BMJ Open AI Policy: ChatGPT and Generative AI Disclosure Rules for BMJ Open Authors
BMJ Open requires AI disclosure under BMJ rules. AI cannot be an author. This guide covers where to disclose, what to disclose, and the consequences of non-compliance for BMJ Open submissions.
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.
BMJ Open at a glance
Key metrics to place the journal before deciding whether it fits your manuscript and career goals.
What makes this journal worth targeting
- IF 2.3 puts BMJ Open in a visible tier — citations from papers here carry real weight.
- Scope specificity matters more than impact factor for most manuscript decisions.
- Acceptance rate of ~27% means fit determines most outcomes.
When to look elsewhere
- When your paper sits at the edge of the journal's stated scope — borderline fit rarely improves after submission.
- If timeline matters: BMJ Open takes ~134 days median. A faster-turnaround journal may suit a grant or job deadline better.
- If OA is required: gold OA costs £2,390 GBP. Check institutional agreements before submitting.
Quick answer: The BMJ Open AI policy follows BMJ's rules calibrated to clinical and population-health research evaluated on methodology rather than perceived novelty submissions. AI tools can be used for manuscript preparation but every use must be disclosed in the Methods section, with BMJ Open's editorial team checking specifics at desk-screen. AI cannot be listed as an author of any BMJ Open paper. AI-generated figures and schematics representing original research data are prohibited under BMJ Open's image-integrity standard. BMJ Open editors treat undisclosed use as a publication-ethics violation per ICMJE + COPE.
Run the BMJ Open submission readiness check which includes an automated AI-disclosure audit, or work through this guide manually. Need broader context? See the BMJ Open journal overview.
The Manusights BMJ Open readiness scan. This guide tells you what BMJ Open's editors look for when verifying AI disclosure at desk-screen. The scan tells you whether YOUR Methods section has the required language before you submit. We have reviewed manuscripts targeting BMJ Open and peer venues; the named patterns below are the same ones Adrian Aldcroft and BMJ Publishing Group's editorial AI committee flag at the desk-screen and editorial-board consultation stages. 60-day money-back guarantee. We do not train AI on your manuscript and delete it within 24 hours.
Editorial detail (for desk-screen calibration). Editor-in-Chief: Adrian Aldcroft (BMJ Publishing Group) leads BMJ Open editorial decisions. Submission portal: https://mc.manuscriptcentral.com/bmjopen. Manuscript constraints: 300-word abstract limit and 5,000-word main-text cap (BMJ Open enforces methodological completeness). We reviewed BMJ's AI policy framework against current BMJ Open author guidelines (accessed 2026-05-08); evidence basis includes both publicly documented BMJ policy and our internal anonymized submission corpus. The applicable word limit at BMJ Open is shown below: 300-word abstract limit and 5,000-word main-text cap (BMJ Open enforces methodological completeness).
Verify exact word and figure limits against the latest author guidelines before submission. The named editorial-culture quirk: BMJ Open reviewers consistently flag CONSORT/STROBE/PRISMA checklist incompleteness; methodology-first review means missing items extend revision.
What does BMJ Open's AI policy require?
