What Should Authors Include in AI Disclosure Statements?

Why AI Disclosure Statements Matter

Authors should state which AI tools they used, for what tasks, and at what stage. A useful disclosure names the product and version, describes whether it assisted with brainstorming, drafting, editing, analysis, image generation, or citation checking, and clarifies whether a human reviewed and approved every final claim. It should explain the scope of AI contribution, distinguishing minor language polishing from substantive generation of text, code, or visuals. For scientific work, authors should identify AI involvement in methods, results, or interpretation, because readers need to know where machine assistance shaped the record. Vague phrases such as “AI was used” are insufficient; specificity protects trust and helps editors assess accountability.

Also worth reading: AI Disclosure Compliance Checklist for Authors: What Do You Need? · What New AI Disclosure Policies Mean for Authors? · How Should AI Disclosure Contract Clauses Protect Authors and Publishers?

A strong statement also addresses data handling and confidentiality. Authors should disclose whether prompts or manuscripts were shared with third-party systems, whether sensitive material was entered, and whether AI-generated content was verified against primary sources. They should confirm that no AI system is listed as an author and note publisher requirements followed. The disclosure should be updated if workflows change. Clear language reduces confusion and supports fair evaluation.

What Authors Should Disclose

Authors should state which AI tools were used, what tasks they assisted with, and how outputs were verified. They should clarify whether AI contributed to drafting, editing, analysis, images, citations, or data handling, and identify any human review that shaped the final work. A useful disclosure names the tool and version when possible, describes the extent of assistance, and explains safeguards against errors, bias, or fabricated references. It should also note if AI was used to generate or transform content that could affect meaning, not merely to check grammar.

Disclosures should be specific enough for readers, editors, and reviewers to understand the author's responsibility. They should avoid vague phrases such as "AI was used" and instead explain the boundary between human judgment and machine assistance. Authors should mention any limitations, such as unverified suggestions or incomplete source checking, and confirm that they remain accountable for accuracy, originality, and ethical compliance. When policies require it, they should include statements about consent, confidentiality, and data protection, especially if unpublished material was entered into a system.

Publisher and Journal Requirements

Authors should state where and how generative AI was used, naming the tool, version, and date, plus the task, such as drafting, editing, translation, analysis, or image generation. They should clarify whether AI influenced research design, methods, results, or conclusions, and confirm that outputs were reviewed and revised by humans. Disclosures should identify AI-generated text, figures, tables, or code, and explain how accuracy, bias, privacy, and copyright concerns were handled. If no AI was used, authors should say so when required. A good statement is concrete, proportionate, and leaves human responsibility clear.

Authors should align disclosures with journal, publisher, funder, and institutional policies, updating them during revision. They should avoid vague phrases, disclose prompts or workflows when relevant, and state that AI is not an author because it cannot take responsibility. Even for language polishing, authors should confirm that meaning, citations, and claims remain accurate. Transparent disclosure protects trust, supports reproducibility, and helps readers distinguish human judgment from machine assistance. It also reduces the risk that undisclosed use is treated as misconduct, while showing authors remain accountable.

How to Write Clear Disclosures

Authors should state which AI tools they used, what tasks those tools performed, and how the output was reviewed or edited before publication. A useful disclosure names the system, version, and date of use, then describes whether it assisted with drafting, editing, translation, data analysis, image generation, or idea generation. It should also clarify the author's own contribution and confirm that they take responsibility for the final work.

Vague phrases like "AI was used" are rarely enough, because readers and reviewers need to understand the scope and nature of assistance. When AI contributed to methods or results, authors should explain validation steps and any limitations. Transparency builds trust, but it also protects authors by showing that human judgment remained central. Clear disclosures should be concise, specific, and placed where readers expect them, such as in an acknowledgments section, methods note, or end-of-article statement.

Common Mistakes to Avoid

Authors should state which generative AI tools they used, including product names, versions, dates, and specific tasks such as brainstorming, drafting, editing, translation, coding, or data analysis. They should describe how much final work relied on AI output, whether prompts or outputs were edited, and how human judgment verified facts, citations, tone, and originality. Disclosures should clarify that AI is not an author because responsibility remains with the human creator, and note any confidential or copyrighted material not entered into a tool.

A useful statement explains limitations and risks, including hallucinations, bias, outdated information, or uncertainty, plus steps taken to reduce them. It should identify who reviewed AI-assisted material, what sources were checked, and whether institutional, publisher, or funder policies were followed. When AI supported research, authors should distinguish analysis assistance from authorship and report reproducibility details. Clear, proportionate disclosure builds trust without overstating AI’s role, helping readers, editors, and reviewers understand the contribution and human accountability behind the finished work.

AI Disclosure Requirements Compared

Authors should disclose AI use clearly, naming tools, versions, tasks, human oversight, and data handling so readers can judge reliability and accountability. Transparent statements reduce confusion, protect trust, and align with publisher expectations without overstating AI's role or implying authorship. Keep disclosures factual, concise, and updated whenever tools, prompts, or workflows change.