Unacceptable risk is the EU AI Act’s most severe regulatory category. It covers AI applications considered incompatible with fundamental rights and therefore prohibited rather than merely monitored or restricted. section 10.30
Prohibited examples include:
- Cognitive or behavioral manipulation targeting people, especially vulnerable groups. section 10.30
- Social-scoring systems. section 10.30
- Real-time and remote biometric identification, with some law-enforcement exceptions. section 10.30
A violent-behavior toy
A voice-activated toy that encourages children to behave violently would fall into the unacceptable-risk category and be banned. section 10.30
Figure 10.7 The four risk categories in the AI Act
Use the four-category overview to compare unacceptable risk with high-, limited-, and minimal-risk AI systems. section 10.30
