Skip to content
BENIBIX Learn

EU AI Act

Also known as AI Act, European AI Act, Artificial Intelligence Act

The European Union's AI law. It sorts AI by risk and sets rules for each level.

Think of it like

It is like a safety stamp on toys. But it is for AI that is sold in Europe.

Example

A company sells a tool in Germany that scores job seekers. Under the EU AI Act, that is high risk. So it must keep logs, test its data, and let a person overrule it.

Why it matters

It applies to anyone who sells AI to people in the EU, wherever the firm is based. So many firms around the world follow it.

Definition

The EU AI Act is the European Union's law for AI, and the first broad law of its kind. It came into force in August 2024, and its rules start in stages over several years. It sorts uses of AI into four levels of risk. A few uses are banned, such as unfair social scoring, by a state or a firm. High-risk uses need strict checks, chatbots and deepfakes need clear labels, and the rest are mostly free. Breaking the bans can cost a firm up to 35 million euros or 7% of its yearly world sales, whichever is higher.
Four levels: banned, high risk, limited risk with labels, and minimal risk.

Common mistake

Thinking it only covers firms in Europe. It covers AI used in the EU, even if the maker is somewhere else.

Key takeaways

  1. 1
    Four risk levels. Banned to minimal.
  2. 2
    In stages. From 2024 on.
  3. 3
    Reaches worldwide. If sold in the EU.