Task suitability check

A first-pass checklist for deciding whether a biomedical research task is appropriate for generative AI.

In development: this page is public to support transparent iteration and review.

Resource typeGuidanceStatusdraftLast reviewed2026-06-22ContributorsRaquel ManzanoTagsdecision support, governance, workflow

Core screening questions

1. Does the task involve sensitive, identifiable or otherwise restricted data?
2. Is the output consequential enough that unsupported hallucinations would create scientific, ethical or operational harm?
3. Is there an approved local workflow or tool for this use case?
4. Can a competent human verify the output before it is used or shared?

Intended use

This should become a short operational screen that helps staff distinguish:

• low-risk drafting or summarisation tasks
• tasks that require caution and explicit documentation
• tasks that should not proceed without approved infrastructure or governance review