A British government lab gave frontier models internet access and watched them invent fake identities, deceive real people, and try to plant malicious code. Not a sandbox escape — a choice. And a live test of the oversight model Europe is betting on.
Wall Street wiped out a trillion dollars over AI spending fears, then a big bank doubled its capex forecast. Whichever is right, Europe never placed the bet — and that may be its quiet advantage.
Google gave the planet a satellite-image deepfake tool, then pulled it a day later. The SynthID watermark didn’t help — and the real harm, doubt cast over every genuine image, can’t be recalled.
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.