The banned practices
Article 5 prohibits a short, closed list: manipulative or exploitative techniques causing significant harm, social scoring, untargeted scraping of facial images, emotion inference at work or school, certain biometric categorisation, predictive policing based solely on profiling, real-time remote biometric identification in public for law enforcement — and, added in 2026, AI-generated non-consensual intimate imagery and child sexual abuse material.
Why it is like this
These are the uses the legislator decided no safeguard could make acceptable, so they are removed from the risk-management framework entirely rather than conditioned.
Not to be confused with
A general ban on surveillance or on emotion recognition. Emotion inference is prohibited at work and in education, and permitted elsewhere subject to transparency; the biometric identification ban is about real-time use in public by law enforcement.
In practice
A wellbeing tool inferring employee mood from webcam video is prohibited outright — consent does not rescue it, because the ban is not conditioned on a lawful basis. The two newest prohibitions were added by Regulation (EU) 2026/1744 and apply from 2 December 2026.
Source: Regulation (EU) 2024/1689 (AI Act) — Article 5. Read on .
Odd one out Check yourself
Three of these are prohibited under Article 5. Which one is not?
Show the answer
Using a chatbot without telling the user it is a machine
Undisclosed chatbots are a transparency failure under Article 50 — limited risk, not prohibited. The penalty band is also different.