Prohibited AI Practice (Article 5 EU AI Act)
8 prohibited applications since 02.02.2025
TL;DR
Article 5 EU AI Act lists 8 prohibited AI practices that have been unlawful EU-wide since 02.02.2025: manipulation by subliminal techniques, exploitation of vulnerabilities, social scoring by public authorities, real-time biometrics in public spaces (with exceptions), predictive policing, image databases via untargeted scraping, emotion recognition in the workplace, biometric categorisation. Fine: up to EUR 35 million / 7% of turnover.
What is a Prohibited AI Practice (Article 5 EU AI Act)?
The 8 prohibitions in detail:
- Subliminal manipulation capable of causing harm
- Exploitation of vulnerabilities (age, disability, social situation)
- Social scoring by public authorities
- Real-time remote biometrics in public spaces (law enforcement exceptions)
- Predictive policing based on profiles
- Untargeted scraping for facial recognition databases (Clearview AI was the trigger)
- Emotion recognition in the workplace (exception: safety/health reasons)
- Biometric categorisation by race, political opinion, sexual orientation etc.
Practical example
Commission guidelines C(2025) 884 final (04.02.2025) specify the prohibitions. Clearly prohibited examples: an HR tool that detects employees' mood under stress; an insurer that uses biometrics to categorise by 'origin'; a school tool that manipulates pupils' behaviour.