Amazon CEO's security fears led to US ban on Anthropic's Fable 5 model

The Wall Street Journal reports that Amazon CEO Andy Jassy directly raised security concerns about Anthropic's models with Treasury Secretary Scott Bessent and other government officials, leading to a swift government crackdown. According to sources, Amazon researchers demonstrated that Anthropic's Claude Fable 5 could be used to obtain information that could aid in cyberattacks. The US government subsequently imposed an export control ban on both Claude Fable 5 and the related Mythos 5 model. Anthropic responded by cutting off worldwide access to these models, including through Amazon's AWS platform, which issued a public update acknowledging the impact. An Amazon spokesperson stated that while governments often seek counsel on security risks, the company does not share details of those discussions. The incident highlights the delicate balance between AI development, national security, and corporate investments.

David Sacks, Trump's former AI czar now co-chairing the President's Council of Advisors on Science and Technology, provided additional context via social media. Sacks claimed that "a highly credible trusted partner of both Anthropic and the USG came forward with a jailbreak" of the models. He asserted that the administration asked Anthropic CEO Dario Amodei to "fix the jailbreak or de-deploy the model," but that Amodei refused. Anthropic defended its position in a blog post, arguing that the capabilities that concerned the government are "already available in other publicly accessible models," suggesting the ban may be unfair. The situation raises questions about the influence of major investors like Amazon (a key Anthropic investor) in shaping AI regulation and the potential for internal security audits to trigger government intervention.