Should Authors Be Told When AI Was Used in Editorial Assessment?

Should Authors Know When AI Has Helped Evaluate Their Manuscript?

You submit a research manuscript to an academic journal. After several days, you receive a desk-rejection decision explaining that the paper does not sufficiently align with the journal's scope.

What you do not know is that an AI system generated the initial scope assessment. The handling editor reviewed the recommendation before issuing the decision, but the journal did not disclose that AI had contributed to the screening process.

Would knowing about the AI assistance matter? Should the journal explain how the technology was used, especially if the assessment influenced whether your manuscript received external peer review?

As AI becomes involved in manuscript screening, reporting checks, and editorial decision support, transparency raises questions about fairness, accountability, and the information authors need to understand consequential decisions.

The central issue is whether authors should be informed about every automated operation or primarily those uses that materially influence the scientific evaluation of their work.

Material AI Involvement Warrants Transparency, but Disclosure Rules Are Not Universal

Why Transparency Matters When AI Influences Editorial Decisions

Editorial Transparency Is Not Simply About Naming the Technology

Authors have a legitimate interest in understanding the basis on which their manuscripts are evaluated. Editorial decisions can affect publication opportunities, career progression, funding applications, and the dissemination of research findings.

When AI contributes to an assessment, merely stating that an automated system was used may provide little meaningful information.

More useful transparency identifies what the system was authorized to do, whether it produced a recommendation or made an administrative classification, and how the editor evaluated its output.

For example, a journal might use software to check whether a submission includes a conflict-of-interest statement. That differs substantially from using a generative AI model to assess the novelty or methodological quality of the manuscript.

The significance of disclosure should therefore reflect the system's actual role in the editorial process.

Not Every Automated Editorial Activity Has the Same Consequences

Editorial workflows already include various forms of software-assisted processing. Some are rule-based, some use machine learning, and others may involve generative AI.

These technologies should not automatically be treated as equivalent.

These are proposed transparency considerations rather than universal publishing requirements. Individual journals may adopt different arrangements.

What Do Publication-Ethics Authorities Require?

Publishing organizations increasingly address AI use in scholarly workflows, but their guidance should be interpreted carefully.

The International Committee of Medical Journal Editors (ICMJE) provides recommendations concerning AI use by authors, reviewers, and editors. Its guidance emphasizes confidentiality, transparency, and human responsibility for scholarly publishing activities.

The Committee on Publication Ethics (COPE) identifies transparency, editorial independence, and fair editorial processes among the principles relevant to publication ethics.

However, these general principles should not be represented as establishing an identical author-notification requirement for every automated editorial task.

Whether a journal must provide specific disclosure depends on its adopted policies, the nature of the AI involvement, and any applicable regulatory or contractual requirements.

Reviewers' obligations to disclose AI use to a journal are also distinct from a journal's obligations to inform authors about AI-assisted editorial assessment.

Why AI-Assisted Desk Rejection Deserves Particular Attention

A desk rejection prevents a manuscript from proceeding to external peer review. Although editors may legitimately reject submissions without external review, the basis for doing so should be scientifically and editorially defensible.

If AI contributes materially to the recommendation, authors may have a legitimate interest in understanding that involvement.

For example, a model may classify an interdisciplinary educational technology manuscript as primarily computer science because it emphasizes algorithmic terminology, overlooking its contribution to teaching and learning.

If the editor adopts that recommendation without sufficiently examining the manuscript, the decision may reflect a classification error rather than a genuine scope mismatch.

Transparent editorial procedures can help distinguish AI-supported assessment from automated decision-making that lacks meaningful oversight.

Does Human Approval Eliminate the Need for Disclosure?

Not necessarily.

An editor may formally approve an AI-generated recommendation while relying heavily on the system's initial assessment.

This creates the possibility of automation bias, in which human decision-makers give excessive weight to automated suggestions.

Meaningful human oversight requires more than clicking an approval button. Editors should examine the manuscript, assess the evidence supporting the recommendation, and be able to reject an incorrect classification or criticism.

When AI materially shapes the assessment, the fact that a human approved the final decision does not necessarily make the system's involvement irrelevant to transparency.

Should Authors Be Told Which AI Tool Was Used?

Identifying the tool may be useful, particularly when its capabilities or limitations materially affect the assessment.

However, a tool name alone does not explain what happened.

For example, stating that a journal used a particular language model does not reveal whether it checked formatting, summarized the manuscript, evaluated methodological weaknesses, or recommended rejection.

A meaningful disclosure may identify the tool or system, its purpose, the scope of its contribution, and the nature of human oversight.

The appropriate level of detail should reflect the consequences of the activity and the journal's disclosure requirements.

What Is the Difference Between Policy-Level and Manuscript-Specific Disclosure?

Journals may communicate AI use at different levels.

