Inside the White House Decision That Cut Off Anthropic's AI Models

A new report from the Wall Street Journal reveals how Amazon's internal security research helped trigger a White House directive that forced Anthropic to block access to its Fable 5 and Mythos 5 models. The chain of events began when Amazon's security team published a paper demonstrating that, through carefully crafted prompts, it could coax Fable 5 into revealing information useful for cyberattacks. Amazon CEO Andy Jassy then discussed these findings directly with White House officials, and shortly after, the government issued an export control order barring foreign nationals from using the models. That move has had an outsized impact on Anthropic, where many researchers are foreign-born and suddenly found themselves locked out of their own product.

Anthropic pushed back, arguing the government was overreacting. The company said the supposed vulnerabilities weren't a true "jailbreak" and that similar issues exist in other publicly available models, including GPT 5. 5. Security researcher Katie Moussouris backed that view, posting on BlueSky that the Amazon paper "is not a jailbreak.

" Some former officials, including ex-Commerce Department staffer Kate Koren, have suggested the White House's long-simmering tension with Anthropic may have played a role. The company has refused to let its AI be used for mass surveillance or lethal autonomous weapons, a stance that put it at odds with the Trump administration. In February, Trump ordered federal agencies to stop using Anthropic's models, and Defense Secretary Pete Hegseth designated the company a supply chain risk. While the two sides had recently seemed to mend fences and even collaborated on expanding access to Mythos, this latest clash suggests the détente was short-lived.

Source: The Verge