Amazon CEO Andy Jassy may have been responsible for the security concerns that led Anthropic to cut off worldwide access to two of its models on Friday.
The Wall Street Journal reported that Mr. Jassy told Treasury Secretary Scott Bessent and other government officials that Amazon researchers used Anthropic's Claude Fable 5 to obtain information that could be used in a cyberattack. Subsequently, the government imposed an export control ban on the Fable 5 and Mythos 5 models.
An Amazon spokesperson told the Journal that “it is not uncommon for governments to seek our advice on potential security risks,” but the company “does not share details of those discussions.”
The Information and Reuters similarly reported that Amazon (a major investor in Anthropic) had raised concerns about the security of Anthropic's models.
David Sachs, President Trump's former AI czar and current co-chair of the President's Council of Advisors on Science and Technology, offered his own explanation of the debate, arguing that it is “a very reliable and trusted partner for both humanity and the U.S. government.” […] came forward as [information about] It's a jailbreak. ”
“The administrator asked,” Sachs added. [Anthropic CEO Dario Amodei] Fix the jailbreak or undeploy the model. Dario refused. ”