Cybersecurity Experts Unite Against U.S. Government Ban on Anthropic’s Powerful Models
A group of cybersecurity veterans has penned an open letter urging the U.S. government to lift export controls on Anthropic’s Fable and Mythos models, arguing that such restrictions are dangerous for defenders.
Admin User
A powerful alliance of cybersecurity experts, including well-known industry veterans like Alex Stamos and Paul Vixie, have joined forces to protest a recent ban imposed by the U.S. government on Anthropic's advanced AI models, Fable and Mythos.
According to an open letter published by these experts, the export control order has removed ‘the best models’ from cybersecurity defenders' hands, preventing them from using these tools to find vulnerabilities and enhance their security measures. The signatories argue that 'this action is dangerous,' as it leaves cybersecurity defenders vulnerable when adversaries are rapidly advancing.
On Friday, Anthropic issued a statement announcing the export restrictions on Fable and Mythos, citing national security concerns without specifying details. As a result, access to these models was suspended worldwide. The letter, signed by 76 notable cybersecurity professionals, including Alex Stamos (former Facebook chief of security), Casey Ellis (founder of Bugcrowd), Jon Callas (famed cryptographer and former Apple security architect), Paul Vixie (computer scientist), Dino Dai Zovi (former head of applied security engineering at Block), Katie Moussouris (founder of Luta Security), and Rachel Tobac (CEO of SocialProof Security), highlights the potential risks posed by these bans.
When Mythos was first released in April, Anthropic emphasized its capabilities for finding security vulnerabilities. The company restricted access to around 50 organizations initially but later expanded this number to include approximately 150 entities across 15 countries. Fable, a public version of Mythos with strict guardrails, was launched last week, yet these restrictions have been so stringent that many cybersecurity experts found them limiting.
The ban has sparked controversy due to a recent paper by Amazon researchers, which Anthropic believes may be the basis for the export control order. However, Katie Moussouris, one of the letter's signatories, disputes this claim, stating that the paper did not demonstrate a real jailbreak but rather asked Fable to fix known vulnerabilities in open-source code.
Despite these concerns, Anthropic’s restrictions on Fable have been so stringent that even other models like OpenAI’s GPT-5.5 and Anthropic’s Claude Opus 4.8 and Sonnet can replicate the capabilities of Mythos. Moussouris argues that 'defenders need to be able to ask AI to fix the bugs in a file, explain why the fix matters, and write tests that confirm the patch works,' which is a critical function for defensive security.
The letter also calls for transparent and fairly enforced regulations created through a democratic process involving scientific research from industry and academic experts. These regulations should only be used minimally to ensure public safety.