BMJ Open authors must follow four rules under BMJ's AI framework, all enforced at desk-screen:
Rule 1: Disclose every AI tool used in manuscript preparation
Authors must name every generative AI tool used, its version, and how it was used. The disclosure goes in the Methods section, not the Acknowledgments. Examples that REQUIRE disclosure at BMJ Open:
- For BMJ Open-targeted manuscripts addressing clinical and population-health research evaluated on methodology rather than perceived novelty: using ChatGPT, Claude, Gemini, or similar to draft, polish, or edit manuscript text passing through BMJ Open editorial review
- For BMJ Open submissions: using AI to generate boilerplate text for limitations, ethics statements, or BMJ Open-specific response-to-reviewers letters that cite BMJ's framework
- For BMJ Open submissions: using AI to translate manuscript text into English from another language, with BMJ expecting disclosure of the source language and translation chain
- For BMJ Open literature reviews: using AI for citation discovery or summarizing prior BMJ Open work; BMJ's policy applies regardless of citation context
- For BMJ Open analytical pipelines: AI-assisted code generation requires Methods + code disclosure under ICMJE + COPE, particularly when code touches clinical and population-health research evaluated on methodology rather than perceived novelty analysis
Examples that do NOT require AI disclosure:
- At BMJ Open, using grammar/spell checkers (Word, Grammarly basic) that do not generate new content for the manuscript
- For BMJ Open submissions, using reference managers (Zotero, EndNote) for citation formatting against BMJ's style guide
- For BMJ Open statistical analysis, using established statistical software (R, Stata, SPSS) where the algorithm is the established tool documented in BMJ Open's methodological norm, not a generative AI
Rule 2: AI cannot be an author
No AI tool can be listed as an author of a BMJ Open paper, particularly for clinical and population-health research evaluated on methodology rather than perceived novelty-class submissions. Under BMJ's policy: authorship requires the ability to take responsibility for the content, agree to be accountable for accuracy, and to consent to publication. AI tools cannot do any of these in BMJ Open's editorial framework. This rule is consistent across all BMJ-published journals and applied at BMJ Open's desk-screen.
Rule 3: AI-generated figures are prohibited for original research data
BMJ Open editorial team does not accept AI-generated images, figures, or schematics that represent original research data in clinical and population-health research evaluated on methodology rather than perceived novelty-class submissions. AI tools may assist with figure layout (axis labeling, color schemes) but the underlying data visualization must come from the actual research. AI-generated diagrams used for conceptual illustrations (e.g., a schematic of a hypothesized mechanism) require explicit disclosure and a statement that the diagram is conceptual.
Rule 4: Disclose AI use in peer review participation
Reviewers writing reports for BMJ Open cannot use generative AI to draft their reports without disclosing it to the editor. Some BMJ journals prohibit AI-assisted reviewing entirely; BMJ Open follows BMJ's default of disclosure-required. The editor decides whether the report is acceptable based on disclosure.
How does BMJ Open's AI policy compare to peer journals?
Rule | BMJ Open stance | BMJ default | ICMJE/COPE alignment |
|---|---|---|---|
AI authorship | Prohibited | Prohibited | ICMJE-aligned |
Disclosure location | Methods section | Methods section | ICMJE-aligned |
AI-generated figures | Prohibited for original data | Prohibited | COPE image-integrity-aligned |
Reviewer AI use | Disclosure required | Disclosure required | COPE peer-review-aligned |
Enforcement intensity | Desk-screen check | Desk-screen check | Pre-publication enforcement |
Source: https://authors.bmj.com/policies/ (accessed 2026-05-08) plus BMJ Open author guidelines.
What does AI disclosure look like in a BMJ Open Methods section?
Acceptable disclosure language for BMJ Open submissions:
"For our clinical and population-health research evaluated on methodology rather than perceived novelty-focused manuscript at BMJ Open, we used ChatGPT-4o (OpenAI, version dated October 2024) to polish English-language phrasing in the Introduction and Discussion sections. We did not use generative AI for data analysis, figure generation, or substantive manuscript content. All authors reviewed and edited the AI-assisted text and take responsibility for the final manuscript."
Or, for AI-assisted code:
"For this BMJ Open submission addressing clinical and population-health research evaluated on methodology rather than perceived novelty, initial Python code for the Bayesian regression analysis was drafted with Claude 3.5 Sonnet (Anthropic, version dated December 2024). All code was reviewed, modified, and validated by the authors before use; the final version is available at [repository URL]. Statistical inference was performed using the established R package brms."
What does NOT pass BMJ Open's desk-screen:
- For BMJ Open addressing clinical and population-health research evaluated on methodology rather than perceived novelty: "AI tools were used in manuscript preparation." Too vague for BMJ editorial review of BMJ Open submissions; the BMJ Open editorial team needs the specific tool name, version, and specific use case
- "We acknowledge AI assistance in the Acknowledgments." (Wrong location; must be Methods)
- "ChatGPT helped write this paper." (Insufficient detail on use case)
- No disclosure when AI was used (publication-ethics violation)
What do pre-submission reviews reveal about BMJ Open's AI-disclosure desk-screen failures?
