Generative AI Use and Disclosure Policy

Scope

Journal Marine Services (JMS) recognizes that automated tools can support legitimate language preparation and data workflows, but they can also introduce misleading content and privacy risks, particularly when manuscripts involve identifiable community participants, unpublished project records, or photographs.

Responsibilities of Authors

  • Authors remain accountable for all facts, evidence, calculations, references, images, and conclusions, whether or not any AI tool was used.
  • Generative AI systems cannot be authors. Their outputs cannot be cited as authoritative scientific sources without tracing and checking the underlying original source.
  • Authors must disclose substantive AI assistance beyond routine spelling, grammar, and basic formatting, including the tool, purpose, and relevant stages of the work.
  • Generative AI must not be used to fabricate questionnaires, participants, evaluation results, implementation photographs, quotations, field observations, institutional approvals, or references.
  • Authors must not upload personally identifying participant records, private partner information, or confidential materials to AI services without a lawful basis and adequate safeguards.

Responsibilities of Reviewers and Editors

Unpublished manuscripts and reviewer reports are confidential. Reviewers and editors must not submit them to external generative AI services for autonomous assessment or manuscript summarization without appropriate authorization and confidentiality controls. A human editor must make the actual editorial decision. Automated flags for similarity, generated text, or image concerns require a contextual human assessment and must not by themselves prove misconduct.

Suggested Disclosure Format

Generative AI Use: The authors used [verified tool and version, where relevant] for [specific task]. The authors reviewed and take full responsibility for all resulting content. No participant data or confidential manuscript materials were uploaded without appropriate authorization.

If no reportable AI assistance occurred, authors may state that no generative AI tools were used in preparing the substantive manuscript. Reports of AI use should reflect what was actually done; they must not be copied indiscriminately between articles.