Cybersecurity experts demand lifting ban on Anthropic models

A group of dozens of cybersecurity experts has approached the White House demanding that export restrictions on Anthropic's flagship models — Fable and Mythos — be lifted. The authors call the ban "dangerous" and warn that it restricts precisely those engaged in protecting software and corporate infrastructure.

The Essence of the Appeal

The letter was signed by dozens of industry veterans — specialists with years of experience from private companies and academic institutions. Their key argument: the attacking side — hackers, cybercriminals, state-sponsored APT groups — is not bound by regulatory barriers. Malicious actors continue to freely use powerful AI models to discover vulnerabilities, write exploits, and automate attacks at an industrial scale. The defensive side, meanwhile, finds itself in an artificially disadvantaged position: security specialists outside authorized jurisdictions are deprived of the same tools — and work with demonstrably smaller capabilities.

What Defenders Lose

Fable and Mythos — Anthropic's flagship models — demonstrate high results precisely on tasks critical to cybersecurity: code analysis for vulnerabilities, interpretation of complex binary files, synthesis of technical threat reports. This makes them valuable primarily for security specialists' work. Specific applications that come under threat:

• Automated code audit and zero-day vulnerability detection
• Security patch generation and verification
• Malware analysis and reverse engineering
• Threat intelligence automation and incident classification
• Development and updating of security playbooks for response teams

The more complex the threat and the more intricate the code — the more critical the quality of the AI assistant. This is precisely why restricting access to the most powerful models hits analysts first and foremost, not malicious actors.

Regulatory Logic and Its Weaknesses

Export restrictions on frontier models are part of a broader policy: American regulators view powerful AI systems as a strategic resource comparable to semiconductor technologies. The formal objective is to prevent the use of cutting-edge developments by competing states for military or intelligence purposes. The authors of the appeal point to a fundamental contradiction in this logic.

The most dangerous applications of AI — automation of cyberattacks, exploit development, deepfake creation — are realized through open models and workarounds. No export control can block them. Meanwhile, the defensive community — MSSPs, independent researchers, corporate SOCs, Bug Bounty teams — bears direct operational losses.

Cybersecurity is inherently global: threats do not recognize borders, and effective defense requires international cooperation. Restrictions that deprive foreign partners and allies of access to the best tools weaken the entire ecosystem response to threats — including those directed against the United States itself.

What This Means

The appeal of cybersecurity veterans is one of the first coordinated industry signals against the new generation of AI export restrictions. It puts regulators before a choice: continue to block powerful models in their entirety or transition to a more targeted logic — restricting specific applications rather than the tools themselves. Until this distinction is made, the control functions as a barrier for defenders but not for attackers.