According to a report by The TechFlow Street Journal on June 14, the US government's export controls and access restrictions on the Anthropic model Fable 5 / Mythos 5 are partly due to Amazon's cybersecurity research and communications between AWS CEO Andy Jassy and the White House.
According to reports, a study submitted by Amazon indicated that through a series of cue word tests, researchers were able to induce Fable 5 to output sensitive information that could be used for cyberattacks, raising security concerns. Andy García subsequently reported the findings to the US government, prompting the White House to take further restrictive measures, including banning foreign users from accessing the model.
Anthropic denied the government's claim that the issue was a "jailbreak," stating that similar vulnerabilities could occur in other publicly available models, including GPT 5.5. Some security researchers also support this view, believing that the issue is closer to a "hint of injection risk" than a serious model breach.
According to former U.S. Commerce Department official Kate Curran, the White House's existing policy stance towards Anthropic may also have influenced this decision, as Anthropic disagrees with the White House on the boundaries of artificial intelligence safety, including refusing to use its models for mass surveillance or lethal autonomous weapons systems. Although the two sides had eased tensions and expanded cooperation earlier this year, this incident could further escalate tensions.
