More than 40 cybersecurity experts, led by former Facebook Chief Security Officer Alex Stamos, have signed an open letter urging the Trump administration to reverse its export ban on Anthropic's most powerful AI models. The letter argues that taking Mythos and Fable 5 offline doesn't eliminate security risks — it strips defenders of critical tools needed to counter automated AI hacking campaigns.
The White House's decision stemmed from a vulnerability report submitted by Amazon security researchers, who warned that Fable 5 could automatically generate proof-of-concept exploit code and, under certain conditions, evolve into a fully autonomous attack chain. But Stamos says the administration's concern is misguided. The standard version of Fable 5 can't scan the entire Linux kernel and identify every security hole. The version with autonomous attack capabilities is actually a limited, unreleased variant called Mythos 5 and Mythos Preview, tested only by select institutions.
The open letter emphasizes that generating exploit code isn't unique to Fable 5. OpenAI's GPT-5.5, Anthropic's own Opus and Sonnet, and even China's Kimi 2.7 can perform comparable vulnerability analysis and code generation. Stamos warns that Chinese open-source models are already closing the gap with Fable 5 in analyzing system flaws. At a time when adversaries are stockpiling and exploiting vulnerabilities, shutting down America's best defensive tools is a self-inflicted wound.