Prove you're American

When the Commerce Department pulled Anthropic's two newest models offline in mid-June, barring every foreign national from them, the company complied the only way it could: it shut both off for everyone on Earth. That was the earlier chapter. This is how it ends. After roughly two weeks fully dark, a June 26 letter from Commerce Secretary Lutnick un-froze exactly one of them — Mythos 5, the restricted cyber-defense model — and only for a private annex of US firms. Fable 5, the model anyone could pay to use three days before the shut-off, stayed offline.

The everyday consumer model stayed the most restricted of the two — the cyber model came back first.

The resolution is the news, because of how Anthropic is engineering the way back. Rather than wait for the government to lift its foreign-national ban, the company is building a path to satisfy it: a privacy policy effective July 8 adds a 'Verification Data' category that collects an image of a government-issued ID, a photo or video of your face, and facial-geometry templates Anthropic concedes 'may be considered biometric data.' The unstated logic is that if Anthropic can prove a user is a US citizen, it can serve them without violating the directive. The check applies to consumer accounts — Free, Pro, Max — not to enterprise or API customers.

So the everyday-consumer model came back last and the cyber-weapon model came back first, to a secret list. And the mechanism that restores a chatbot to ordinary Americans is a biometric citizenship gate bolted onto the login — a frontier lab scanning faces to enforce an export rule the public still can't read. Anthropic has argued the underlying standard 'would essentially halt all new model deployments for all frontier model providers,' which is less a defense than an admission of how brittle the deployment regime has become.