Immunity's AI Policy: Cell Press Rules for Immunology's Top Journal

Immunity is a Cell Press journal, and Cell Press is owned by Elsevier. Two layers of policy exist:

Elsevier's company-wide position covers all ~2,800 Elsevier journals: no AI authorship, mandatory disclosure, author responsibility. Elsevier doesn't mandate a specific disclosure location.

Cell Press adds specificity: STAR Methods placement is required, not optional. Cell Press provides example disclosure language. The editorial team screens actively during review.

Immunity follows Cell Press exactly. No journal-specific modifications. If you know the Cell Press policy, you know Immunity's rules.

This two-layer structure means Immunity's policy is marginally more prescriptive than what you'd find at a non-Cell-Press Elsevier journal. The STAR Methods requirement adds structure that a general Elsevier journal doesn't impose.

High-dimensional immune profiling

Immunity publishes cutting-edge immunology that increasingly relies on high-dimensional data: CyTOF (mass cytometry), spectral flow cytometry, CITE-seq, spatial transcriptomics, and multiplexed imaging. These analyses involve complex computational pipelines where AI tool use is becoming common.

The key distinction:

• Research analysis tools (FlowJo, Scanpy, Seurat, CATALYST, CytofRUV): These are described in your standard STAR Methods as part of your research methodology. They don't fall under the AI manuscript preparation policy.

• AI writing/code tools (ChatGPT, Copilot, Claude): If these helped you write analysis scripts, generate code, or edit manuscript text, they require disclosure in STAR Methods under Method Details.

Example of clear separation:

"CyTOF data was preprocessed and analyzed using CATALYST (R/Bioconductor) with clustering performed using FlowSOM (see STAR Methods: CyTOF Analysis). GitHub Copilot (Microsoft) was used to assist with writing custom R scripts for the differential abundance analysis between patient groups. All code was validated by the bioinformatics team (S.K. and L.M.) against manually computed results."

Immune cell illustrations and pathway diagrams

Immunity papers frequently include:

• T cell activation pathway diagrams
• Cytokine signaling network illustrations
• Immune cell differentiation schematics
• Tissue microenvironment cartoons

The rules are clear:

• BioRender, Illustrator, PowerPoint: Fine, no AI disclosure needed
• Midjourney, DALL-E, Stable Diffusion: Prohibited for generating these images
• Hybrid (AI-generated, then refined): Still prohibited

If you aren't sure whether your illustration tool counts as generative AI, stick with BioRender. It's designed for scientific illustration and doesn't use generative AI to create images.

Vaccine and therapeutic immunology

Immunity publishes basic science that informs vaccine design and immunotherapy development. If your paper connects fundamental immune mechanisms to therapeutic applications, keep AI away from the translational implication sections. These are where your expertise matters most, and AI-generated claims about therapeutic potential can mislead readers.

For a mechanistic immunology paper:

"During preparation of this manuscript, the authors used ChatGPT (GPT-4, OpenAI) to improve the clarity of the Introduction and Discussion sections. The experimental results, data interpretation, and mechanistic conclusions were generated entirely by the research team. All AI-suggested text edits were reviewed by the corresponding author (A.M.) and senior author (R.S.). The authors take full responsibility for the published content."

For a paper with extensive computational analysis:

"Single-cell RNA-seq data was analyzed using Scanpy (v1.9) and scvi-tools as described in STAR Methods: scRNA-seq Analysis. ChatGPT (GPT-4, OpenAI) was used to improve the language of the Results section. GitHub Copilot (Microsoft) assisted with writing Python scripts for custom trajectory analysis. All analysis code was validated against published reference datasets (see Data and Code Availability). The authors take full responsibility for the content."

For a short paper (Immunity Reports):

"The authors used Claude (Claude 3.5, Anthropic) to edit the manuscript for language clarity. All content was reviewed by both authors. The authors take full responsibility for the published content."

The reviewer response letter point is an edge case. Cell Press doesn't explicitly require disclosure of AI use in revision response letters (these aren't part of the published manuscript). But if AI helped you rewrite significant portions of the manuscript during revision, update the STAR Methods disclosure to reflect this.

Cell Press formalized its AI policy in early 2023, shortly after ChatGPT's widespread adoption:

The Cell Press policy has been more stable than some competitors. Science (AAAS) initially banned all AI-generated text before switching to a disclosure model. Cell Press started with disclosure from the beginning and hasn't changed course. This consistency is helpful for immunology labs running long-term projects, the rules you follow today are the same rules you'll face when you submit next year.

Cell Press enforcement follows the standard process:

During review:

• Editor contacts corresponding author
• Disclosure must be added to STAR Methods
• Deliberate concealment can lead to rejection

After publication:

• Correction for minor language editing non-disclosure
• Expression of concern if AI affected data interpretation
• Retraction for fabricated content or false claims
• COPE investigation for systematic issues

Immunity's community dynamics: Immunity publishes approximately 150-200 research articles per year. The journal's editorial board and reviewer pool overlap substantially with Nature Immunology and JEM. A publication ethics flag at Immunity doesn't stay contained, it becomes known across all three major immunology journals.

The STAR Methods distinction means Immunity and Cell have the same formatted disclosure structure, while Nature Immunology and JEM use free-form Methods sections. If you're preparing manuscripts for multiple journals simultaneously, note which format you need, STAR Methods has specific subsections (Key Resources Table, Resource Availability, Experimental Model and Study Participant Details, Method Details) that don't apply at non-Cell-Press journals.

For high-dimensional data papers:

• Separate research software (Seurat, FlowJo, CATALYST) from AI writing tools in your STAR Methods
• If AI helped with dimensionality reduction visualization code, say so, but clarify that the biological interpretation of clusters was done by the investigators
• Deposit all analysis code in a public repository

For mechanistic studies:

• AI can polish your writing, but the mechanistic model or signaling cascade interpretation should be your own
• Don't use AI to generate hypothetical mechanisms in the Discussion, Immunity's reviewers are domain experts who will recognize generic AI-generated immune mechanism descriptions

For translational immunology:

• Keep AI away from sections discussing therapeutic implications
• If your paper bridges basic immunology with clinical applications, the clinical relevance assessment should come from the investigators, not an LLM

For all submission types:

• Draft the STAR Methods AI disclosure during writing, not after
• Make sure all co-authors review the disclosure
• If you used AI during revisions, update the disclosure before resubmission
• Check your graphical abstract, BioRender is safe; generative AI isn't

Before submission checklist:

• [ ] AI disclosure in STAR Methods → Method Details
• [ ] Tool names, versions, and specific use cases listed
• [ ] Research tools in standard STAR Methods sections (not AI disclosure)
• [ ] No generative AI images or graphical abstract
• [ ] Analysis code validated and deposited
• [ ] All co-authors reviewed AI disclosure
• [ ] Mechanistic interpretations are human-generated

A free manuscript assessment can help verify your Immunity submission meets Cell Press standards before you submit.