Cybersecurity experts challenge US export ban on Anthropic's Fable and Mythos models

The U.S. government ordered Anthropic to restrict exports of its Fable and Mythos models on security grounds, prompting a coalition of 76 prominent cybersecurity experts to challenge the decision publicly. According to TechCrunch AI, the signatories argue the ban removes critical vulnerability-detection tools from legitimate defenders at a moment when security threats are accelerating globally.

The Government Order and Anthropic’s Response

On Friday, the U.S. government issued an export control order against Anthropic’s two most capable models without publishing detailed justification, according to TechCrunch AI. In response, Anthropic suspended worldwide access to both Fable and Mythos. According to the company, the ban may have been triggered by a report indicating a method existed to circumvent Fable’s guardrails and recover Mythos-level capabilities.

Mythos and Fable: Design and Deployment

Mythos, which launched as a preview in April 2026, was engineered to excel at identifying software security vulnerabilities. According to TechCrunch AI, Anthropic restricted access to prevent adversaries from weaponizing the model. The company granted initial access to roughly 50 organizations, later expanding to approximately 150 organizations across 15 countries.

Fable, a public-facing version released more recently, incorporates strict guardrails intended to block use in biology, chemistry, and cybersecurity domains, and to prevent model distillation. According to TechCrunch AI, the guardrails proved so restrictive that cybersecurity experts found them blocking most security-related requests.

The Expert Coalition’s Opposition

According to the open letter cited by TechCrunch AI, the signatories contend that “this action has taken the best models away from defenders” conducting legitimate vulnerability research. The letter argues: “To pull the best capabilities away from defenders without a good reason when our adversaries are rapidly advancing is dangerous.”

Signatories include Alex Stamos, former chief security officer at Facebook; Casey Ellis, founder of the Bugcrowd bug-bounty platform; Jon Callas, cryptographer and former Apple security manager; 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.

Why This Matters

The dispute reflects a tension in U.S. artificial intelligence policy between national security imperatives and industrial capability. Restricting dual-use models—tools that serve both defensive and offensive functions—requires balancing the risk to adversaries against the cost to domestic security research. If the government’s restriction stands without public justification, it may signal a precedent for broader export controls on frontier AI systems, affecting how AI vendors design access policies and how U.S. security teams source vulnerability-detection infrastructure. The open letter’s high-profile signatories suggest the policy may face sustained industry pressure, particularly if alternative tools for vulnerability research remain unavailable to defenders operating within U.S. jurisdiction.