Prohibited AI Practices: Article 5 Explained
Understanding which AI practices are completely banned under the EU AI Act and how to ensure your systems comply.
Prohibited AI Practices: Article 5 Explained
Article 5 of the EU AI Act completely bans certain AI practices that violate fundamental rights.
Prohibited Practices
1. Social Scoring
AI systems that evaluate or classify people based on social behavior or personality traits are prohibited. This includes social credit scoring systems.
2. Manipulative Techniques
AI systems using subliminal techniques or manipulative/deceptive methods that distort behavior in a harmful way are banned.
3. Real-Time Remote Biometric Identification
Live facial recognition and other real-time biometric identification in public spaces is prohibited, except for specific law enforcement purposes with authorization.
4. Emotion Recognition in Sensitive Contexts
Using emotion recognition in the workplace or educational institutions is prohibited.
5. Biometric Categorization
Categorizing people based on biometric data related to race, religion, sexual orientation, or other protected characteristics is banned.
Exceptions
Limited exceptions exist for:
- Law enforcement with proper authorization
- National security purposes
- Research and development (with safeguards)
Compliance
If your system uses any prohibited practice, you must:
- Immediately cease the practice
- Redesign the system to remove prohibited functionality
- Ensure compliance before deployment
Assessment
Use EUAIActReady's assessment tool to check if your system uses any prohibited practices.