A general editorial policy might explain that automated systems assist with administrative screening and reporting checks. This informs all submitting authors about the journal's routine procedures.

A manuscript-specific disclosure might explain that AI-assisted scope classification contributed to a particular editorial assessment.

The second provides more direct information about how the author's submission was evaluated.

These approaches are complementary. A journal may reasonably use both, depending on the activity and its significance.

Can Disclosure Help Authors Challenge Incorrect Assessments?

Potentially, although disclosure alone does not establish a right to appeal.

Suppose an AI system incorrectly identifies a manuscript as outside the journal's scope. If the author knows that automated screening contributed to the assessment, they may be better positioned to request clarification through the journal's established procedures.

However, an appeal should be evaluated according to the manuscript's actual contribution and the journal's editorial standards, not merely because AI was involved.

Journals should consider whether their processes allow consequential automated errors to be identified and corrected.

The availability and scope of appeals remain matters of journal policy.

Why Transparency Matters for Unconventional Research

AI systems may reproduce patterns that favor familiar research topics, conventional methodological approaches, or dominant scholarly language.

Such tendencies are possible rather than inevitable and require empirical assessment.

Nevertheless, manuscripts employing unfamiliar theoretical perspectives or addressing underrepresented educational contexts may be particularly vulnerable to inappropriate classification.

For example, research examining technology use in resource-constrained schools may employ context-specific measures that differ from those common in well-resourced institutions.

An AI-generated assessment that treats those differences as methodological deficiencies could misrepresent the study's contribution.

Transparency can support scrutiny of these risks, although disclosure alone cannot eliminate bias in AI-assisted peer-review decisions.

Could Disclosure Create Misunderstandings About AI's Role?

Yes, if the wording is imprecise.

A statement that "AI evaluated your manuscript" may imply that the system independently made a scientific judgment, even when it only flagged missing administrative information.

Conversely, describing substantive AI-generated criticisms as "routine technical assistance" may understate their influence.

Journals should describe the actual function rather than use broad labels that obscure the distinction between administrative automation and scientific evaluation.

Clear disclosure can also prevent authors from assuming that every automated operation determined the final editorial outcome.

Does Disclosure Require Revealing Confidential Reviewer Information?

Not necessarily.

A journal may be able to describe AI involvement without revealing reviewer identities, confidential editorial deliberations, or protected information about other submissions.

Transparency should be designed alongside the journal's confidentiality obligations.

For example, a journal might disclose that an approved AI tool assisted with reporting-completeness screening without exposing internal reviewer correspondence.

The objective is to explain consequential technological involvement while preserving legitimate editorial confidentiality.

Should Journals Preserve Records of AI-Assisted Assessments?

Where AI materially influences editorial decisions, retaining appropriate records can support accountability and quality assurance.

Such records might include the system version, authorized task, generated recommendation, date of use, and the editor's documented assessment of consequential flags.

Retention should follow the journal's policies and applicable data protection requirements.

Records may help investigate disputed decisions, evaluate systematic errors, and determine whether a model update changed the system's performance.

Without adequate documentation, a journal may struggle to establish how an automated recommendation influenced an editorial decision.

Who Remains Responsible for the Decision?

The journal's authorized editors remain responsible for editorial decisions.

An AI system cannot independently explain the scholarly rationale for rejecting a manuscript, evaluate an appeal, or accept professional responsibility for an unfair assessment.

Disclosure should therefore identify AI as a supporting technology rather than as an accountable editorial decision-maker.

The distinction follows the broader principles governing AI-assisted editorial manuscript evaluation.

Transparent use is most defensible when the journal can explain both the system's contribution and the independent basis for the final decision.

Disclosing AI Assistance in a Desk-Rejection Decision

This example illustrates a possible transparency practice, not a statement that all journals currently provide or are required to provide such disclosures.

The disclosure is useful because it explains the system's limited role while identifying the human authority responsible for the decision.

Misconceptions About Author Disclosure of Editorial AI

How Journals Can Decide When Authors Should Be Informed

A proportionate disclosure policy should consider the purpose of AI processing and its influence on editorial outcomes.

The more directly AI affects scientific evaluation or publication opportunities, the stronger the case for specific disclosure and documented human oversight.

Explain the Decision, Not Merely the Presence of AI

Authors benefit most from understanding why the manuscript received a particular assessment.

A statement that AI was used should not replace a meaningful explanation of the journal's scientific or editorial reasoning.

Similarly, a journal should not present an AI-generated score as sufficient justification for rejecting a manuscript.

Disclosure is most valuable when it supports accountable editorial decisions and gives authors an accurate understanding of how their work was assessed.

Checklist for Author Transparency About Editorial AI

Frequently Asked Questions About AI Disclosure in Editorial Assessment

Authors Deserve Meaningful Transparency About Consequential AI Assessment

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.