Prohibited AI: what's absolutely forbidden?
The AI Act bans certain AI applications outright: social scoring, subliminal manipulation, emotion recognition in the workplace and scraping faces for databases. The fine: up to 7% of annual turnover.
Article 5 of the AI Act describes AI practices considered so harmful that they are completely prohibited in the EU. These bans have been in force since 2 February 2025.
Prohibited AI includes: (a) subliminal manipulation that distorts behaviour, (b) exploitation of vulnerable groups, (c) social scoring, (d) profiling-based risk scoring for criminal behaviour, (e) scraping faces for databases, (f) emotion recognition in workplace/education, (g) biometric categorisation based on protected attributes.
For webshops the most relevant cases are: dark patterns that use AI to nudge you unconsciously towards a purchase, and marketing that specifically targets vulnerable groups (elderly people with phishing-style offers, children with addictive gamification).
The fine for prohibited AI is the heaviest: up to €35 million or 7% of global annual turnover, whichever is higher.
Want to know if this applies to your website?
Our scanner automatically checks whether your website meets these AI Act requirements and gives concrete recommendations.
Scan your website