Cyber experts warn US ban on Mythos and Fable for foreigners not good, here is why
On Friday, the US government banned all foreigners from using Anthropic’s Mythos 5 and Fable 5 AI models due to cybersecurity concerns – with Anthropic
On Friday, the US government banned all foreigners from using Anthropic’s Mythos 5 and Fable 5 AI models due to cybersecurity concerns – with Anthropic ultimately suspending the models for all users. The White House believes that these models could be jailbroken by bad actors, and then they can use the AI to potentially hack any software out there. However, according to a group of leading cyber experts, this ban on the use of Mythos 5 and Fable 5 may do more harm to cyber defenders. Read Full Story A group of 160 cyber experts wrote an open letter to US Commerce Secretary Howard Lutnick, explaining why this ban may be dangerous for the US and its allies. To make the point clear, the group adds that the US was taking away the best AI models out there from those who were working to fix cybersecurity flaws at a time when bad actors may be getting better models. The letter read, “To pull the best capabilities away from defenders without a good reason when our adversaries are rapidly advancing is dangerous.” Some of the notable names in this group include Facebook's former chief security officer Alex Stamos, cybersecurity expert Bruce Schneier, who is often called the godfather of cybersecurity, and Philip Zimmermann – the creator of PGP, the most widely used email encryption software in the world.
Experts say China may already have access to advanced models Reports indicate that a reason behind the ban on Mythos 5 and Fable 5 was the fear that China-linked groups may jailbreak these models and use them to potentially plan cyberattacks. However, according to this group of cyber experts, Chinese AI models are already advancing at a rapid pace. “The Chinese open-weight models are only months behind the best American models, and those are the models we know about,” the group wrote. The group argues that it is possible that the Chinese government already has advanced models that we don’t know about. They added, “It seems likely that the PRC government has access to private capabilities beyond what has been published.” If true, this would mean that the Chinese officials may already have AI models similar to Mythos in terms of cybersecurity, which would make the ban ineffective. Using Mythos for cyber attacks is not so easy While some may believe that these AI models are so advanced that one can easily give a simple prompt and see magic happen, Alex Stamos insists that things are not as simple.
He told Axios, "You cannot give Fable the entire Linux kernel and say 'Find all the security bugs.’” Stamos claimed that the Fable 5 capability that appears to have alarmed the White House was the model's ability to create a "proof of concept" for vulnerabilities. Proofs of concept can serve as a blueprint for code that allows adversarial hackers into a system, but they also help security teams understand how to defend their own systems. "For us to shut down our best capabilities at the moment we know the Chinese are using and stockpiling these vulnerabilities is dangerous – absolutely foolish," Stamos added. "We are in a race right now to fix these bugs as fast as possible." It is believed that the trigger behind the ban came after researchers at Amazon were able to jailbreak the models by giving certain prompts to bypass safeguards. However, the group says that one can easily use “GPT-5.5, Opus, Sonnet and even Chinese models like Kimi 2.7” to find cybersecurity flaws.
The BriefWire