EU AI Act enforcement begins, granting regulators powers

The European Union's landmark Artificial Intelligence Act enters its most consequential enforcement phase today, August 2, 2026, granting regulators sweeping new powers to oversee the world's most advanced AI models. The move cements Brussels' position as the globe's foremost AI regulator and initiates a complex, high-stakes effort to govern a fast-evolving technology.

The AI Act, passed in 2024, adopts a risk-based approach, imposing stricter obligations on systems deemed to pose higher risks to citizens' rights, health, or safety. It explicitly bans certain AI applications, including predictive policing, emotion recognition systems in workplaces and schools, and technology designed to manipulate human behavior. Starting today, new transparency requirements take effect: AI-generated content such as deepfakes must be clearly labeled, and companies must ensure users know they are interacting with an AI system like a chatbot. Existing AI systems have until December 2, 2026, to comply, with exemptions for artistic and creative work.

Enforcement authority rests with the newly expanded European AI Office, a regulatory body composed of dozens of tech experts, lawyers, and economists....