Amazon CEO Andy Jassy may have been responsible for the security issues that led Anthropic to block global access to two models on Friday.
The Wall Street Journal reported that Jassy told Treasury Secretary Scott Bessent and other government officials that Amazon researchers used Anthropic’s Claude Fable 5 to obtain information they could use in cyberattacks. The government has since imposed an export control ban on the Fable 5 and Mythos 5 models.
An Amazon spokesperson said in a statement that while “it is not uncommon for governments to seek our advice on potential security risks,” the company “does not share the details of such discussions.”
The spokesperson also pointed to an update from AWS that said it had been impacted by the model outage.
Information and Reuters similarly reported that Amazon (a major investor in Anthropic) had communicated concerns about the security of Anthropic’s model.
David Sacks, current co-chair of the President’s Council of Advisors on Science and Technology and Trump’s former AI czar, offered his own account of the discussion, claiming that “very trusted partners of both Intropic and USG (…) presented information about the jailbreak.”
“Administrators asked (Anthropic CEO Dario Amodei) to fix the jailbreak or de-deploy the model, but Dario refused,” Sacks added.
This post has been updated with a statement from an Amazon spokesperson.
