White House imposes export controls on Anthropic's Fable model after security concerns
The Trump administration placed export restrictions on Anthropic's newly released Fable AI model following a series of high-level meetings and concerns about potential security vulnerabilities.
Select a version of the text written from a presumed ideological perspective. This is not the original text, but a hypothetical version — how someone with that viewpoint might have phrased it. Tapping the current version again will return to the original or select cleaned version.
The U.S. Treasury and White House officials imposed export controls on Anthropic's Fable and Mythos 5 models on Friday, citing national-security risks. The action followed a rapid, 24-hour effort by senior officials to persuade the company to voluntarily withdraw the models after doubts emerged about the effectiveness of their safety guardrails.
According to unnamed administration sources, Treasury Secretary Scott Bessent, White House Cyber Director Sean Cairncross and other senior staff held multiple calls with Anthropic CEO Dario Amodei. The discussions, which also involved the Commerce Secretary and senior White House aides, focused on recent findings—reported by Amazon, an Anthropic investor—that the model could be bypassed in certain scenarios.
Anthropic initially defended its safeguards, stating in a blog post that no universal “jailbreak” had been discovered and that the model’s protections were intentionally robust. The company later said it was complying with the export-control order, which it described as disproportionate and lacking a transparent, fact-based process.
"Export controls were a last resort after begging them for hours to work with us," a senior White House official said.
The administration’s decision effectively disabled the models for all customers, including foreign nationals, until the identified vulnerabilities are addressed. Amazon declined to provide details of its discussions with the government, noting that such consultations are common.
The episode highlights ongoing tensions between regulators and AI developers as the U.S. seeks to balance rapid technological advancement with security considerations.
