When to apply When manuscripts, grant applications, or repository deposits need to carry a structured disclosure of generative-AI involvement that goes beyond free-text "we used ChatGPT" sentences.
Before you start
Prerequisites
What needs to be in place before you operationalise Generative AI use and disclosure terminology in your CRIS or repository.
- Editorial or institutional policy that distinguishes disclosable AI use from background tools (spell-check, reference managers)
- A controlled vocabulary for AI-use categories — writing assistance, language polishing, code generation, image generation, analysis, ideation
- Agreement that LLMs are not authors (ICMJE, COPE, WAME alignment) and a structured way to capture model and version separately
- A submission system or deposit form capable of carrying repeatable structured statements, not just a free-text box
- Reviewer-side workflow for handling absent or evasive disclosure
Deployment
Five steps to deploy
Each step is small enough to land in a single sprint or a single sitting with the relevant CRIS administrator. Follow in order.
Adopt a structured disclosure schema
Each AI-use declaration carries: ai_system_name, ai_system_version, ai_system_provider, use_category (closed list), affected_sections (closed list — abstract, methods, results, code, figures, all), human_oversight_description, prompt_archive_url (optional). Multiple declarations per work are normal.
Bake the closed-list into the submission UI
Replace the free-text "did you use AI" box with a multi-row table the depositor fills row-per-system. Each row uses the controlled use_category and affected_sections lists. Provide examples of disclosable versus non-disclosable use in inline help.
Emit machine-readable disclosure
JATS: a structured ai-disclosure block or custom-meta tags; JSON-LD: a schema:ContributorRole-style entity per declaration; Crossref: relationship type and ai-disclosure narrative in the metadata payload. Avoid burying disclosures inside the article PDF only.
Add reviewer-side gating
When the disclosure block is empty, surface a checkbox: "Confirm: no generative-AI tools were used in the preparation of this manuscript beyond background grammar-and-spell-check". This forces a positive affirmation rather than silent absence.
Pilot with a single journal or department
Run for one editorial cycle. Measure: percentage of submissions with any declaration, the distribution across use_categories, the number of reviewer-requested re-disclosures. Iterate the controlled list before broad rollout.
Worked example
Sample workflow
A realistic walk-through of a single record passing through the Generative AI use and disclosure pipeline once the checklist is in production.
Integration points
CRIS and repository systems
Vendor-specific notes on where this vocabulary fits in real research-information systems. Names appear here only where there is public field evidence — they are not vendor partnerships.
Add the disclosure block via a custom plugin or the metadata fields configuration; surface it on the published article view template.
Both accept structured questions on submission; build the closed-list disclosure as a configured submission step.
Add AI-disclosure fields to the deposit form via the configurable submission, mapped to a custom Schema.org-aligned metadata schema.
Use a custom metadata template on the Output entity; surface on the Pure Portal as a publicly visible disclosure block.
Crossref accepts AI-disclosure narrative as part of the metadata deposit; align with their recommended fields rather than inventing a local format.
What goes wrong in the field
Common pitfalls
The patterns that show up repeatedly when this checklist is skipped or misapplied. Address these before they become entrenched.
- Treating a one-line "we used ChatGPT" sentence as adequate disclosure — it carries no machine-readable affordance
- Disclosing the model family ("a large language model") without the version, so the disclosure cannot be audited or reproduced
- Listing the AI system as an author or co-author, contradicting ICMJE and COPE policy
- Hiding the disclosure inside the PDF only, so meta-analysis and downstream indexers cannot detect it
- Letting the controlled list of use categories drift between journals in the same publisher portfolio, breaking cross-corpus analysis
Frequently asked
Implementation FAQ
- Who maintains this checklist?
- The Generative AI use and disclosure working group maintains the checklist alongside the dictionary terms in the same domain. It is reviewed each release cycle (March and September) and updated when a working-group consultation, a vendor product change, or a federation-partner schema update materially changes the operational guidance.
- What if my CRIS or repository is not listed?
- The integration points listed name the systems CASRAI has direct field experience with — Pure, Symplectic Elements, Worktribe, Converis, DSpace and DSpace-CRIS, EPrints, VIVO, Dataverse, Invenio-RDM. The CERIF mapping in the checklist is vendor-neutral and applies equally to other CRIS or repository products. If your system supports the underlying entities (Person, Project, Output, Funding, plus the domain-specific extensions), the steps transfer.
- How do I validate my implementation?
- Three validation surfaces. First, the deposit form should refuse a record missing required fields rather than warn and accept. Second, the resulting metadata should round-trip through the federation layer your institution uses (OpenAIRE Guidelines 4.0 for European federation, DataCite Commons for DOI-anchored discovery, Crossref for article-anchored discovery) without upstream errors. Third, walk a real-world record through the sample-workflow path on this page and confirm the structured fields capture what the prose describes.
- Where do I report errors in the checklist?
- Open a comment via the dictionary-feedback flow at /dictionary/contribute. Editorial corrections — wrong vendor module names, deprecated standards, broken integration paths — are queued into the next release cycle. Substantive disagreements on the operational guidance are routed to the working group for review and may motivate a checklist revision.
- Is this checklist enough to certify my implementation?
- No. The checklist gives you the operational baseline; certification against federation profiles (CoreTrustSeal, OpenAIRE-compliant, COAR-aligned) is a separate process with its own audit. Treat the checklist as the engineering scaffolding and the certification as the institutional sign-off that the scaffolding is being used.
