# How to write an AI disclosure statement for medical journals?

Brooklyn Bishop · November 7, 2026

> Understanding CASRAI AI disclosure guidelines Crafting a robust AI disclosure statement requires precise alignment with specific journal mandates and...

## Understanding CASRAI AI disclosure guidelines

Crafting a robust AI disclosure statement requires precise alignment with specific journal mandates and emerging ethical frameworks. Writers must explicitly detail the extent of artificial intelligence involvement in manuscript preparation, distinguishing between automated language polishing and substantive content generation. This transparency ensures readers can accurately assess the originality and reliability of the published work. Journals increasingly demand clear attribution of AI tools used for drafting, editing, or data analysis, preventing misleading claims of human authorship. By adopting standardized phrasing from recognized bodies like CASRAI, authors can navigate the complex landscape of publication ethics effectively. This approach fosters trust among peer reviewers and readers who rely on accurate representation of research methodologies and intellectual contributions.

**Also worth reading:** [Can AI Disclosure Requirements for Journals Bridge the Gap Between Law and Publisher Policy?](https://storywriter.pro/knowledge/can_ai_disclosure_requirements_for_journals_bridge_the_gap_between_law_and_publisher_policy.php) · [What Must an AI Disclosure Checklist for Journals Include to Meet Rising Ethical Standards?](https://storywriter.pro/knowledge/what_must_an_ai_disclosure_checklist_for_journals_include_to_meet_rising_ethical_standards.php) · [Generative AI Author Disclosure Template: What Should Scientific Journals Require?](https://storywriter.pro/knowledge/generative_ai_author_disclosure_template_what_should_scientific_journals_require.php)

Furthermore, integrating these disclosures into CRediT authorship contribution statements clarifies individual roles in the research process. Authors should specify whether AI assisted in literature reviews, statistical analysis, or text generation, ensuring full accountability. This granular level of detail helps maintain integrity within medical publishing, where precision and veracity are paramount. As guidelines evolve, staying informed about current best practices is essential for compliance. Ultimately, a well-crafted disclosure statement not only satisfies regulatory requirements but also upholds the highest standards of scholarly communication. It empowers the scientific community to evaluate the true scope of human versus machine contribution, preserving the credibility of medical literature in an era of rapid technological advancement.

## Drafting AI assistance statements for authors

One unresolved issue is how specific an author must be when describing the use of generative tools. A strong statement names the tool and version, identifies the exact stage of work where it was used, and clarifies that the technology did not contribute to study design, data analysis, or interpretation. For example, an author might note that a language model assisted with copyediting for clarity and grammar, while all scientific content was created and verified by the research team.

Equally important is stating where human responsibility begins and ends. Medical journals expect authors to confirm that they reviewed every AI-assisted passage, checked citations against original sources, and accept full accountability for the manuscript’s accuracy and integrity. Placing this disclosure in the cover letter, methods section, or acknowledgments—according to the journal’s instructions—helps editors assess compliance without ambiguity. Clear, task-based wording protects transparency and preserves trust in the published record.

## Handling AI generated medical images

AI-generated or AI-enhanced medical images should be disclosed separately from text assistance, even when no generative tool was used in drafting the manuscript. The statement should name the tool and version, describe whether it created, segmented, denoised, colorized, or annotated the image, and state that a qualified clinician or imaging specialist reviewed the final figure for anatomical accuracy and diagnostic fidelity. If the image is illustrative rather than patient-derived, that distinction should be explicit. This prevents readers from mistaking synthetic visual support for original clinical evidence and gives editors enough detail to assess provenance.

A concise disclosure can read: "Figure X was generated with [tool/version] from [source data/prompt] and reviewed by [role] for accuracy; no patient-identifying information was included." If multiple figures differ, list each. When AI only adjusted contrast or removed artifacts, say so rather than using vague wording such as "AI-assisted imaging." The key is proportionate specificity: enough to show human oversight and source, not a full technical appendix. Journals can then apply their image-integrity checks without guessing what the tool did.

