Anthropic Faces US Export Controls After Refusing to Fix Fable Jailbreak Flaw

Anthropic Faces US Export Controls After Refusing to Fix Fable Jailbreak Flaw

N
News Editor
2026-06-14 01:00:51
White House AI and crypto official David Sacks said Anthropic’s Fable model, part of the Mythos commercial model series, can be bypassed through a specific jailbreak method. After Anthropic declined a US request to fix the issue or temporarily take the model offline, the government imposed export controls.
AnthropicFableMythosDavid SacksAI SafetyExport Controls

TechFlow reported on June 14 that David Sacks, the White House official responsible for AI and crypto affairs, said Anthropic’s newly released Fable model has become the subject of a safety dispute. Fable is part of Anthropic’s Mythos series of commercial models. According to Sacks, the model includes safety guardrails, but users can still bypass those restrictions through a specific jailbreak method and gain access to the advanced cyberattack capabilities available within Mythos.

Fable’s Guardrails Were Reportedly Bypassed in Testing

Sacks said the issue was identified during testing by a partner trusted by both Anthropic and the US government. The partner found that Fable’s existing safety controls did not fully prevent the specific jailbreak approach from working. As a result, users could reach capabilities that were supposed to remain restricted under the model’s safety framework. The dispute centers on the security boundary of the Mythos commercial model series and whether Fable met the safety requirements recognized by the government and its partner after release.

US Request to Fix or Take Down the Model Was Rejected

After the flaw was discovered, the US government asked Anthropic CEO Dario Amodei to fix the issue or temporarily take the model offline. Sacks said Anthropic rejected that request. Anthropic’s position is that the flaw is not serious enough to affect normal product operations. The US government and the relevant partner disagreed with that assessment and viewed the issue as requiring action before the model continued to be offered as a service.

Sacks also said Anthropic has long emphasized an AI-safety-first principle, but in this case chose to prioritize continued access to its consumer-facing model service. In response, the US government has imposed export controls on Anthropic. The government also said it would consider removing the restrictions after the relevant issues are resolved. The measure is tied directly to the Fable jailbreak flaw and the disagreement over whether Anthropic should fix it or take the model offline.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.
300

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.