Cancer Cell's AI Policy: Cell Press Rules for Oncology Authors

Cancer Cell uses STAR Methods, which has a specific structure. Your AI disclosure goes in the Method Details subsection, not in Acknowledgments or a separate AI statement.

Proper disclosure format:

"During the preparation of this manuscript, the authors used ChatGPT (GPT-4, OpenAI) to improve the clarity of the Discussion and Introduction sections. The authors also used GitHub Copilot (Microsoft) to assist with writing Python scripts for the RNA-seq differential expression analysis. All AI-generated text was reviewed and revised by the authors, and all computational code was independently validated against known datasets before application to experimental data. The authors take full responsibility for the content of this article."

What makes this work for Cancer Cell specifically:

• It separates language editing from computational assistance, important for a journal where both are common
• It confirms that bioinformatics code was validated independently
• It explicitly states human review of all AI outputs
• It names specific tools and versions

What wouldn't work:

"The authors acknowledge the use of AI during manuscript preparation."

This is too vague for any Cell Press journal, but it's especially inadequate for Cancer Cell, where reviewers and editors need to understand whether AI touched the scientific content versus the prose.

Disclosure for common Cancer Cell submission types

Cancer Cell publishes several article types, each with different AI disclosure considerations:

Research Articles (original data): Disclose AI use in language editing and computational code. Be explicit that experimental design, data collection, and biological interpretation were human-driven.

Resource papers: If your resource involves AI/ML as the research subject (e.g., an AI model for cancer detection), that's different from using AI to write the paper. Clearly separate the AI that's the subject of research from AI used in manuscript preparation.

Reviews and Perspectives: AI can help organize literature and structure arguments, but Cancer Cell's editors expect original intellectual contribution. Disclose AI assistance in literature organization if applicable.

Clinical and Translational papers: Same heightened considerations as Nature Medicine. Don't use AI to interpret clinical outcomes, and be explicit in your disclosure that clinical data wasn't AI-processed.

The gene pathway diagram point deserves attention. Cancer Cell papers frequently include pathway schematics, signaling cascade diagrams, and graphical abstracts. If you used any generative AI tool to create these, even as a starting point that you then modified, Cell Press's policy requires disclosure, and purely AI-generated versions aren't acceptable.

Cell Press's escalation process for undisclosed AI use:

During review:

• Editor flags concern and contacts corresponding author
• Authors asked to add disclosure to STAR Methods
• If AI use appears to have affected scientific claims, additional review may be required
• Deliberate concealment can lead to rejection

After publication:

• Minor cases (language editing only): Published correction adding disclosure to STAR Methods
• Moderate cases (undisclosed AI in analysis code): Expression of concern while the editorial office investigates whether results are affected
• Serious cases (AI-generated data or fabricated content): Retraction, with notification to the authors' institution and funding agencies

For Cancer Cell specifically, the oncology research community is tight-knit. A correction or retraction at Cancer Cell circulates fast. Researchers in the field watch Retraction Watch and PubPeer, and an integrity issue at this journal affects your reputation disproportionately because the readership is concentrated among the leaders in your subfield.

Real-world example of what can go wrong: Imagine you use an LLM to help write the clinical implications section of a paper on a novel immunotherapy target. The AI generates a sentence claiming "this target has shown preliminary efficacy in Phase I trials" based on its training data, but your paper is about preclinical work. If that sentence survives editing and makes it into the published paper, it's a factual error that could mislead clinicians. Cancer Cell's editors would issue a correction, and the corresponding author would need to explain how a false clinical claim ended up in their paper.

For bioinformatics-heavy papers:

• If AI helped write analysis code, include the validated code in your GitHub repository or supplementary materials. Cancer Cell increasingly expects code availability, and reviewers may test it.
• Separate AI-written code from human-written code in your documentation. This makes the disclosure clearer and the code more auditable.
• Don't use AI to generate synthetic datasets for method validation unless you disclose this explicitly and explain why real data wasn't available.

For papers with clinical implications:

• Keep AI tools away from clinical interpretation sections. Reviewers at Cancer Cell include practicing oncologists who will catch AI-generated clinical language.
• If your paper includes treatment recommendations or patient stratification claims, these must be entirely human-authored and data-driven.

For all submissions:

• Draft your STAR Methods AI disclosure during writing, not after submission
• Circulate the disclosure to all co-authors before submission, Cancer Cell papers often have 15+ co-authors, and surprises about AI use create problems
• If you didn't use AI, you don't need to say so. There's no requirement for a negative disclosure statement.
• Check your graphical abstract. If you made it with Biorender or Illustrator, you're fine. If you used Midjourney or DALL-E, replace it.

Before submission checklist for Cancer Cell:

• [ ] AI disclosure in STAR Methods → Method Details
• [ ] Tool names, versions, and use cases specified
• [ ] No AI-generated images in figures or graphical abstract
• [ ] Bioinformatics code independently validated
• [ ] Clinical claims written by human investigators, not AI
• [ ] All co-authors have reviewed the AI disclosure

A free manuscript assessment can help you confirm that your Cancer Cell submission meets the journal's formatting and ethical requirements before you submit.