Authorship

Generative artificial intelligence (AI) and large language models (LLMs) do not meet the criteria for authorship as defined by the International Committee of Medical Journal Editors (ICMJE), the World Association of Medical Editors (WAME), and The Committee on Publication Ethics (COPE). Similarly, AI cannot hold copyright, sign legal agreements, or be held accountable for its outputs. As such, generative AI tools, LLMs, and other AI systems cannot be listed as authors or co-authors on any manuscript submitted to Ciencias Marinas. Human authors retain full responsibility and accountability for all content in a manuscript submitted to Ciencias Marinas, including written content, images, artwork, data analysis, results, citations, and references. 

Manuscript preparation

Text

AI tools may be used solely to improve the readability, grammar, and language of the manuscript written by the authors. This includes language editing, improvement, or translation that does not introduce new substantive scientific content. If you have used AI, you must disclose its use in your manuscript (consult the MS template for detailed instructions). 

Authors are strictly prohibited from using AI to generate original substantive text, citations, or references. Substantive scientific text includes content that introduces or develops scientific reasoning, interpretation, argumentation, hypotheses, conclusions, descriptions of research findings, or other original intellectual contributions.

Authors are responsible for reviewing and verifying all AI-assisted language modifications and ensuring the accuracy and integrity of the final text.

Images

Generative AI may not be used to create, alter, or enhance research data, measurements, graphs, or analysis outputs.

Ciencias Marinas allows authors to use AI to create purely illustrative or presentational elements that do not alter, generate, or represent research data or scientific findings. Permitted uses may include graphical abstracts, conceptual models, or schematic illustrations, although the authors are solely responsible for the content of these elements.

AI may not be used to generate substantive scientific interpretations or conclusions through graphical abstracts, conceptual models, schematic illustrations, or other visual elements.

Data analysis

Authors may use AI for data processing and analysis when its use constitutes a defined component of the research methodology. However, such use must be described in the Methods section of the paper with sufficient detail to ensure reproducibility, and an appropriate statement must be included in the Declarations section of the article. Authors must identify the AI tool and version used, describe its role in the analysis, and report relevant procedures, parameters, or settings.

Ciencias Marinas reserves the right to request original, unaltered versions of images; the raw data underlying figures, images, and analyses; and the original AI prompts, AI-generated outputs, tool name and version, relevant settings, and other records necessary to evaluate the AI-assisted procedure at any stage during the peer review or publishing processes.

Use of AI for data processing or analysis that violates this policy may result in rejection or, if identified after publication, retraction of the article.

Disclosure requirements

All manuscripts submitted to Ciencias Marinas must include a subsection titled "Disclosure of AI Use" located in the Declarations section immediately before the References found in the MS Template.

When AI has been used for language improvement, translation, permitted illustrative purposes, or another use requiring disclosure, the declaration must identify the tool, version, and specific purpose for which it was used:

• “The authors used [tool and version] for [specific purpose]. The authors reviewed all resulting material and accept responsibility for its accuracy, appropriateness, and compliance with the AI policy of Ciencias Marinas.”

When AI has been used as part of the research methodology, the declaration must identify its use, and the Methods section must provide the additional information necessary to describe the procedure transparently and reproducibly:

• “The authors used [tool and version] for [specific research procedure or analysis]. The application of this tool is described in the Methods section, including the information required to evaluate the procedure.”

If no AI or AI-assisted technologies were used in a manner requiring disclosure under this policy, the declaration must state:

• “No use of AI requiring disclosure under the Ciencias Marinas AI policy occurred in the preparation or conduct of this work.”

Authors are responsible for ensuring that their use of AI complies with applicable institutional requirements, licensing agreements, copyright restrictions, intellectual property requirements, and other legal or ethical obligations.

Peer review

Confidentiality 

Reviewers and editors are strictly prohibited from uploading any part of an unpublished manuscript, including text and images, to public generative AI tools (e.g., large language models [LLMs] like ChatGPT) to protect the intellectual property of the authors. If a reviewer is found to have uploaded a manuscript or any part of it to a public generative AI tool, the reviewer should be removed from the journal reviewer pool.

Authorized use by reviewers

Reviewers may use AI only to improve the language of their own review documents. Reviewers must transparently declare the tool name, version, and purpose of use to the corresponding editor in their confidential report. 

Suspected AI use by authors (undeclared)

Reviewers and editors may use approved AI detection tools to evaluate potential AI use. However, reviewers and editors should interpret any results from AI detection tools with extreme caution, as these tools can be unreliable and return false positives. 

If a reviewer suspects that a manuscript contains information generated by AI, they should bring it to the attention of the corresponding editor. If an editor is confident that AI was used by the authors and not disclosed appropriately, they should determine whether the AI tool was used in accordance with the AI policy of the journal. If AI use is found to be in violation of the guidelines, the editor should determine the appropriate action. 

• If the violation is undisclosed text editing, language improvement, or translation, the corresponding editor should request a correction in the Declarations section of the manuscript. 

• If AI is listed as an author, the corresponding editor should remove the AI tool from the author list and request clarification of the roles of each co-author based on the established CRediT taxonomy.

• If content was generated by AI and this use was not disclosed, the corresponding editor should request full disclosure from the authors and evaluate whether the use of AI violates the AI guidelines of Ciencias Marinas. The decision to request revision or to reject the article remains the prerogative of the corresponding editor.

• Hallucinations are one of the most problematic issues associated with the use of generative AI for writing. Hallucinations occur when an AI (e.g., an LLM) generates fabricated or unsupported information and presents it as factual. Reviewers and editors should be aware that hallucinations in scientific papers often take the form of fictitious citations and references. If hallucinations are present, the corresponding editor should reject the manuscript. However, depending on the particular case, the corresponding editor may allow resubmission as a new manuscript rather than revision of the rejected manuscript. 

• If data have been falsified or inappropriately manipulated using AI, the corresponding editor should reject the manuscript.