AI Governance Explorer

Scientific integrity and Test, Evaluation, Verification & Validation considerations documented.

Approaches and metrics for measuring AI risks are selected.

Test sets, metrics and methodology details documented for evaluation.

AI system is regularly evaluated for validity, reliability and performance.

AI system is evaluated for safety risks.

Security and resilience including adversarial robustness evaluated.

Appropriate level of accuracy, robustness and cybersecurity throughout lifecycle.

Model evaluations, systemic risk assessment & mitigation, incident reporting, cybersecurity.