Amazon CEO Andy Jassy raised security concerns tied to Anthropic’s Claude Fable 5 and Mythos 5 models with U.S. government officials before a government crackdown. Anthropic cut worldwide access to both models on Friday after an export control ban was imposed
On Friday, Anthropic cut worldwide access to two models—Claude Fable 5 and Mythos 5—after a government export control ban was imposed on Fable 5.
The move immediately turned into a question of who first raised the alarm. The Wall Street Journal reports that Amazon CEO Andy 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 cyberattacks.
An Amazon spokesperson later pushed back on how much the public would learn about what was said in those meetings. In a statement. the spokesperson said it’s “not uncommon for governments to seek our counsel on potential security risks. ” but added that the company does not “share the details of those discussions.” The spokesperson also pointed to an update stating that AWS has been affected by the model cut off.
The story’s pressure points don’t stop with the government’s decision. The Information and Reuters similarly reported that Amazon, a major Anthropic investor, had communicated concerns about the security of Anthropic’s models.
David Sacks—Trump’s former AI czar who now co-chairs the President’s Council of Advisors on Science and Technology—offered a sharper version of how those concerns reached the White House circle. He claimed that a “highly credible trusted partner of both Anthropic and the USG […] came forward with [information about] a jailbreak.”.
Sacks added that “The Admin asked [Anthropic CEO Dario Amodei] to fix the jailbreak or de-deploy the model. Dario refused.”
Taken together. the facts sketch a tense chain: Amazon’s internal work and security concerns reach government officials; an export control ban lands on Fable 5; Anthropic responds by cutting worldwide access to Fable 5 and Mythos 5. In parallel. Amazon maintains that it won’t disclose the details of its government discussions. even as its own cloud business is described as affected by the cutoff.
Amazon’s spokesperson also reiterated that point by addressing the public accounting that followed. The company, in that statement, did not confirm the specifics of what it shared—only the existence of ongoing government engagement on potential risks.
For now, Anthropic’s decision is already public and immediate. The question that remains—how much of the government action depended on which warning, and at what exact point—has become the real battleground behind the model cutoffs.
Amazon Andy Jassy Anthropic Claude Fable 5 Mythos 5 AWS export control ban cyberattacks jailbreak Scott Bessent Dario Amodei David Sacks cybersecurity AI models
So they banned it because it could help hackers, right? Seems like common sense.
Wait I’m confused—did Amazon “warn” them or did Anthropic get caught first? The article makes it sound like multiple people were talking to Treasury and then suddenly poof, export ban.
Honestly sounds like corporate drama. Like if Amazon is also using Claude for cybersecurity stuff then why are they acting shocked? Also this “jailbreak” thing… so they refused because they didn’t want to admit it worked?
I don’t trust any of it. If the government says “export ban” then why is Anthropic still supposedly cutting access “worldwide” like that fixes the cyberattack issue? It’s weird that AWS is “affected” too, like Amazon is the one who benefits from all this attention… and then they won’t share details with the public 🙄