Anthropic's Most Powerful AI Was Released, Then Halted Over a Foreign-User Ban

An AI model once kept from general release because it was considered powerful enough to excel at hacking and potentially escape containment was finally released, only to be pulled back from public use almost immediately. The reversal happened within an extremely short time window.

The video captures the pace bluntly: it was released, and then almost at once it was made private again. The change happened fast.

The trigger for the withdrawal was US government intervention. According to the video, the government told Anthropic that it could not allow foreign nationals to use Fable, whether they were inside or outside the United States.

That policy creates two obvious problems. First, a nontrivial number of the Anthropic employees who built Fable are themselves foreign nationals. Second, enforcing model access restrictions on the basis of nationality would be operationally extremely difficult in practice.

Because drawing that nationality-based line was not realistic, Anthropic chose not a limited restriction but a full disablement for all users. In effect, its most powerful model was placed entirely out of reach.

Reaction to the episode has been divided. The video says Grok, a rival model, welcomed the development and sarcastically commented that AI models should not be left unconstrained and instead need to be held back.