Cybersecurity Leaders Urge White House to Reverse Anthropic AI Model Ban

A coalition of more than 40 prominent cybersecurity leaders is pushing back against the White House ban on Anthropic's most advanced artificial intelligence models, calling the move a self-inflicted wound to national security. The open letter, led by former Facebook chief security officer Alex Stamos, demands the Trump administration immediately restore access to both Mythos 5 and Fable 5. The letter has rapidly gained momentum since its release, with new signatories joining throughout the weekend.

Defenders Pay the Price

The signatories include CISOs, security researchers, and senior executives from major technology companies like Adobe, Zoom, and Sophos. Their central argument is blunt: the ban kneecaps the good guys while barely inconveniencing the bad ones. Defensive cybersecurity teams rely heavily on these advanced AI tools to detect threats, analyze vulnerabilities, and respond to attacks at machine speed. A large language model like Mythos can scan millions of lines of code in hours, a task that would take human analysts weeks or months. Without access, security teams must fall back on slower and less capable alternatives while adversaries simply switch to other powerful models or open-source tools readily available online.

Why the Ban Happened

The White House ordered Anthropic to cut off access to both models for all foreign nationals after months of growing concern about their extreme capabilities. During testing, Mythos reportedly discovered serious security flaws across every major operating system and web browser, and could chain those weaknesses into complete end-to-end cyberattacks. Officials feared these abilities could be weaponized by hostile actors if access was not tightly controlled. Anthropic said it took the systems offline to comply with the directive, calling it a misunderstanding it hopes to resolve quickly.

The letter's authors argue that the capabilities Amazon researchers flagged exist across other leading AI models too, making the targeted ban on Anthropic both unfair and ineffective. They propose a middle path: let United States agencies and major corporations safely use the strongest defensive AI available, with controlled access and independent risk assessments rather than broad restrictions that also shut out security teams. Whether the administration will reconsider remains an open question, but the growing pressure from the cybersecurity community signals that the current approach may not hold for long.