Human Oversight
Requiring a human to supervise or step in for important AI decisions.
Explainer
Regulators often require that a human be able to supervise, intervene, or override AI decisions that significantly affect people. This is especially true for high-risk AI.
What this means for you
For important decisions you can usually ask for a human to review what the AI did.
Related regulations
-
EU AI Act
Germany · Enacted
-
NIST AI RMF
United States · Guidance
-
UK GDPR Art. 22
United Kingdom · Enacted
-
Australian AI Ethics
Australia · Voluntary