Security experts protest Anthropic ban
Seventy-six cybersecurity experts signed an open letter urging the US to lift its export ban on Anthropic's Fable and Mythos models.
Seventy-six cybersecurity professionals signed an open letter on June 14 asking the US government to lift its export-control order on Anthropic’s Fable and Mythos models. Signatories include Alex Stamos, Facebook’s former security chief; Katie Moussouris, founder of Luta Security; Bugcrowd founder Casey Ellis; and cryptographer Jon Callas. Their argument: the order “has taken the best models away from defenders,” who used them to find and fix software vulnerabilities, while adversaries keep advancing. The letter, posted at freefable.org, calls the move “dangerous.”
The order landed on June 12, when the government told Anthropic to limit exports of the two models on national-security grounds without giving specifics. Anthropic suspended access worldwide in response. The company has said the order may rest on a report describing a way to jailbreak the public Fable model to reach Mythos-level capability. Moussouris, who reviewed the non-public Amazon paper behind that claim, says it shows no real jailbreak: researchers asked Fable to fix open-source code with planted bugs, which she calls the ordinary find-fix-test loop of defensive security, not a bypass.
Why this matters
When a government can switch off a model overnight, the people who lose access are not only attackers. The dispute also shows how thin the public evidence is: a ban reaching every user worldwide may rest on one contested, unpublished paper. If your security or development work leaned on a top model, treat sudden access loss as a live risk, not a hypothetical, and keep a tested fallback. The original access freeze and Anthropic’s visible Fable guardrails set up this fight, and the sovereign-AI scramble in India is partly a response to exactly this.