In our pre-submission review work on BMJ Open-targeted manuscripts, three patterns most consistently predict AI-policy desk-screen flags at BMJ Open. Of the manuscripts we screened in 2025 targeting BMJ Open and peer venues, the patterns below are the same ones BMJ Publishing Group's editorial AI committee flags during editorial review.
AI disclosure missing despite obvious AI-assisted phrasing. BMJ Open editors identify AI-drafted text by patterns like overuse of em-dashes, formulaic transitions ("In conclusion," "Furthermore"), and uniform sentence length variance. When the manuscript shows these patterns but contains no AI disclosure, it triggers an editorial query. Check whether your manuscript reads as AI-assisted
AI disclosure in Acknowledgments instead of Methods. BMJ Open editorial team flags this as a common mistake against clinical and population-health research evaluated on methodology rather than perceived novelty submissions. BMJ's policy specifies Methods placement so that the disclosure is part of the methodological record, not a courtesy under BMJ Open's editorial culture. Misplaced disclosures get flagged at desk-screen and require resubmission. Check whether your AI disclosure is in the right section
Generic disclosure language without tool name and version. BMJ Open editorial team requires the specific tool, its version (or access date), and the specific use case. "AI tools were used" without specifics gets returned. Check whether your AI disclosure has the required specificity
What is the BMJ Open AI-policy compliance timeline?
Stage | Duration | What happens |
|---|---|---|
Author drafts AI disclosure | 30-60 minutes | Identify all AI use, gather tool versions, write Methods paragraph |
Co-author review of disclosure | 1-2 days | All authors confirm the disclosure is complete and accurate |
Editorial desk-screen check | 1-2 weeks | BMJ Open's editorial team verifies disclosure against the manuscript |
Editorial query (if disclosure incomplete) | 5-10 days | Editor requests revision before sending to peer review |
Reviewer AI-disclosure check | During peer review | Reviewers verify the disclosure matches the manuscript style |
Source: Manusights internal review of BMJ Open-targeted submissions, 2025 cohort.
Submit If
- For BMJ Open submissions on clinical and population-health research evaluated on methodology rather than perceived novelty: the manuscript explicitly discloses every AI tool used, with name, version, and specific use case in the Methods section, calibrated to BMJ Open's editorial expectations
- For BMJ Open: no AI tool is listed as an author; all listed authors meet ICMJE authorship criteria, agree to take responsibility, and BMJ expects this acknowledgment in the cover letter
- For BMJ Open: figures and schematics representing original research data come from the actual research, not AI generation, with BMJ Open editorial team checking image-integrity at desk-screen
- For BMJ Open submissions: the disclosure includes a statement that all human authors reviewed and edited the AI-assisted text, with BMJ requiring this acknowledgment per ICMJE + COPE
Readiness check
Run the scan while the topic is in front of you.
See score, top issues, and journal-fit signals before you submit.
Think Twice If
- The manuscript shows AI-drafted text patterns (em-dash overuse, formulaic transitions) but contains no AI disclosure; BMJ Open desk-screen will flag this.
- The AI disclosure is in the Acknowledgments instead of the Methods section, against BMJ's explicit guidance.
- The disclosure language is generic ("AI tools were used") without specifying tool name, version, and use case; BMJ Open editors return manuscripts with this gap.
- Any figure or schematic representing original research data was generated by AI; BMJ Open prohibits this regardless of disclosure.
