The EU AI Act entered into force in August 2024 and phases in over 2 years.
**Risk tiers:**
- **Unacceptable risk** (banned): social scoring by governments, real-time biometric surveillance in public
- **High risk**: AI in hiring, credit scoring, critical infrastructure, medical devices — strict requirements
- **Limited risk**: chatbots must disclose they're AI
- **Minimal risk**: most AI products — no restrictions
**GPAI rules** (General Purpose AI): Foundation models like GPT and Claude face transparency requirements. Systemic risk models (above compute threshold) face additional auditing.
**Penalties:** Up to €35M or 7% of global revenue.
**The global impact:** Companies often comply globally, so the EU law effectively shapes products worldwide.