Regulating AI

The EU AI Act, passed in 2024 and phasing in through 2027, is the world's first comprehensive AI law — and the template others are copying. Its core idea: regulate by risk tier, not by technology.

• Unacceptable — banned outright: government social scoring, systems that manipulate people into harm, most real-time biometric ID in public spaces
• High-risk — allowed under strict rules: AI in hiring, credit scoring, medical devices, policing, admissions. Requires quality data, documentation, and human oversight
• Limited risk — transparency duties only: chatbots must say they're AI; deepfakes must be labeled
• Minimal — no new rules: spam filters, game AI, most recommenders