Does AI Assistance in a Response-to-Reviewers Letter Require Disclosure?

If AI Helps Write Your Response to Reviewers, Must You Tell the Journal?

You receive reviewer comments requesting substantial revisions to your manuscript. After completing the changes, you use generative AI to improve the wording of your point-by-point responses. Perhaps the tool reorganizes your explanations, suggests more diplomatic language, or drafts responses from notes describing the revisions you made.

The final letter accurately represents your work, and you have checked every claim. But should you disclose that AI helped prepare it?

Many researchers are familiar with AI declarations in manuscripts. The position of response-to-reviewers letters is less obvious because they are submitted during editorial correspondence and may not appear in the published article.

The distinction matters. A journal may require disclosure of AI assistance in submitted materials even when the assistance does not directly affect the research findings. Other policies may distinguish basic proofreading from substantive content generation.

Disclosure May Be Required, Depending on the Journal and How AI Was Used

How AI Disclosure Rules Apply to Response-to-Reviewers Letters

Why a Response Letter Creates a Distinct Disclosure Question

A response-to-reviewers letter is part of the journal's editorial process. It explains how authors addressed reviewer concerns and may influence whether the revised manuscript is accepted, returned for further revision, or rejected.

Unlike the manuscript itself, the response letter is often not published. Nevertheless, it contains substantive representations about the research and the authors' revision decisions.

Generative AI may contribute to the letter in different ways. It can correct punctuation, reorganize comments, suggest diplomatic wording, or generate complete responses from information supplied by the authors.

These activities are not necessarily equivalent from a disclosure perspective.

For example, correcting a grammatical error generally does not introduce a new intellectual contribution. Drafting an explanation of why a statistical analysis was not performed may involve substantive argumentation, even when the authors subsequently revise and approve the text.

The appropriate disclosure decision should therefore consider what the AI actually contributed, not merely whether the finished letter contains AI-generated wording.

What Do Major Publishing Authorities Say About AI Disclosure?

Publishing organizations increasingly distinguish between routine language assistance and substantive generative AI use. Their requirements are not identical, and a policy addressing manuscript preparation may not explicitly specify every type of editorial correspondence.

The International Committee of Medical Journal Editors (ICMJE) recommends transparency about AI use throughout the publishing process. Its guidance calls for authors to describe relevant AI assistance in submitted work and accompanying disclosures.

Elsevier's generative AI policy, updated in 2026, requires disclosure of substantive AI assistance in manuscript preparation while exempting basic spelling, grammar, and punctuation checks. Its guidance distinguishes manuscript-preparation assistance from AI used in the research methodology.

These policies provide important principles, but authors should not infer from them that every publisher has adopted an identical rule specifically for response-to-reviewers letters.

Publisher policies can change, and individual journals may impose additional requirements. Authors should consult the current official instructions rather than rely on an AI-generated summary of a policy.

Does Basic Grammar Correction Require Disclosure?

Not always.

Some publishers explicitly exempt basic spelling, grammar, and punctuation checks from their generative AI declaration requirements. Elsevier's current policy, for example, makes this distinction.

However, the boundary between proofreading and substantive rewriting can become unclear.

Correcting an incorrectly placed comma is different from rewriting an argument about the adequacy of a research method. A tool may begin with language correction but substantially change the meaning, emphasis, or organization of the response.

Researchers should therefore consider the actual extent of AI intervention rather than classify every use of a writing assistant as routine proofreading.

Where a journal provides an explicit exemption, authors may rely on it within its stated scope. Where the policy is silent, clarification may be preferable to assuming an exemption.

When Does AI Assistance Become Substantive?

Substantive assistance generally involves more than correcting surface-level language errors.

It may include generating a complete response from reviewer comments, developing arguments to justify a methodological decision, restructuring an extended explanation, or proposing claims about how a revision addresses a concern.

For example, asking AI to correct grammar in a completed paragraph differs from asking it to write a persuasive justification for retaining a particular analytical approach.

The latter may influence the content and reasoning of the correspondence, even if the authors retain final editorial control.

Authors should be especially attentive when using AI to explain a disagreement with a reviewer. The reasoning must reflect the authors' actual methodological position rather than an argument invented to sound convincing.

Substantive AI assistance is generally a stronger reason to disclose, particularly when the journal's policy calls for transparency about AI-generated content.

Where Should AI Use in a Response Letter Be Disclosed?

The appropriate location depends on the journal's instructions.

Some submission systems provide a dedicated AI-use declaration field. Others require a statement within the manuscript, cover letter, or accompanying submission documents.

If the journal explicitly requires disclosure in the response-to-reviewers letter, authors should place the statement where instructed.

If the journal requires a manuscript declaration but does not address revision correspondence, authors should determine whether the AI assistance also affected the manuscript itself. A response-letter-only use should not automatically be described as AI-assisted manuscript drafting.

