Reports suggest that Amazon CEO, Andy Jassy, may have tipped off Treasury Secretary Scott Bessent about potential security risks posed by Anthropic’s models. After receiving this intel, the US government slapped an export control ban on two of Anthropic's AI models: Claude Fable 5 and Mythos 5. While Amazon denies sharing specifics of these discussions, it does confirm that its operations have been impacted.

David Sacks, a former advisor to President Trump who now co-chairs the President’s Council of Advisors on Science and Technology, adds his version of events. He claims that a ‘highly credible trusted partner’ brought attention to a jailbreak issue, leading to Anthropic CEO Dario Amodei being pressured to address or completely remove the model.

Anthropic maintains that the functionalities cited as problematic are already available in other models and is undeterred. This saga raises questions about who should be responsible for regulating AI: governments, tech giants, or both?

The incident also highlights the growing tension between innovative AI development and national security concerns, with major players like Amazon and Anthropic caught in the crossfire.