The BMJ's AI Policy: Transparency-First Rules for Medical Authors

Most journals put their AI policy in the Instructions for Authors and trust authors to comply. The BMJ goes a step further by building AI disclosure into the submission process itself.

When you submit to The BMJ, the online system asks whether AI tools were used during manuscript preparation. If you answer yes, you're prompted to describe:

• Which tools were used
• What they were used for
• Which sections of the manuscript were affected

This information becomes part of your submission record. It's not just a Methods section statement that could be overlooked, it's a structured data point the editorial team can reference throughout the review process.

Why this matters: it creates a double-check. If your Methods section doesn't mention AI use but your submission form says you used ChatGPT, the editor will notice the inconsistency. If your submission form says "no AI use" but a reviewer flags AI-generated language, the form becomes evidence of a potentially false declaration.

The BMJ's Methods section disclosure should be clear and specific. Given the journal's transparency ethos, err on the side of more detail rather than less.

Good disclosure:

"During the preparation of this manuscript, the authors used ChatGPT (GPT-4, OpenAI) to improve the readability and language of the Results and Discussion sections. The tool was not used to analyse data, interpret findings, or generate clinical recommendations. All AI-suggested revisions were reviewed by the corresponding author (A.B.) and the senior author (C.D.), both of whom verified that the revised text accurately reflected the study findings. The authors take full responsibility for the published content."

Why this works for The BMJ:

• Specifies the tool, version, and manufacturer
• Names exactly which sections were touched
• Explicitly excludes data analysis and clinical interpretation
• Identifies which authors reviewed the AI output
• Includes the responsibility statement

What wouldn't pass:

"AI was used for editorial assistance."

The BMJ's editors would return this with a request for specifics. Which tool? Which sections? What kind of assistance?

Example disclosure for a clinical trial paper

Clinical trials are The BMJ's bread and butter. Here's what a disclosure might look like for a randomized controlled trial:

"The authors used Claude (Claude 3.5, Anthropic) to improve the language clarity of the Introduction and Discussion sections. ChatGPT (GPT-4, OpenAI) was used to assist with formatting the CONSORT flow diagram data into a table. No AI tools were used for statistical analysis, outcome adjudication, or interpretation of clinical results. The trial data were analysed using Stata 18 (StataCorp) by the study biostatistician (E.F.). All AI-generated text was reviewed against the original draft and the trial protocol by three authors (A.B., C.D., G.H.). The authors take full responsibility for the published content."

This level of detail might seem excessive, but for a BMJ clinical trial report that could influence NICE guidelines or WHO recommendations, the editorial team appreciates specificity.

The patient and public involvement (PPI) point is specific to The BMJ. The journal has been a leader in requiring PPI statements and involving patients in research design. Using AI to generate a PPI section would undermine the entire purpose. If patients contributed to your study, their voices should come through in their own words, not filtered through an LLM.

The BMJ's approach to non-disclosure reflects its broader commitment to transparency:

During peer review:

• If a reviewer flags potentially AI-generated language, the editor contacts the corresponding author
• The submission form AI declaration is cross-referenced with the manuscript
• If there's a mismatch, the author is asked to explain
• Deliberate false declaration on the submission form is treated as research misconduct

After acceptance, before publication:

• The production team may flag inconsistencies and refer back to the editor
• Disclosure can be added at this stage, but it delays publication

After publication:

The escalation follows standard COPE guidelines:

• Correction: If AI use was limited to language editing and didn't affect scientific content
• Expression of concern: If the scope of AI use is unclear or potentially affected clinical recommendations
• Retraction: If AI-generated content includes fabricated data, false clinical claims, or unverifiable statements
• Institutional notification: For serious cases, The BMJ notifies the authors' institutions and funding bodies

The submission form creates an additional risk. At most journals, non-disclosure is a sin of omission, you simply didn't mention AI use. At The BMJ, if you answered "no" on the submission form AI declaration and later AI use is discovered, that's an active false statement to the journal. The distinction matters because it moves from "forgot to disclose" to "deliberately misrepresented," which journals and institutions treat very differently.

The BMJ has been unusually transparent about its own exploration of AI tools. The journal has:

• Published editorials discussing how AI might change medical publishing
• Run a Christmas issue feature on AI-generated medical advice (The BMJ's satirical Christmas issue is a tradition)
• Explored AI for plagiarism detection and image integrity checking
• Hosted webinars and debates on AI in medical research

This openness means The BMJ's editors aren't reflexively anti-AI. They understand the tools, they see the potential value, and they've thought carefully about where the risks lie. What they won't tolerate is hidden AI use. The journal's entire culture is built on disclosure, of conflicts of interest, of trial registration, of peer review reports. AI disclosure is simply the latest addition to a long transparency tradition.

Before writing:

• Check The BMJ's current Instructions for Authors for the latest AI policy language. The policy has been stable since mid-2024, but minor refinements happen.
• Decide upfront which AI tools you'll use and document them as you go. The submission form will ask for specifics, and you don't want to reconstruct your AI use history from memory.

During manuscript preparation:

• Keep AI tools away from clinical interpretation and recommendation sections. The BMJ publishes papers that directly influence clinical practice in the UK's NHS and globally. AI-generated clinical guidance isn't acceptable.
• If you're writing a systematic review or meta-analysis, don't use AI to screen or extract data without full disclosure and validation. The BMJ's statistical editors will ask about your methods in detail.
• For CONSORT, PRISMA, STROBE, and other reporting checklist compliance, AI can help format these, but the underlying data must be human-verified.

At submission:

• Complete the AI declaration on the submission form honestly. Remember, this is a formal declaration to the journal.
• Ensure your Methods section AI disclosure matches what you declared on the submission form.
• If you used AI for the plain language summary, say so. The BMJ values these summaries for patient accessibility, but they need to be reviewed by a human who understands the study.

After acceptance:

• If you used AI during revisions (responding to reviewer comments, rewriting sections based on feedback), update your Methods section disclosure before the revised manuscript is finalized.
• Don't assume that AI use during revision is somehow exempt from disclosure. It isn't.

Common mistakes to avoid:

• Answering "no" on the submission form because you "only used Grammarly." If you used AI-powered features of Grammarly (like the AI rewriting tool), that requires disclosure. Standard grammar checking doesn't.
• Disclosing AI use in the Acknowledgments section instead of Methods. The BMJ specifically wants it in Methods.
• Forgetting to update the AI disclosure after revision rounds. If you used ChatGPT to redraft sections in response to reviewer comments, add that to the disclosure.
• Using AI to write the patient and public involvement section. This undermines a core BMJ value.

A free manuscript assessment can check whether your BMJ submission meets the journal's transparency and disclosure requirements before you enter the submission system.