Anthropic says the U.S. government has forced it to abruptly suspend access to its newest AI models, Fable 5 and Mythos 5, following a national security directive issued on Friday.
According to the company, the order requires Anthropic to block access to the models for all foreign nationals, both inside and outside the United States. Rather than attempt to sort users by nationality, Anthropic says it has disabled access to the models entirely while it works to comply with the directive.
The company claims it received the government’s order at 5:21 p.m. ET and was given little information regarding the specific national security concerns behind the action.
Anthropic believes the government’s concerns stem from an alleged jailbreak technique that could be used to bypass some of the models’ safeguards. However, the company strongly disputes the severity of the issue.
In a lengthy statement, Anthropic argued that the reported jailbreak is narrow rather than universal and that the capabilities demonstrated are already available in other publicly accessible AI systems. The company even specifically referenced OpenAI’s GPT-5.5, saying the reported behavior does not provide unique capabilities beyond what competing models can already do.
“We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users,” Anthropic said. “However, we disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.”
The company also warned that applying the same standard across the AI industry could effectively halt future frontier model deployments.
While Anthropic is clearly unhappy with the decision, the situation raises broader questions about how governments should oversee increasingly powerful AI systems.
Personally, I don’t hate the idea of governments refusing to blindly let AI companies do whatever they want. The AI industry has largely operated under the assumption that companies can release first and address problems later. Even if this particular action ultimately proves to be an overstep, I would rather see regulators scrutinizing powerful AI models than simply taking the industry’s word that everything is fine.
At the same time, transparency matters. If the government is going to intervene at this level and effectively pull a major AI model from the market, there should be a clear process and a clear explanation. Right now, Anthropic claims it has seen little evidence that the alleged jailbreak represents a significant threat.
That leaves observers with more questions than answers.
What makes this story particularly noteworthy is that it may represent the first major instance of the U.S. government effectively recalling a frontier AI model after deployment. Whether the government’s concerns prove justified or not, this could become a defining moment in the ongoing debate over how advanced AI systems should be regulated.
For now, Fable 5 and Mythos 5 are offline, Anthropic is pushing back publicly, and the rest of the AI industry is undoubtedly watching very closely.
Support independent tech journalism
NERDS.xyz is independently owned and operated. If you enjoy my coverage of Linux, AI, hardware, cybersecurity, and tech culture, consider supporting the site on Ko-fi.
Support NERDS.xyz