A report says cybersecurity research from Amazon, plus conversations between Amazon CEO Andy Jassy and the White House, played a role in an export control directive that led Anthropic to block foreign access to Fable 5 and Mythos 5. Anthropic disputes the gove
It wasn’t an abstract policy debate—it was a directive that quickly reshaped what people could do with Anthropic’s newest models.
The export control order that led Anthropic to cut off foreign access to Fable 5 and Mythos 5 was triggered in part by cybersecurity research from Amazon and by conversations between Amazon CEO Andy Jassy and the White House, a report states. Amazon has yet to respond to a request for comment.
In the report. Amazon’s paper argues that. through a series of prompts. it was able to get Fable 5 to produce information that could be used in cyberattacks. Shortly after Jassy shared the company’s findings with the government. Amazon’s position moved from discovery to restriction: the company made the call to block its use by foreign nationals.
What makes the timing sting for Anthropic’s own team is that many of the company’s researchers are foreign-born. The result, according to the reporting, is that they were barred from accessing their own product.
Anthropic says the government is telling the story wrong. In a statement, the company disputed the government’s characterization of the issue as a “jailbreak.” Anthropic argued that many of the same vulnerabilities could be discovered using other publicly available models, including GPT 5.5.
Some security researchers appear to back Anthropic’s reading. Katie Moussouris, the founder and CEO of LutaSecurity, posted on BlueSky: “I’ve seen the paper. It’s not a jailbreak.” Another perspective came from former Commerce Department official Kate Koren. who speculated to the Wall Street Journal that the White House’s dislike of Anthropic may have influenced the decision.
That backdrop isn’t new. Anthropic and the Trump administration have been at odds for some time. The conflict has included the company’s refusal to allow its AI to be used for mass surveillance of Americans or to power lethal autonomous weapons. In February, Trump instructed federal agencies to stop using Anthropic’s AI. And just hours later, Secretary of Defense Pete Hegseth designated the company a supply chain risk.
After that. there was a brief thaw: the government and the company seemed to have made amends. and they worked together to expand access to Mythos. Now. based on the latest reporting. the relationship appears headed for another confrontation—this time over how a cybersecurity finding becomes a boundary on who can use a model. and who gets locked out when the restriction lands.
On the ground level, the central contradiction is hard to miss. Amazon’s research is framed as a way to show how Fable 5 could be pushed toward harmful ends. Anthropic and some researchers counter that the same weaknesses can be found elsewhere and that the government’s label doesn’t fit. For Anthropic’s staff—especially those foreign-born—the policy doesn’t just limit outsiders; it reaches inside the company too. And for the administration. the latest steps suggest the dispute isn’t only about technical risk. but about trust—who is allowed to control advanced tools. and under what terms.
Amazon cybersecurity Anthropic Fable 5 Mythos 5 export controls White House Andy Jassy jailbreak dispute LutaSecurity Katie Moussouris GPT 5.5 AI safety policy