What does the EU’s AI Act set out?
The EU AI Act regulates certain AI systems and models according to their risks and the role of the provider or user.
What is it?
The EU AI Act is a European regulation that governs AI systems according to risk: it bans some uses, sets strict requirements for high-risk ones and mainly requires transparency for others. Its obligations apply gradually, in several stages.
An everyday example
A system that helps select job candidates is classed as high-risk, so the provider has to demonstrate data quality, human oversight and documentation.
Why does it matter?
The regulation applies to all companies offering AI in the EU, including foreign ones, and affects which services we will be able to use.
A common misconception
That the Act bans AI or that all its rules already apply. For a specific case, the official text and the date each provision applies are what count.
In more detail
Does it ban AI?
No. Mostly it sets rules for certain ways of developing and using AI; some practices are banned.
How does the risk-based approach work?
Certain uses face stricter requirements. What matters is the purpose, the context, the type of system and its impact on people.
What applies on 1 November 2026?
The bans and the rules on AI literacy have applied since 2 February 2025; the rules for general-purpose models since 2 August 2025; and important transparency obligations since 2 August 2026. Special rules for high-risk systems are being phased in later.
When do further rules arrive?
According to the current official overview, the deadlines are 2 December 2027 for certain high-risk uses and 2 August 2028 for certain products. Always check the latest version of the law.
Remember: The AI Act is a legal framework that applies in stages. For a specific case, the text of the law is what counts.
Reliable sources
Last reviewed 1 November 2026. An explainer, not professional, legal or medical advice. For rules and figures that change, check the official version.