The US government has forced AI company Anthropic to completely disable its most powerful artificial intelligence models after ordering the company to block foreign nationals from accessing them. The Commerce Department directive, citing national security concerns, has effectively taken Anthropic’s newest AI systems offline just days after their release.
Anthropic said the government instruction forced it to suspend all access to Fable 5 and Mythos 5 for everyone, not just foreign nationals, because the company cannot technically separate users by nationality. The directive affects both foreign nationals working at Anthropic and international customers worldwide.
The suspension marks an unprecedented government intervention in AI model deployment and signals growing federal concern about advanced AI capabilities falling into foreign hands. Commerce Secretary Howard Lutnick issued the directive with help from the Bureau of Industry and Security, highlighting the administration’s willingness to use export control authorities to restrict AI access.
Anthropic had released both models just Tuesday, calling them the most powerful AI systems the company had ever created. The timing of the government action suggests officials were closely monitoring the release and quickly determined the models posed national security risks.
The two suspended models represent different approaches to AI safety and access:
- Fable 5 was released to the general public with strong restrictions on cybersecurity and biology-related questions
- Mythos 5 went to select trusted partners including cybersecurity and infrastructure companies without safety guardrails
- Both models share the same underlying technical foundation but differ in their safety restrictions
When announcing the models Tuesday, Anthropic itself acknowledged the security risks, stating that “Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage” without proper safeguards. This admission may have contributed to the government’s rapid response.
The directive represents a significant escalation in government oversight of AI development. Previous regulatory discussions have focused on voluntary safety commitments and future legislation, but this action shows the administration is prepared to use existing export control laws to immediately restrict AI access when national security concerns arise.
For the broader AI industry, the suspension sets a new precedent. Other companies developing similarly powerful models now face the possibility of government intervention if their systems are deemed too capable or risky for international access. This could fundamentally change how AI companies approach model releases and international business operations.
The action also raises questions about the balance between innovation and security in AI development. While the US seeks to maintain its AI leadership, overly restrictive measures could push international partners and customers toward competitors in other countries that may have fewer restrictions.
