Manuscript Preparation12 min readUpdated Mar 17, 2026

Journal AI Policies in 2026: What Authors Need to Know Before Submission

83% of high-impact journals now have AI policies. Here is what you must disclose, what is prohibited, and how to stay compliant across different journals.

Senior Researcher, Oncology & Cell Biology

Author context

Specializes in manuscript preparation and peer review strategy for oncology and cell biology, with deep experience evaluating submissions to Nature Medicine, JCO, Cancer Cell, and Cell-family journals.

Next step

Choose the next useful decision step first.

Use the guide or checklist that matches this page's intent before you ask for a manuscript-level diagnostic.

Open Journal Fit ChecklistAnthropic Privacy Partner. Zero-retention manuscript processing.Run Free Readiness Scan

Decision cue: If you used ChatGPT, Claude, Gemini, Copilot, or any AI tool during manuscript preparation, most journals now require you to disclose it. The policies differ by publisher: some require disclosure in the methods section, some in the cover letter, some in a separate declaration. Getting this wrong can delay your submission or, worse, trigger a post-publication investigation. The rules are evolving fast, and what was acceptable in 2024 may not be in 2026.

Check your manuscript readiness, including AI compliance signals, with the free scan.

The current landscape

As of 2026, approximately 83% of high-impact journals have explicit AI policies. The consensus across major publishers:

  • AI can assist with writing. Drafting, editing, language improvement, and literature searching are generally permitted with disclosure.
  • AI cannot be listed as an author. No journal accepts AI as a co-author. Authorship requires accountability, and AI cannot be accountable.
  • AI-generated content must be disclosed. How and where varies by publisher.
  • Authors are responsible for everything. If AI generates an error, fabricates a citation, or produces misleading content, the human authors bear full responsibility.

How policies differ across major publishers

Publisher
Disclosure required?
Where to disclose
AI as author?
Key requirement
Nature Portfolio
Yes
Methods or Acknowledgments
No
Describe how AI was used and which tool
Elsevier
Yes
At submission + in manuscript
No
Authors must verify all AI-assisted content
Springer Nature
Yes
Methods or Acknowledgments
No
AI use in manuscript preparation only, not research
AAAS (Science)
Yes
Methods section
No
Must specify exact tools and how they were used
AMA (JAMA)
Yes
At submission
No
Disclosure of all AI-assisted technologies
NEJM
Yes
At submission
No
Must be able to assert no plagiarism
Cell Press
Yes
Acknowledgments
No
Describe the role of AI in manuscript preparation
Wiley
Yes
Acknowledgments or Methods
No
Author-level responsibility for accuracy
PLOS
Developing
Varies
No
Transparency about AI use encouraged

What you must do before submitting to any journal

1. Check your target journal's specific AI policy

Policies are not standardized. Some journals require disclosure in the methods section. Others want it in the acknowledgments. Some require it during the online submission process in a separate form field. Check the author guidelines for your specific target journal.

2. Document which AI tools you used and how

Be specific: "We used ChatGPT (OpenAI, GPT-4) to assist with language editing of the discussion section. All output was reviewed and revised by the authors." is acceptable at most journals.

"We used AI to help write the paper" is too vague.

3. Verify everything the AI produced

This is not just a policy requirement. It is a practical necessity. AI tools hallucinate citations, fabricate statistical claims, and generate confident-sounding statements that are factually wrong. Every claim, every reference, and every data point that an AI tool touched must be verified by a human author.

The Manusights AI Diagnostic ($29) verifies citations against 500M+ live academic papers (CrossRef, PubMed, OpenAlex, Semantic Scholar, bioRxiv, medRxiv). This catches fabricated references that ChatGPT or other AI tools may have generated during drafting. At $29, this is the most cost-effective way to ensure your references are real before a reviewer discovers they are not.

4. Do not use AI-generated images without disclosure

Some journals explicitly prohibit AI-generated images (figures, diagrams, or photographs) unless disclosed and justified. If you used DALL-E, Midjourney, or similar tools to create any visual content, check whether your target journal permits it.

5. Do not claim AI-generated text as entirely your own

If an AI tool wrote substantial portions of the manuscript, this must be disclosed. Not disclosing is a form of plagiarism under most journal policies. The risk is not just rejection but retraction and reputational damage if discovered post-publication.

Common compliance mistakes

Forgetting to disclose AI use in the cover letter or submission form

Many journals have added an AI disclosure checkbox or text field to their online submission systems. Missing this is a procedural error that can delay processing. Check the submission form carefully.

Using AI to generate the reference list

This is one of the highest-risk uses of AI in manuscript preparation. AI tools generate plausible-looking references that may not exist. A 2025 analysis found over 100 hallucinated citations in papers accepted at a top machine learning conference. If you used AI to suggest or generate references, every single one must be verified.

Run the free readiness scan to check citation integrity. If the scan flags citation issues, the $29 diagnostic provides detailed verification against live databases.

Assuming the same policy applies across journals

A manuscript prepared for Nature (which requires disclosure in Methods or Acknowledgments) may need a different disclosure format for Elsevier (which requires disclosure at submission and in the manuscript). If you resubmit a rejected paper to a different publisher, update the AI disclosure to match the new journal's requirements.

Not disclosing because "everyone uses it"

The fact that over 50% of peer reviewers now use AI during review does not exempt authors from disclosure requirements. The policies are asymmetric: reviewers are largely unregulated while authors must disclose. This may change, but as of 2026, the disclosure burden falls on authors.

What journals are actually looking for

Editors and reviewers are not trying to ban AI use. They are trying to ensure:

  • Accuracy. AI-assisted content has been verified by humans.
  • Transparency. Readers know which parts of the work involved AI.
  • Accountability. Human authors take responsibility for everything in the manuscript.
  • Integrity. Citations are real. Data are real. Claims are human-verified.

A manuscript that uses AI thoughtfully, discloses it transparently, and verifies everything carefully is not penalized. A manuscript that uses AI carelessly, does not disclose it, and contains AI-generated errors will be.

How Manusights handles AI in its own output

Manusights is transparent about its approach:

  • the AI Diagnostic is powered by AI (built on Anthropic technology)
  • every citation in the diagnostic report is verified against live databases, not generated from training data
  • manuscripts are processed once, then deleted (Anthropic Privacy Partner, zero-retention)
  • the scoring rubric was trained on actual Cell, Nature, and Science peer review documents
  • no manuscript content is used for model training

This matters because "AI-powered review" means very different things across services. A tool that generates plausible-sounding feedback from training data is different from one that verifies claims against live databases. Ask any AI-powered review service how they handle citations, and compare the answer.

References

Sources

  1. Nature Portfolio AI policy
  2. Elsevier AI policy
  3. JAMA AI disclosure requirements
  4. AAAS/Science editorial policies
Navigate

On this page

Reference library

Use the core publishing datasets alongside this guide

This article answers one part of the publishing decision. The reference library covers the recurring questions that usually come next: how selective journals are, how long review takes, and what the submission requirements look like across journals.

Open the reference library

Before you upload

Want the full journal picture?

Scope, selectivity, what editors want, common rejection reasons, and submission context, all in one place.

These pages attract evaluation intent more than upload-ready intent.

Anthropic Privacy Partner. Zero-retention manuscript processing.

Internal navigation

Where to go next

Open Journal Guide