The EU AI Act: Europe's law on artificial intelligence
The EU AI Act is the European law that regulates artificial intelligence systems according to the risk they pose. It entered into force on August 1, 2024 and applies in stages through 2028.
Data as of October 22, 2026
Key figures
| Level | Beginner |
|---|---|
| Origin | 2024 |
| Introduced by | Union européenne, règlement (UE) 2024/1689, entré en vigueur le 1er août 2024 |
A law built around risk levels
Regulation (EU) 2024/1689, known as the AI Act, doesn’t treat all AI the same way. The more harm a use could cause, the heavier the obligations:
- Unacceptable risk: banned practices, such as social scoring of citizens or manipulation that exploits vulnerable people.
- High risk: AI used to hire people, grant loans, grade students, run critical infrastructure and so on. These systems will have to prove data quality, limit bias, keep logs and allow human oversight.
- Transparency: a chatbot must tell you it’s a machine; deepfakes and certain texts published to inform the public must be disclosed.
- Minimal risk: the vast majority of uses (spam filters, video games), with no specific obligations.
Makers of general-purpose AI models, the ones powering the big assistants, also have their own documentation duties.
The timeline, step by step
According to the European Commission’s official page, checked on October 22, 2026:
| Date | What applies |
|---|---|
| August 1, 2024 | The regulation enters into force |
| February 2, 2025 | Bans on unacceptable-risk practices |
| August 2, 2025 | Rules for general-purpose AI models and governance |
| August 2, 2026 | General application, including transparency rules |
| December 2, 2026 | Deadline for technical marking of generated content; new ban on AI creating non-consensual intimate images |
| December 2, 2027 | High-risk systems (hiring, education, biometrics…) |
| August 2, 2028 | High-risk systems built into regulated products |
The last dates come from a simplification regulation, known as the « omnibus », that entered into force on July 27, 2026. The timeline could still change, so always check the official text.
What it changes for you
Day to day, the most visible effect is transparency: knowing when you’re talking to an AI and spotting generated content more easily, notably through watermarking. If you apply for a job or a loan in the EU, the systems assessing you will face tighter rules.
The AI Act doesn’t replace the GDPR: your rights over your personal data stay the same.
Common misconceptions
- « The AI Act bans ChatGPT in Europe. » No, it regulates its use without banning it.
- « Everything has applied since 2024. » No, it phases in gradually through 2028.
- « It only affects European companies. » Wrong: any company offering an AI system in the EU is covered.
Related terms
- GDPR and AI · The EU's General Data Protection Regulation, which also applies to AI tools that process personal data.
- AI watermarking · A mark, often invisible, embedded in an image, audio, video or text to signal that it was generated by AI.
- Deepfake · An AI-made or AI-altered image, video or voice recording that makes a real person appear to say or do something they never did.
- AI ethics · The principles for designing and using AI in a way that is fair, transparent and respectful of people and their rights.
- AI bias · A systematic error that makes an AI treat some groups unfairly or repeat stereotypes, often inherited from its training data.
- Chatbot · A program you talk to in writing or by voice, which replies in everyday language like a conversation partner.
Full glossary
Frequently asked questions
Does the AI Act apply to ChatGPT or Gemini?
Yes, as long as these services are offered in the European Union. Their makers are covered by the rules on general-purpose AI models, and chatbots must make it clear that you're talking to a machine.
Which AI practices are banned?
The law bans, among other things, social scoring of individuals, manipulative techniques that exploit people's vulnerabilities, and emotion recognition at work and in schools, with some exceptions. These bans have applied since February 2, 2025.
Have the rules been delayed?
Partly. A simplification regulation that entered into force on July 27, 2026 pushed the high-risk obligations back to December 2, 2027, and to August 2, 2028 for systems built into regulated products. The transparency rules were not postponed as a whole.
Does it matter if my company is in the US?
It can. The AI Act applies to any provider that places an AI system on the EU market, wherever it is based. If you serve EU users, it's worth checking which category your tools fall into.
Data sources
Data as of October 22, 2026
Report an error
Does a figure, a date or a fact look wrong? Tell us: every report is checked.