European Heart Journal's AI Policy: ESC Rules for Cardiology's Top Journal

Cardiac imaging and AI research

EHJ publishes substantial research on AI-assisted cardiac imaging: echocardiographic analysis, cardiac MRI segmentation, CT calcium scoring algorithms, nuclear cardiology AI applications. If your paper is about an AI tool for cardiac imaging, that AI is your research subject. Describe it in your standard Methods section as methodology.

The manuscript preparation AI policy is separate. If you developed a deep learning model for left ventricular ejection fraction estimation and also used Claude to polish your Discussion section, these are two distinct items:

• The research AI goes in standard Methods (model architecture, training data, validation approach)
• The writing AI goes in the AI disclosure statement

Don't combine them. Reviewers and editors need to distinguish between AI as a research tool and AI as a writing tool.

Clinical trial reports

EHJ publishes major cardiovascular clinical trials, studies on anticoagulation strategies, heart failure drugs, device therapies, and interventional procedures. These papers directly influence ESC guideline updates and clinical practice across Europe. AI involvement in interpreting trial endpoints, describing adverse events, or drawing clinical conclusions isn't acceptable under EHJ's framework.

You can use AI to clean up your prose and improve readability. You shouldn't use it to draft the sections where clinical claims are made. There's a meaningful difference between "AI helped me express my findings more clearly" and "AI helped me figure out what my findings mean."

ESC guideline papers

EHJ is the primary publication venue for ESC Clinical Practice Guidelines. These documents directly determine how millions of patients across Europe are treated. The AI stakes here are as high as they get in cardiology publishing. While the formal policy doesn't have separate rules for guidelines versus original research, the editorial sensitivity is much higher. If AI involvement in a guideline paper came to light after publication, the credibility damage would extend well beyond the authors.

Here's what a proper disclosure looks like for different types of EHJ submissions.

For an original research article (e.g., observational cohort study):

"During the preparation of this manuscript, the authors used ChatGPT (GPT-4, OpenAI) to improve the clarity and readability of the Introduction and Discussion sections. All AI-generated suggestions were reviewed and edited by the corresponding author (A.B.) and the senior biostatistician (C.D.). No AI tools were used in study design, data collection, statistical analysis, or interpretation of results. The authors take full responsibility for the content of this article."

For an imaging/AI research paper:

"The deep learning model described in this study (EchoNet-LV) was developed using PyTorch and trained on the institutional echocardiographic dataset described in Methods. Separately, during manuscript preparation, the authors used Claude (Claude 3.5, Anthropic) to edit the Results and Discussion sections for language clarity. The research methodology and the manuscript editing tool are entirely separate systems. All AI-suggested text edits were reviewed by all authors. The authors take full responsibility for the published content."

For a meta-analysis:

"The authors used ChatGPT (GPT-4o, OpenAI) to improve the readability of the Results section and to check the consistency of numerical values between tables and text. All outputs were verified against the original data extraction spreadsheets by two reviewers (E.F. and G.H.). The systematic search strategy, study selection, data extraction, quality assessment, and statistical analyses were performed entirely by the authors without AI assistance."

These examples share three features: they name the tool and version, they specify what was and wasn't done with AI, and they state that authors take responsibility. Don't skip any of these elements.

The consequences of undisclosed AI use at EHJ follow a standard escalation path that's consistent with COPE guidelines:

Stage 1: Discovery during peer review. If a reviewer or editor suspects undisclosed AI use during the review process, the editor will contact the corresponding author and request clarification. This doesn't automatically mean rejection, but it does mean your paper gets flagged for closer scrutiny. The editorial team may run the manuscript through AI detection tools, though they'd be the first to acknowledge these tools aren't perfectly reliable.

Stage 2: Discovery after acceptance but before publication. If AI use comes to light during production, the paper can be pulled before publication. The authors would need to provide a satisfactory disclosure statement and explain why it wasn't included originally. This delays publication and creates an uncomfortable record in the editorial system.

