Anthropic Pulls Fable 5 and Mythos 5 AI Models Amid White House Export Controls
Anthropic has taken its Fable 5 and Mythos 5 artificial intelligence models offline in response to export controls imposed by the Trump administration. The White House cited national security authorities, expressing concerns that Fable 5 could be bypassed or jailbroken. Anthropic CEO Dario Amodei defended the models, stating that disclosed potential jailbreaks are minor or benign, while industry experts have urged the government to lift the controls.
Anthropic has withdrawn its Fable 5 and Mythos 5 artificial intelligence models from public access following the imposition of strict export controls by the Trump administration.
According to Anthropic, the White House cited "national security authorities" in its directive, leading the company to pull the models offline to ensure compliance. The administration reportedly believes that Fable 5, released days earlier, could potentially be bypassed or jailbroken.
Anthropic CEO Dario Amodei defended the firm's Mythos-class AI models, stating that any potential jailbreaks disclosed were either benign responses or minor findings that offered no significant uplift specific to Mythos models.
Prior to the official crackdown, senior Trump administration officials reportedly attempted to persuade Anthropic to voluntarily remove Fable 5 from circulation. These efforts included a series of phone calls between Anthropic CEO Dario Amodei and officials such as Treasury Secretary Scott Bessent and White House Cyber Director Sean Cairncross. The officials' concerns were reportedly prompted by Amazon CEO Andy Jassy.
Anthropic officials have since traveled to Washington, DC, for meetings with White House representatives to address the issue. The dispute is seen as a potential precedent for the AI industry, testing the extent to which government officials will intervene in the release of AI systems deemed national security risks. The export controls also barred access to Fable 5 and Mythos 5 for foreign nationals, including some of Anthropic's non-US employees, though access to other Anthropic models remains unaffected.
An open letter signed by numerous AI industry and cybersecurity leaders, including specialists from Nvidia and Adobe, urged the government to lift the export controls. The letter argued that Anthropic had incorporated aggressive protections into Fable 5 to prevent its use for cyber offensive purposes and that restricting access to such capabilities for defenders could be dangerous.
According to Business Insider, key questions, such as the exact evidence that prompted the government's intervention, remain unanswered.