Where no location is specified, a short statement in the response letter or accompanying editorial correspondence may provide transparency without unnecessarily interrupting the point-by-point responses.

Authors should avoid creating multiple inconsistent declarations across submission documents. The description of AI use should accurately reflect what occurred in each document.

What Should an AI Disclosure Statement Include?

A useful declaration should identify the AI tool, explain the purpose of its use, and clarify that the authors reviewed and approved the resulting material.

It should describe the actual assistance rather than use vague language such as "AI was involved in preparing this submission."

For example, when AI was used to improve the organization and wording of responses already developed by the authors, the declaration can say so directly.

If AI generated initial response drafts from the authors' revision notes, that more substantive contribution should not be disguised as simple proofreading.

The declaration need not reproduce every prompt unless the journal requests such documentation or the prompts are methodologically relevant.

Does Disclosing AI Use Mean the Journal Will Reject the Revision?

Disclosure itself does not establish that the assistance was inappropriate.

Many publishers permit certain forms of generative AI assistance while requiring transparency and human oversight. The important question is whether the use complied with the applicable policy and whether the submitted material remains accurate and accountable.

However, researchers should not assume that disclosure guarantees acceptance of every AI-assisted practice. A journal may restrict particular uses, especially those involving confidential material, unsupported content generation, or prohibited processing of manuscripts.

Authors should distinguish between permission to use AI and the requirement to disclose that use. These are related but separate policy questions.

Does Disclosure Resolve Confidentiality Concerns?

No. A declaration cannot retroactively authorize an inappropriate disclosure of confidential material.

Reviewer comments, decision letters, and unpublished manuscripts may be subject to contractual, editorial, or institutional confidentiality obligations.

Authors should determine whether the intended AI service is permitted to process those materials before uploading them.

This includes examining data retention, provider access, model-training terms, institutional arrangements, and the journal's instructions.

Even when the final AI-generated wording is accurate, the processing of confidential information may create a separate compliance concern.

The question of whether confidential reviewer comments can be uploaded to AI tools should therefore be resolved before the tool is used, not merely addressed in a later disclosure statement.

Does Disclosure Replace Independent Verification?

Absolutely not.

Disclosure explains the role of AI in preparing the correspondence. It does not establish that the AI-generated material is accurate.

A response letter may openly acknowledge AI assistance while still containing an invented analysis, incorrect manuscript reference, or exaggerated claim about a completed revision.

Authors should therefore independently verify the revised manuscript and response letter against the reviewer comments and underlying research records.

This responsibility remains with the authors regardless of whether the AI contribution was disclosed.

Should AI Use Be Disclosed Again During a Second Revision?

That depends on the journal's instructions and whether additional AI assistance was used.

If a journal requires an AI declaration for each submission or revised version, authors should comply with that requirement.

If the AI-assisted activities change substantially between revisions, the earlier declaration may no longer describe the complete extent of assistance.

For example, an initial declaration limited to language editing would be incomplete if AI was subsequently used to develop substantive responses or assist with research analysis.

Researchers should maintain an accurate record of AI use throughout the revision process so the declaration can be updated when necessary.

Deciding Whether to Disclose AI Assistance in a Major Revision

Now consider a different situation. The researcher uses a spelling checker solely to correct typographical errors in an otherwise completed response letter.

If the journal explicitly exempts routine spelling and grammar checks, a separate declaration may not be necessary. The answer changes because the nature of the assistance and the applicable policy are different.

The practical lesson is that disclosure should accurately describe the work performed by AI rather than depend on a blanket assumption that every AI-assisted action is either always reportable or always exempt.

Misconceptions About AI Disclosure in Reviewer Responses

How to Decide Whether Your Response Letter Needs an AI Declaration

Begin with the journal's official instructions, then examine the actual role AI played in preparing the correspondence.

The most defensible approach is to distinguish routine language correction from substantive drafting while avoiding assumptions about requirements that the journal has not explicitly addressed.

Keep the Declaration Specific and Proportionate

A useful declaration should be factual rather than defensive. It should identify the tool, describe its contribution, and explain the authors' oversight.

There is no need to suggest that AI performed scientific analyses when it only assisted with language. Equally, substantial drafting should not be described as basic grammar correction.

Authors should also avoid implying that the AI independently resolved reviewer concerns. The actual intellectual and methodological decisions must remain attributable to the researchers.

Transparency about AI assistance should accompany, not replace, human accountability for every claim in the response letter. A declaration explains how the document was prepared; the authors must still ensure that it accurately represents the work completed.

AI Disclosure Checklist for Response-to-Reviewers Letters

Frequently Asked Questions About AI Disclosure in Reviewer Responses

Disclose Substantive AI Assistance According to the Journal's Rules

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.