Law

EU AI Act

The European Union’s law regulating artificial intelligence by risk level, in force since August 2024 and applying in stages until 2030.

Last reviewed · 4 sources · Suggest a correction

Contents
  1. What it is
  2. The four levels
  3. The 2026 changes
  4. What it means for you
  5. Timeline
  6. Sources

What it is

The EU AI Act is the world’s first broad law for artificial intelligence. Instead of regulating the technology as a whole, it sets rules based on how risky a use is.

The four levels

  • Banned. Uses the EU considers unacceptable, such as social scoring by governments. These bans have applied since February 2, 2025.
  • High-risk. AI used in areas like hiring, education, credit and critical products must meet strict requirements. After the 2026 Digital Omnibus, these apply from December 2, 2027 for stand-alone systems and August 2, 2028 for AI in regulated products.
  • Transparency. People must be told when they are talking to an AI or looking at a deepfake. These duties have applied since August 2, 2026.
  • Minimal risk. Most everyday AI, like spam filters, has no new obligations.

The 2026 changes

The Digital Omnibus on AI (Regulation (EU) 2026/1744) entered into force on July 27, 2026. It delays the high-risk deadlines rather than removing them, and adds new bans on AI built to produce child sexual abuse material or non-consensual intimate imagery, scheduled to apply from December 2, 2026.

What it means for you

If you live in the EU, you should increasingly see labels on AI-generated content and chatbots. Outside the EU, many global apps adopt the same labels rather than build separate versions.

Timeline

  1. The EU AI Act enters into force.

  2. The AI Act’s bans on prohibited AI practices start to apply.

  3. Obligations for general-purpose AI models start to apply.

  4. The Digital Omnibus on AI enters into force, delaying high-risk deadlines.

  5. Transparency duties (Article 50), including disclosure of deepfakes and AI chatbots, apply.

  6. New bans on AI that generates child sexual abuse material or non-consensual intimate imagery are scheduled to apply.

  7. High-risk rules for stand-alone systems (Annex III), such as hiring, education and credit, are scheduled to apply.

  8. High-risk rules for AI in regulated products (Annex I), such as toys and medical devices, are scheduled to apply.

Sources

  1. Regulation (EU) 2024/1689 (Artificial Intelligence Act), EUR-Lex
  2. Digital Omnibus on AI, artificialintelligenceact.eu
  3. EU AI Act Update: Digital Omnibus Finalizes 8 Compliance Changes, Orrick
  4. EU AI Act Omnibus Agreement: Postponed High-Risk Deadlines, Gibson Dunn

Every fact on this page comes from the sources above. Entries are re-checked when things change. Spotted something out of date? Tell us. Read our editorial policy.