Skip to content
BENIBIX Learn

Prohibited AI Practices

Also known as Banned AI Uses, Unacceptable Risk AI, EU AI Act Article 5

The uses of AI the EU AI Act bans outright. Social scoring is one.

Think of it like

It is like the list of things you may never take on a plane. No permit makes them okay.

Example

A shop wants cameras that read how its staff feel during a shift. In the EU, that is a banned use of emotion reading at work.

Why it matters

The law judged these uses too harmful to fix with rules. So nothing can make them legal.

Definition

Prohibited AI practices are the AI uses banned outright by Article 5 of the EU AI Act. They include harmful tricks, and using a person's age, illness or being poor against them. Unfair social scores are banned, and so is guessing who will break the law from a profile. Making face files from the web or cameras is out. So is reading feelings at work or school, except for health or safety. The bans apply since 2 February 2025, and fines go up to 35 million euros or 7% of world sales.
The bans came first, on 2 February 2025, before any other rule in the Act.

Common mistake

Thinking a ban only hits the maker. Using a banned system breaks the law too, not only selling one.

Key takeaways

  1. 1
    Banned outright. No checks can make it legal.
  2. 2
    A short list. All of it in Article 5.
  3. 3
    Huge fines. Up to 7% of world sales.