, summarized, processed, or shared through public generative AI tools, chatbots, unsecured external translation systems, automated analysis platforms, or any service that may store, reuse, train models on, or transfer submitted content to third parties.
Reviewers must not use generative AI tools to create, draft, or replace their scientific assessment of a manuscript. Peer review must result from the reviewer’s expert, independent, and responsible judgment. AI tools may produce comments that are superficial, biased, inaccurate, insufficiently specific, or based on false information, and cannot replace critical evaluation of methodology, data, results, interpretation, originality, and relevance.
Limited use of support tools may be acceptable to improve the linguistic clarity or formal organization of a review report already prepared by the reviewer, provided that manuscript confidentiality is not breached and sensitive content is not uploaded to unauthorized external systems. If a reviewer uses an AI or automated tool to edit or improve the wording of their review report, this must be disclosed to the editor in the confidential comments field.
Reviewers must evaluate manuscripts based on scientific integrity, methodological quality, data coherence, bibliographic relevance, originality, ethical compliance, clarity of presentation, and suitability for the scope of MAFIS. They must not base their recommendations exclusively on scores generated by AI detectors, automated-text detectors, similarity-checking tools, or other automated systems.
If a reviewer suspects undisclosed AI use, image manipulation, data fabrication, false citations, plagiarism, duplication, inappropriate authorship, artificial literature review, methodological inconsistencies, or any other integrity issue, they must report the concern to the editor in a clear, specific, and evidence-based manner. Reviewers must not directly accuse authors or initiate independent investigations.
The reviewer’s recommendation must focus on the scientific merit of the work. Declared AI use by authors is not, in itself, grounds for rejection, provided that it is transparent, limited, verified by the authors, and compatible with the scientific and ethical integrity of the manuscript.
3. Instructions for Editors
MAFIS editors are responsible for ensuring that the use of artificial intelligence and automated tools within the editorial process is transparent, proportionate, secure, and subject to human oversight. The final editorial decision regarding acceptance, rejection, revision, or referral to external review must always be made by a human editor and never by an automated tool.
Editors must not use public generative AI tools to create editorial assessments, make decisions, summarize confidential manuscripts, select reviewers, analyze unpublished content, or process manuscript data when such use may compromise confidentiality, author or reviewer privacy, copyright, research originality, or the integrity of peer review.
The use of automated tools by the journal may be appropriate for editorial support tasks, such as text similarity checking, reference verification, metadata checking, preliminary detection of image duplication or manipulation, formal file inspection, compliance checks with author guidelines, support in identifying potential reviewers, ORCID verification, editorial integrity analysis, or administrative assistance. These tools must not replace human review or automatically determine the editorial outcome of a manuscript.
Any alert generated by automated tools must be verified by an editor or authorized member of the editorial team before any action is taken. Alerts concerning similarity, possible figure manipulation, undisclosed AI use, questionable references, authorship irregularities, or automatic reviewer suggestions must be interpreted as indicators requiring human analysis, not as conclusive evidence.
The editorial priority is to identify substantive problems: inconsistent or impossible data, inadequate sampling designs, inappropriate statistical analyses, unsupported conclusions, nonexistent or irrelevant references, duplicated or manipulated images, undeclared conflicts of interest, authorship problems, insufficient ethical permits, plagiarism, redundant publication, or indicators of fabricated manuscripts. If potential irregularities are detected through automated tools or editorial review, authors should be given an opportunity to respond before the journal refers the case to institutions, ethics committees, or other external bodies, unless there are justified reasons to proceed otherwise under applicable publication ethics guidance.
Editors must apply these policies consistently, proportionately, and without discrimination. AI use by authors must not be treated as an ethical breach when it has been transparently disclosed, has not replaced the authors’ intellectual responsibility, and does not compromise the scientific validity of the work. However, undisclosed, misleading, or deceptive AI use intended to fabricate, manipulate, or conceal information may be considered a breach of MAFIS editorial policies.
