The US government issued an emergency export control directive ordering Anthropic to immediately suspend access to its two most advanced AI models, Claude Fable 5 and Claude Mythos 5. The order cites national security concerns regarding a potential jailbreak vulnerability that could allow foreign nationals to bypass safety guardrails.
The directive bars all foreign nationals from accessing the models, compelling Anthropic to disable them for its entire customer base to ensure compliance. While the government identified a method to jailbreak the publicly available Fable 5, Anthropic disputed the severity of the finding. The company argued the vulnerability is relatively simple and already replicable using other publicly available models, including OpenAI’s GPT-5.5.
Anthropic warned that this directive sets a dangerous precedent. If applied industry-wide, such standards could effectively halt all new frontier AI model deployments. The company stated it is working to restore access as soon as possible, noting that other Anthropic models remain unaffected.
David Sacks, co-chair of the President’s Council of Advisers on Science and Technology, commented on the situation via social media. He claimed a trusted partner discovered the jailbreak and that Anthropic CEO Dario Amodei refused to fix the issue or de-deploy the model. Sacks criticized Anthropic for prioritizing consumer offerings over safety, contradicting the firm's public stance on AI regulation.
This incident follows earlier tensions between Anthropic and the administration. The AI firm previously refused to sign agreements allowing mass domestic surveillance and autonomous lethal weapons systems, leading President Donald Trump to label Anthropic a supply chain risk. Although recent reports suggested a cooling of hostilities as the government sought access to Claude Mythos, this latest order marks a significant escalation in regulatory enforcement.