Reviewer AI Policy

Artificial intelligence (AI) and AI-assisted technologies may create risks involving confidentiality, accuracy, bias, transparency, intellectual property, and human accountability when used during peer review. Because submitted manuscripts are confidential scholarly communications, the Celebes Nursing Journal (CNJ) requires all scientific evaluation, interpretation, reviewer comments, recommendations, and editorial judgments to remain human-led and human-accountable.

This policy applies to generative AI systems, large language models, chatbots, AI-assisted writing or translation tools, automated analytical systems, and other technologies capable of processing, retaining, reproducing, interpreting, or generating content. It is informed by the ICMJE Recommendations on the Use of AI by Reviewers and the COPE Ethical Guidelines for Peer Reviewers.

1. Core Principles
Reviewers must ensure that:

• Manuscript confidentiality is protected throughout and after the review process;
• Scientific and methodological evaluation is performed by the assigned human reviewer;
• AI systems are not treated as reviewers, co-reviewers, authors, or accountable contributors;
• Reviewer independence and impartiality are not compromised by automated outputs;
• No confidential manuscript information is disclosed to an unauthorized system or third party; and
• The reviewer remains personally responsible for the accuracy, integrity, originality, reasoning, and recommendation contained in the submitted review.

2. Prohibited Uses
Reviewers must not use generative AI or other unauthorized external technologies to:

• Upload or process a manuscript, manuscript excerpt, abstract, table, figure, supplementary file, dataset, reviewer report, author response, or confidential editorial correspondence;
• Summarize, translate, paraphrase, interpret, or extract information from confidential manuscript content;
• Assess scientific quality, methodological rigor, theoretical coherence, ethical compliance, originality, or research integrity;
• Generate, substantially draft, or determine reviewer comments, critiques, revision requests, or recommendations;
• Make or support decisions concerning acceptance, revision, or rejection;
• Conduct automated fact-checking, reference verification, similarity screening, or AI-content detection when doing so requires confidential manuscript content to be uploaded or disclosed;
• Identify or infer the identity, affiliation, personal characteristics, or location of authors in manuscripts undergoing double-blind peer review;
• Reproduce or retain confidential manuscript content outside the authorized peer-review system; or
• Delegate any substantive part of the assigned scientific or scholarly evaluation to an AI system.

This prohibition applies regardless of whether an AI service is free, paid, publicly available, institutionally licensed, or accessed through a personal or organizational account, unless its use has been expressly authorized by CNJ and confidentiality can be appropriately assured.

3. Limited Permitted Technological Assistance
Reviewers may use routine, non-generative, locally operated functions for limited non-substantive purposes, such as:

• Basic spelling and grammar checking of reviewer-written comments;
• Basic document formatting; or
• Accessibility functions that do not transmit, retain, reproduce, or externally process confidential manuscript information.

Such assistance is permitted only when it does not involve uploading or transmitting manuscript content, confidential editorial information, or substantive reviewer comments to an external system. Generative rewriting, substantive language generation, summarization, translation, interpretation, or analytical assistance is not considered routine non-substantive support.

If a reviewer is uncertain whether a particular tool or function is permitted, the reviewer must not use it and should seek guidance from the editor before proceeding.

4. Prior Permission and Disclosure
Reviewers must obtain prior permission from the handling editor before using any AI-assisted technology beyond routine non-generative and locally operated spelling, grammar, formatting, or accessibility functions.

When permitted AI-assisted technology is used, the reviewer must disclose the following information confidentially to the editor:

• The name and version of the tool, where known;
• The specific purpose for which it was used;
• The type of content processed;
• Whether any information was transmitted to an external system; and
• The measures taken to protect confidentiality and verify the accuracy of the resulting output.

Routine non-AI spelling, grammar, formatting, or accessibility functions that operate locally and do not transmit confidential content do not require disclosure.

5. Human Verification and Accountability
Permission to use a limited technological function does not transfer responsibility to the tool or its provider. Reviewers must independently verify all content included in their report and remain fully responsible for:

• The accuracy and validity of their comments;
• The appropriateness of their methodological and scholarly evaluation;
• The originality of the submitted review;
• The identification and correction of inaccurate, incomplete, fabricated, biased, or misleading output; and
• Compliance with CNJ’s confidentiality, ethical, and conflict-of-interest requirements.

6. Suspected AI Use by Authors
Reviewers should not conclude that authors have used AI solely on the basis of writing style, automated AI-detection scores, or unsupported assumptions. Reviewers must not upload the manuscript to external AI-detection systems or third-party services.

If a reviewer identifies a specific concern involving potentially undisclosed, inaccurate, fabricated, inappropriate, or unethical AI-assisted content, the concern should be reported confidentially to the editor with a clear explanation and relevant manuscript location. Reviewers must not directly accuse or contact the authors. The editor will assess the concern in accordance with CNJ’s AI Use Policy and publication-ethics procedures.

7. Breaches of This Policy
Unauthorized uploading of confidential material, AI-generated reviewer reports, delegation of scientific evaluation to AI, failure to disclose a required use, or other prohibited use may constitute a breach of reviewer confidentiality and professional conduct.

Depending on the nature and seriousness of the breach, CNJ may disregard the affected review, obtain an additional independent review, discontinue the reviewer’s assignment, suspend or discontinue future reviewer invitations, or take other proportionate editorial action.

8. Related CNJ Policies
This policy should be read together with CNJ’s Reviewer Ethics, Reviewer Guidelines, Peer Review Policy, and journal-wide AI Use Policy.