How does AI Control assess AI risk?

The AI Control Risk Engine combines weighted factors — identity, capability, context, behaviour, communication, target, environment, history, policy and consequence — into an explainable score from 0 to 100 and a level of LOW, MEDIUM, HIGH or CRITICAL. The score is an input to policy and human oversight, not a perfect prediction, and it never grants authority by itself.

Frequently asked questions

Can a low risk score approve an action?

No. A low score means fewer restrictions from risk-based policies, but identity, capability and policy must still be satisfied.

Are risk scores explainable?

Yes. Each decision stores the factor breakdown so reviewers can see why a score was high.