HO-010GemiddeldAanbieder & GebruikerDetectief
Beoordeling van doeltreffendheid van toezicht
Aanbieders en gebruikers moeten periodiek de doeltreffendheid van menselijke toezichtsmaatregelen beoordelen, inclusief de vraag of de aangewezen toezichthouders zinvol kunnen ingrijpen en of de toezichtsinstrumenten geschikt blijven voor hun doel.
Artikelen:Article 14(1)Article 72
Bewijs voorbeelden
- Oversight effectiveness review report
- Oversight drill exercise results
- Corrective actions from review findings
Normen
ISO 42001:2023 §9.3
Gerelateerde controles
- HO-001Human Oversight Mechanism DesignArticle 14(1) requires that high-risk AI systems are designed and developed in a way that enables natural persons to effectively oversee the system's…
- HO-002Override and Intervention CapabilityHigh-risk AI systems must include a capability for human overseers to intervene in or interrupt the system's operation, and providers must document this…
- HO-003Stop Function ImplementationProviders must implement a stop function that enables the immediate cessation of the AI system's operation when required by human overseers, with a clear,…
- HO-004Automation Bias Awareness TrainingArticle 14(4) requires that oversight persons are aware of the tendency to over-rely on AI outputs (automation bias); deployers must implement training…
- HO-005Output Interpretation CapabilityDeployers must ensure that persons assigned to oversee high-risk AI systems have the capability to correctly interpret the system's outputs, including the…
- HO-006Decision Reversal and Correction ProcessDeployers must establish documented procedures for reversing or correcting AI-assisted decisions that affect natural persons, including how affected…
Gerelateerde AI-verordening-begrippen
- Post-Market MonitoringProactive collection and review of experience gained from the use of high-risk AI systems, to identify any need for corrective or preventive actions and ensure continued compliance throughout the system lifecycle.
- Human OversightMeasures built into high-risk AI systems enabling natural persons to understand, monitor, and — where necessary — override or shut down the system. Must be proportionate to the risks and must ensure that deployers can intervene in the system's output before it takes effect.
Upgrade wanneer het juridisch verdedigbaar moet zijn.
De gratis vragenlijst geeft voorlopige signalen. De Volledige Beoordeling zet echte systeeminformatie om in een controleerbaar besluitdossier: extractie, componentscheiding, bewijs, nationale wetgeving en een auditklaar dossier.
Start gratis risicovoorbeeld