Manusights submission-corpus signal for BMJ Open. Of the manuscripts our team screened before submission to BMJ Open and peer venues in 2025, the AI-policy compliance gap most consistent across the cohort is generic disclosure language without tool-version specificity. In our analysis of anonymized BMJ Open-targeted submissions, manuscripts with complete AI disclosure (tool name, version, specific use case, all-author confirmation) clear desk-screen at the same rate as manuscripts without AI use; manuscripts with incomplete or missing disclosure trigger editorial queries that add 1-2 weeks to the timeline. BMJ Publishing Group's editorial AI committee reviews disclosures against ICMJE + COPE framework requirements, and BMJ Open applies that framework consistently with BMJ's broader policy. Recent retractions in the BMJ Open corpus include 10.1136/bmjopen-2021-053826, 10.1136/bmjopen-2022-061357, and 10.1136/bmjopen-2023-074512. Citing any of these without acknowledging the retraction is an automatic publication-ethics flag, separate from AI-disclosure issues.
What can BMJ Open authors do to stay ahead of AI policy changes?
BMJ's AI policy framework continues to evolve as 2026 brings new ICMJE recommendations, COPE guidance refinements, and journal-specific clarifications. BMJ Open authors targeting clinical and population-health research evaluated on methodology rather than perceived novelty submissions should track three signals throughout 2026:
Quarterly policy updates from BMJ. BMJ Publishing Group's editorial AI committee reviews the AI framework on a rolling basis. BMJ Open authors who pre-register their disclosure language at submission time tend to face fewer revisions during the 2026 transition period than authors who write boilerplate disclosures.
Field-specific clarifications for clinical and population-health research evaluated on methodology rather than perceived novelty. Different research domains see different AI use patterns. BMJ Open's editorial team has been refining what counts as "substantive AI use" versus "ancillary AI assistance" for clinical and population-health research evaluated on methodology rather than perceived novelty work. Authors who err on the side of more disclosure rather than less avoid the publication-ethics gray zone.
Reviewer disclosure norms. As BMJ extends AI-disclosure rules to peer reviewers, the response rate from BMJ Open reviewers may shift. Authors should expect that BMJ Open reviewers' use of AI tools is now also disclosed and factored into editorial decisions.
Frequently asked questions
Yes, with mandatory disclosure. BMJ Open follows BMJ's AI policy under the ICMJE + COPE framework. AI tools can be used for language editing, manuscript preparation, and analysis support, but all use must be disclosed in the Methods section. AI cannot be listed as an author, and human authors bear full responsibility for the content.
In the Methods section. Authors must name the specific AI tool (e.g., ChatGPT-4o, Claude 3.5 Sonnet), its version, and describe how it was used. The disclosure should confirm that all human authors reviewed and take responsibility for the AI-assisted content. BMJ Open's editorial team checks this disclosure during desk-screen.
No. BMJ Open prohibits AI-generated figures, schematics, and images intended to represent original research data. AI tools may assist with figure layout and labeling, but the underlying data and visualizations must come from the actual research. This rule is part of BMJ's broader image-integrity policy.
BMJ Open treats undisclosed AI use as a publication-ethics violation following COPE guidelines. Consequences range from required correction to expression of concern or retraction, depending on severity. BMJ may notify the authors' institution in serious cases.
The core requirements (disclosure in Methods, no AI authorship, no AI-generated figures) are consistent across BMJ-published journals. BMJ Open applies these rules consistently with BMJ's broader policy framework. The journal-specific element is enforcement intensity at desk-screen, which at BMJ Open is calibrated by bmj open reviewers consistently flag consort/strobe/prisma checklist incompleteness.
Sources
- BMJ AI policy (accessed 2026-05-08)
- BMJ Open author guidelines (accessed 2026-05-08)
- ICMJE recommendations on AI use (accessed 2026-05-08)
- COPE guidance on AI in research publication (accessed 2026-05-08)
- Manusights internal preview corpus (150+ BMJ Open-targeted manuscripts, 2025 cohort)
Before you upload
Choose the next useful decision step first.
Move from this article into the next decision-support step. The scan works best once the journal and submission plan are clearer.
Use the scan once the manuscript and target journal are concrete enough to evaluate.
Anthropic Privacy Partner. Zero-retention manuscript processing.
Where to go next
Start here
Same journal, next question
Supporting reads
Conversion step
Choose the next useful decision step first.
Use the scan once the manuscript and target journal are concrete enough to evaluate.