Anthropic Suspends Advanced AI Models After U.S. Security Order

Government Order Forces Model Shutdown

Anthropic announced it must immediately disable access to its most advanced AI models, Fable 5 and Mythos 5, for all users worldwide following a June 12 government order. The directive requires the AI company to suspend access for foreign nationals due to national security concerns, though the order provided limited specifics about the threat, according to a company statement first reported by USA Today.

The company believes the government's concern centers on a potential jailbreaking method for Fable 5. Jailbreaking allows users to circumvent safety features designed to prevent AI systems from completing sensitive or dangerous tasks.

Anthropic stated it has received only verbal evidence of what it describes as a "narrow, non-universal jailbreak" that involves asking the model to analyze a specific codebase and identify software vulnerabilities. The company launched Claude Fable 5 earlier this week as part of a new capability tier it calls "Mythos-class."

Why It Matters

This marks the first known instance of the U.S. government ordering a complete suspension of commercial AI models over security concerns. The precedent could reshape how frontier AI companies approach model releases and government oversight, particularly as advanced systems gain capabilities in cybersecurity domains. The incident also highlights growing tensions between commercial AI development timelines and national security considerations.

Cybersecurity Risks and Safety Guardrails

Mythos-class models represent a significant capability leap that experts warn could accelerate sophisticated cyberattacks if misused. Banking and other sectors relying on complex, legacy technology systems face particular vulnerability. Anthropic built guardrails into Fable 5 to prevent use in high-risk areas like cybersecurity, though some users have criticized these restrictions as overly broad.

The company expressed disagreement with the government's approach, arguing that a narrow potential jailbreak should not justify recalling a commercial model accessible to hundreds of millions of users. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers," Anthropic stated.

Escalating Government Tensions

The suspension order arrives amid deteriorating relations between Anthropic and federal authorities. Earlier this year, the company declined to allow U.S. military use of its AI models for domestic surveillance and fully autonomous weapons systems. In response, the government placed Anthropic on a supply chain blacklist scheduled to take effect later in 2026.

Pentagon Chief Information Officer Kirsten Davies defended the government's position in a social media post, stating the Defense Department fully supports prioritizing national security. "Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation," Davies wrote, adding "America First. Always."

Anthropic confidentially filed for a U.S. initial public offering in May, positioning itself ahead of competitor OpenAI in reaching public markets.

Details were first reported by USA Today, with contributions from Reuters.