Stage 3: Discovery after publication. This is where consequences get serious. EHJ follows COPE flowcharts for post-publication integrity concerns. Options include:

• A published correction adding the AI disclosure
• An expression of concern if the scope of AI use calls the paper's integrity into question
• Retraction if AI involvement was extensive and materially affected the scientific content

Stage 4: Institutional notification. In serious cases, the journal may notify the authors' institution. For researchers at European academic medical centers, this can trigger a formal research integrity investigation. That's a process that can take months and has career implications far beyond the single paper.

The reality is that most AI disclosure omissions aren't malicious, authors simply didn't realize the policy applied to their use case, or they forgot. But "I didn't know" isn't a defense that editorial boards find compelling in 2026.

If you're considering where to submit your cardiology paper, understanding the AI policy landscape across the field's top journals helps you prepare correctly the first time.

A few things stand out in this comparison:

Circulation's dual-disclosure requirement is unique among the top five. AHA journals require AI disclosure in both the Methods section and the cover letter. EHJ, JACC, and Heart don't require the cover letter mention, though there's nothing stopping you from including it.

JACC's position reflects the ACC's partnership with Elsevier. The American College of Cardiology sets clinical policy, while Elsevier provides the publishing infrastructure. The AI rules come primarily from the ACC side but are implemented within Elsevier's systems.

Heart follows BMJ Publishing Group rules, which are consistent across the BMJ portfolio. If you've submitted to The BMJ, Gut, or Thorax, you'll find the AI policy familiar.

All five journals prohibit AI authorship and AI-generated images. These are the two non-negotiable rules across the entire cardiology journal landscape. There isn't a major cardiology journal that allows either.

Before you start writing

Decide upfront what role AI tools will play in your manuscript. Will you use ChatGPT for language editing? GitHub Copilot for statistical code? Nothing at all? Making this decision early means you can track your AI use as you go rather than trying to reconstruct it months later when you're filling out the submission form.

During writing

If you're using AI tools, keep a log. It doesn't need to be elaborate, a simple note like "Used Claude to edit Discussion paragraphs 3-5 on March 12" is enough. This log will make your disclosure statement accurate and specific rather than vague and generic.

The formatting details

EHJ uses ScholarOne for manuscript submission. The submission system may include specific questions about AI use in the manuscript preparation workflow. Don't skip these or answer them reflexively. They're part of the formal record.

Common mistakes to avoid

Mistake 1: Disclosing too vaguely. "AI tools were used during manuscript preparation" tells the editor nothing. Name the tool, name the version, describe the scope.

Mistake 2: Forgetting about co-authors' AI use. You didn't use AI, but your third co-author used Grammarly's AI features to edit the Methods section they drafted. The disclosure requirement covers all AI use on the manuscript, not just yours.

Mistake 3: Assuming Grammarly doesn't count. Standard grammar checking is exempt. But Grammarly's AI-powered rewriting and paraphrasing features cross the line into generative AI territory. If the tool is generating new text rather than flagging errors, it likely requires disclosure.

Mistake 4: Not distinguishing research AI from writing AI. If your paper uses machine learning for cardiac image analysis, that's research methodology. If you also used ChatGPT to write the paper, that's manuscript preparation AI. These are separate disclosures with separate purposes. Don't combine them into one confusing paragraph.

EHJ's readership is global, and many authors aren't native English speakers. The ESC is a European society with members from countries where English isn't the primary language. The AI policy is explicitly designed to allow language improvement, that's the core permitted use case.

If English isn't your first language, you should feel comfortable using AI tools to improve your grammar, sentence structure, and overall readability. That's exactly what the policy permits. Just disclose it honestly. There's no stigma attached to using AI for language editing; it's the same function that professional language editing services have provided for decades, just faster and cheaper.

What you shouldn't do is use AI to generate scientific arguments, interpret your data, or draft clinical conclusions. The line between "improving how I express my findings" and "generating findings for me to express" is the critical distinction.

A free manuscript assessment can help ensure your manuscript meets EHJ's editorial standards and that your AI disclosure statement is properly formatted before you submit.