Europe Passes the World's First Comprehensive Law on Artificial Intelligence

The European Parliament approved the AI Act by a wide margin. It bans some uses outright, imposes strict rules on 'high-risk' ones like hiring and policing, and requires AI-generated content to be labeled. Most of it takes effect over the next two years.

The European Parliament voted 523 to 46 this afternoon to approve the Artificial Intelligence Act, the first broad law anywhere governing the technology. It has been three years in the making — it was proposed before ChatGPT existed and rewritten after — and it will apply to any company selling AI products to the European Union's 450 million people, wherever the company is based.

The law sorts AI by risk. Some uses are banned: social scoring of citizens, systems that manipulate people's behavior, most real-time facial recognition by police, and reading emotions in workplaces and schools. "High-risk" uses — AI in hiring, credit decisions, education, medical devices, critical infrastructure — are allowed but must meet requirements for testing, documentation, and human oversight. The most powerful general-purpose programs, the ones behind chatbots, must disclose what they were trained on and meet safety standards. And AI-generated images, audio, and video must be labeled as such.

Companies that break the rules can be fined up to 7 percent of their worldwide revenue. The bans take effect in six months; most of the rest phases in over two years.

Reaction split predictably. Civil-liberties groups say the police exemptions are too wide; technology companies say the rules will slow European innovation and push it elsewhere. The law's supporters point to Europe's privacy rules, which the rest of the world complained about and then largely copied. That is the bet: that the first rulebook becomes the template.