Dozens of Security Experts Urge the White House to Reverse Restrictions on Anthropic's Fable 5

In a notable escalation of the controversy over the decision to restrict Anthropic's latest models, dozens of senior cybersecurity experts and technology leaders signed an open letter demanding that the US administration reverse the restrictions imposed on the Claude Fable 5 and Mythos 5 models. The letter, dated June 14, 2026, and addressed to Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross, described the decision as "dangerous," warning that it harms cyberspace defenders more than it deters attackers.

Who Signed the Letter?

The letter campaign was led by Alex Stamos, the former chief security officer of Facebook, and by early in the week it had been joined by nearly 150 signatories among the industry's most prominent names. Among the signatories were the renowned cryptographer Bruce Schneier, Luta Security CEO Katie Moussouris, SocialProof Security CEO Rachel Tobac, Veracode co-founder Chris Wysopal, and the prominent computer scientist Paul Vixie. The list also included security leaders and officers from major companies such as Adobe, Nvidia, Zoom, and Sophos, giving the letter weight that extends beyond individuals to substantial institutions.

Background of the Decision

The crisis traces back to last Friday (June 12), when Anthropic received an export control directive barring access by any foreign national to the two models, forcing it to disable them for all users to ensure compliance, just three days after their launch. According to reports, the directive came after warnings the administration received from Amazon researchers, in addition to a call from Amazon CEO Andy Jassy, over concerns about potential Chinese access and a jailbreak vulnerability that might allow safety guardrails to be bypassed.

The Signatories' Arguments

The letter's arguments centered on several interconnected points. The first is that the jailbreak vulnerability cited by White House officials is not exclusive to Fable 5, but applies to other publicly available models such as OpenAI's GPT-5.5, Anthropic's own Opus and Sonnet models, and Chinese models like Kimi 2.7; meaning that restricting Fable 5 alone does not address the core of the problem. The second is that Anthropic has already built multiple protections into the model to prevent its use for offensive purposes.

The third argument, the strongest, is that removing a model of such capability in cybersecurity tasks "while US adversaries are rapidly advancing is dangerous," as it deprives defenders of the best available tools. Katie Moussouris explained that the model's capabilities that alarmed the administration, such as creating "proofs of concept" for vulnerabilities, stem originally from its skill at fixing code and verifying patches; a defensive capability that cannot be removed without making the model worse at fixing bugs. Stamos also stressed that open-weight Chinese models are only a few months behind the best American models, so the restriction grants no meaningful strategic advantage.

A Legal and Regulatory Dimension

The objection was not limited to security experts; export control experts questioned whether the US government even has the legal authority to impose these restrictions in their current form. This legal dimension adds a new layer to a dispute that now combines technical, security, and regulatory questions at once. Meanwhile, Anthropic met with White House officials this week seeking to cancel or narrow the measures, and is working with the government to reach a solution.

A Precedent That May Shape the Future

The significance of this case goes beyond the two models in question. As Stamos put it, the decision "set a precedent that American models can't do defensive security research." The outcome of the meeting between the company and the government may determine how aggressively governments will regulate the next generation of advanced AI systems. The letter warned that the action "took the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it."

Conclusion

This incident reveals a fundamental tension in AI governance: how does the state balance national security concerns against defenders' need for the most powerful tools? With nearly 150 senior security experts aligned behind a single position, the tech community appears to view this decision as a danger to security rather than a protection of it. Regardless of the dispute's outcome, it sets sensitive standards for the future relationship between governments and AI companies.