AI clinical-safety check.
If your product uses AI to inform clinical decisions, the audit includes a spot-check of how its clinical-safety claims compare with the evidence behind them. It is a review of claims and evidence, not a test of your model.
SCOPE
What it checks
- Whether performance claims match the validation evidence you cite
- Whether the validation population matches the people the product is meant for
- Whether limitations and failure modes are disclosed next to the claims
- Whether claims about clinician oversight match how the product is described to work
What it does not certify
- That the product is clinically safe
- That the model performs as described: I do not test or re-run your model
- Regulatory clearance or compliance
- Any outcome: no finding is promised