Enforcement is live. Europe can now require an independent look inside a frontier model — just as this summer’s sandbox escapes at OpenAI and Anthropic reveal how little control their own makers have.
On 2 August 2026 Brussels stops asking nicely. The AI Office can finally fine, evaluate and withdraw AI models — and a second, quieter clock on chatbot and deepfake disclosure starts the same day.
The EU AI Omnibus deal of May 7 delayed high-risk AI compliance to December 2027. But Brussels has not slowed down. On May 8, the Commission opened consultation on AI transparency obligations. On May 19, it sought feedback on high-risk AI classification guidelines. Italy has already passed its own national AI law. Ireland has established a national AI office. The EU AI Act is not waiting for 2027 — it is being built right now, one guideline at a time. gafam.ai explains what changed, what is coming, and what every GAFAM company operating in Europe must do today.