## Integrating AI into CRediT role statements

One unresolved question is whether AI assistance belongs inside the CRediT taxonomy itself, since CRediT defines fourteen human contributor roles and journals remain divided on how to record a tool's involvement. The most workable approach is to add AI as a footnote rather than a contributor: authors list their own roles as usual, then append a statement such as "AI tools were used under the supervision of the author fulfilling the Writing – Original Draft role." This preserves the principle that accountability attaches to people, not software, while still signaling where machine assistance occurred.

Journals differ here: some, like JACC, have decided no disclosure is required at all, while others expect explicit naming of tools and versions. For medical journals that do require disclosure, the safest wording ties the AI mention to a specific CRediT role, because it tells editors and readers who took responsibility for verifying the output. A typical line reads: "Conceptualization, A.B.; Methodology, A.B.; Writing, A.B. and C.D., with generative AI assistance for language editing, reviewed and verified by C.D." This satisfies transparency requirements without distorting the authorship record.

## Addressing generative AI in medical writing

A workable AI disclosure statement should not merely say “AI was used.” For medical journals, disclose every generative tool that materially shapes drafting, editing, data interpretation, figure creation, or reference handling, naming the tool and version, access date, purpose, and affected sections. Then state which author reviewed and verified the output and that all authors retain full responsibility for accuracy, originality, and scientific integrity. This threshold resolves the common ambiguity about editing-only use: minor spelling or grammar checks need no statement, but substantive language editing, literature synthesis, or image generation does.

Place the statement in the manuscript’s acknowledgments or a dedicated declaration, and mirror it in cover letters, CRediT statements, and funder reports without contradiction. For images, specify whether AI generated, altered, or only labeled them; for text, distinguish assistance from authorship. Avoid vague assurances such as “AI-assisted writing” without scope. A concise, verifiable statement gives editors, reviewers, and institutions the same auditable record, while preserving human accountability as the non-negotiable boundary.

## AI Disclosure Statement Comparison

| Criterion | Prose AI disclosure | Structured/CRediT-style disclosure |
| --- | --- | --- |
| Best use | Brief note in acknowledgments or cover letter when AI only assisted with language editing or formatting. | Role-by-role statement when AI contributed to drafting, analysis, literature synthesis, figures, or translation. |
| Journal fit | Acceptable for simple assistance if the journal permits it; name the tool, version/date, purpose, and author responsibility. | Safer for ICMJE-aligned journals and publishers using CRediT; map AI use to writing, review, visualization, or editing while stating AI cannot be an author. |
| Required details | Tool/version, date accessed, task, extent of use, confidentiality safeguards, and human verification. | Same details plus explicit role category, affected section or output, and a named accountable author. |
| Main risk | Vague “AI was used” wording can fail editorial screening or confidentiality checks. | Over-claiming AI authorship or using CRediT categories that imply accountability AI cannot hold; keep disclosure factual and author-owned. |

For medical journals, start with the target journal’s policy, then use a structured role-by-role statement for substantive AI use and a concise prose note only for minor editing. Name the tool/version, date, task, affected section, and accountable human author; confirm confidentiality and verify all output. Never list AI as an author. This satisfies most ICMJE-aligned and publisher requirements while keeping responsibility transparent.

## Quick answers

### What is the standard wording for AI disclosure?

Authors should state the specific AI tool used and how it assisted in the manuscript preparation.

### Do medical journals require AI disclosure?

Most major medical journals now mandate disclosure of generative AI use in writing or image generation.

### How to disclose AI in CRediT statements?

List AI tools as non-human contributors and clarify the human author's oversight of the output.

### What about AI generated medical images?

Disclose if images are AI generated and specify the model used for transparency and reproducibility.

Canonical: https://storywriter.pro/knowledge/how_to_write_an_ai_disclosure_statement_for_medical_journals.php
Markdown: https://storywriter.pro/knowledge/how_to_write_an_ai_disclosure_statement_for_medical_journals.php/index.md