Fable 5 SHUT DOWN: Why the US Government just pulled the plug on Anthropic’s top AI.
On June 12, 2026, the US government issued an emergency directive to suspend access to Fable 5 and Mythos 5 immediately.
The order cites "national security authorities," forcing Anthropic to disable these models for all users—including their own foreign national employees.
Why the sudden ban? The government claims to have found a "jailbreak" method that bypasses Fable 5’s strict safety protocols.
Anthropic disagrees. They’ve reviewed the "jailbreak" and say it only reveals minor, well-known vulnerabilities found in almost every AI model today.
Before launch, Fable underwent thousands of hours of "red-teaming" with the UK AISI and US government to ensure industry-leading safety.
Anthropic admits "perfect" jailbreak resistance is impossible. Instead, they use a "Defense in Depth" strategy to catch and shut down attacks.
The government's evidence? A "narrow" exploit where the AI identifies flaws in a codebase—a capability common in models like GPT-5.5.
Anthropic warns that if this narrow standard is applied to everyone, it could halt all new AI deployments across the entire industry.
We believe the government should have the power to block unsafe AI, but through a process that is transparent and fair.
Anthropic is complying with the law but working to clear up this "misunderstanding" to restore access for hundreds of millions of users.
Is this a necessary safety measure or government overreach? Read the full statement and stay updated on the Fable 5 crisis.