Commerce Department Threatens Anthropic With Criminal Charges

The U.S. Commerce Department issued an export control directive on June 12, 2026 ordering Anthropic to suspend all access to Fable 5 and Mythos 5 for any foreign national, whether inside or outside the United States, including Anthropic's own non-citizen employees, per Anthropic's public statement. Anthropic received the directive at 5:21pm ET and, given its scope, disabled both models for all customers to ensure compliance; access to other Anthropic models including Claude Opus 4.8 was unaffected. Fable 5 launched June 9 alongside Mythos 5 - Fable is a consumer-facing model built on the Mythos base model, whose advanced cybersecurity capabilities were the primary national security concern.

Commerce Secretary Lutnick named the order per PYMNTS citing Bloomberg. Fox Business reported that Amazon and five other companies testing Fable 5 discovered a jailbreak that opened the full cybersecurity capabilities of Mythos, alarming administration officials. A senior administration official told Fox Business the government had signaled to Anthropic before the launch that Fable 5 would be a test case for a recently signed AI guardrails executive order, and that Anthropic's response to pre-launch concerns created a "lack of trust." The official characterized the company's handling as "recklessness." The Trump administration has a contentious history with Anthropic: the DoD declared it a "supply chain risk" in March 2026 after Anthropic refused contract terms permitting models to be used for autonomous weapons systems; the company is challenging that designation in federal court.

Anthropic called the disclosed jailbreak "narrow" and "non-universal," stating it unlocks cybersecurity capabilities only in one specific instance and that the same vulnerability exists in other publicly available models, including OpenAI's GPT-5.5, per Anthropic's statement. The company wrote: "We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers." Senior technical staff were sent to Washington for White House meetings; Anthropic said it believed the action was "a misunderstanding" and was working to restore access.

Anthropic filed confidentially for an IPO at a $965 billion valuation in early June 2026, per Fortune; the forced withdrawal of its two newest frontier models adds regulatory uncertainty to that offering. Industry observers note this is the first known instance of the U.S. government using export controls to force a commercial AI model recall over a reported jailbreak, setting a potential precedent for how frontier model vulnerabilities are handled. If the standard is applied broadly, it could pressure all frontier-model providers to implement stricter pre-launch government clearance processes.

• •Clarification from Commerce on the legal basis and scope of the export control, including any license pathway for foreign nationals.
• •Whether other frontier-model providers receive similar directives or proactively alter access controls.
• •Technical disclosures on the jailbreak method and any Anthropic mitigations that could restore Fable 5 and Mythos 5 access.
• •Anthropic's IPO timeline and investor response to the government's regulatory posture.