🟡🟡🟡: US limits use of Anthropic's Fable 5 and Mythos
Semafor Technology · Semafor Technology · 2026-06-13
A White House directive restricts Anthropic's Fable 5 and Mythos models to US nationals only, a move Anthropic publicly contested and which could make frontier AI development unprofitable or legally untenable given the prevalence of foreign nationals at major labs.
Appears in
Extraction
Topics: ai-regulationexport-controlsfrontier-ai-safetyanthropicjailbreaking
Claims
- The White House directed Anthropic to limit access to Fable 5 and Mythos to US nationals, including restricting access for Anthropic's own foreign-national employees.
- Anthropic publicly disagreed with the government's decision in a blog post, declining to comment further.
- The government claims to have found a jailbreak in Fable 5 that bypasses guardrails, which Anthropic says is not serious enough to warrant the action taken.
- If applied broadly, the directive could make developing frontier AI models unprofitable or illegal given how many foreign nationals work at major labs.
- Anthropic's own history of lobbying the government to take AI dangers seriously may have contributed to the government's aggressive response.
Key quotes
The impact of the new rule: It will be unprofitable to develop frontier AI models, and potentially even illegal to develop them in the first place, given how many foreigners work on these models at all the major frontier labs.
Anthropic will have to take some responsibility for that overreaction. It has been the leading lobbyist for the government taking AI dangers seriously. The government was listening.
The truth is that no AI model has ever avoided being jailbroken.