In a plot twist that screams "be careful what you wish for," Amazon’s cybersecurity research has apparently led to the White House slapping an export control directive on Anthropic’s Fable 5 and Mythos 5 models. According to the Wall Street Journal, the chain of events began when Amazon CEO Andy Jassy shared the company’s findings - claiming that through a series of prompts, Fable 5 could be tricked into serving up cyberattack-friendly information - with the government. Amazon has yet to respond to a request for comment, presumably because they're busy congratulating themselves.
Shortly after Jassy's little chat, the White House decided to block foreign nationals from accessing Anthropic's models. This would be a minor inconvenience, except that many of Anthropic’s researchers are foreign-born. So the company that built the AI now can't use it. Oops.
Anthropic, unsurprisingly, disputes the government's characterization of the issue as a "jailbreak." In a statement, they argued that similar vulnerabilities could be found in other publicly available models, including GPT 5.5. Security researchers seem to agree: Katie Moussouris, founder and CEO of LutaSecurity, posted on BlueSky that "I’ve seen the paper. It’s not a jailbreak." Former Commerce Department official Kate Koren speculated to the WSJ that the White House’s dislike of Anthropic may have influenced the decision. Because nothing says "fair and balanced regulation" like a grudge.
Anthropic and the Trump administration have been feuding for a while over the company's refusal to let its AI be used for mass surveillance of Americans or to power lethal autonomous weapons. In February, Trump ordered federal agencies to stop using Anthropic’s AI, and Secretary of Defense Pete Hegseth designated the company a supply chain risk. The two sides recently seemed to patch things up, working together to expand access to Mythos. But now they're headed for another clash. Pass the popcorn.