Cybersecurity Leaders Tell the Government to Reverse the Anthropic Ban, Saying It Disarms Defenders

The people who actually defend software are pushing back on the Anthropic shutdown. Dozens of cybersecurity leaders signed an open letter urging the Trump administration to reverse the export controls that took Fable 5 and Mythos 5 offline, arguing the order backfires on US security. The letter went to Commerce Secretary Lutnick and Cyber Director Cairncross, the two officials who can undo it.

The signers are not random. The list includes Alex Stamos, Katie Moussouris, and security chiefs from companies like Sophos, Adobe, Zoom, NVIDIA, and Veracode, the kind of people who run defense at real organizations. Their core claim is blunt: the directive has taken the best models away from defenders, created market uncertainty, and risked America's AI leadership without any real risk to justify it. That is a direct shot at the government's security rationale.

Their technical point is the heart of it. Mythos is genuinely good at finding software flaws, but it is not uniquely good, and the same jailbreak the government worried about reportedly works on other models too, including OpenAI's GPT-5.5 and China's Kimi. Anthropic also built guardrails into Fable specifically to block offensive use. After reviewing the research, Moussouris said Fable was doing basic, harmless cybersecurity work, not anything dangerous. The trigger, by several accounts, was as mundane as asking the model to fix some code.

The administration sees it differently, and that gap has not closed. Officials cited national security, a jailbreak demonstrated by Amazon, and reports that a Mythos model reached an entity tied to the Chinese Communist Party. Anthropic disputes the framing and says it only ever got verbal evidence of a narrow flaw. So you have the government calling it a real threat and much of the security industry calling it an own goal, with the same facts in front of both.

The letter matters because of who signed it. When the defenders a policy is supposed to protect say the policy hurts them, that is hard for officials to ignore, especially with Anthropic staff already in Washington negotiating a fix. It also reframes the whole fight. This stopped being only about one company's models and became a test of whether the US can restrict frontier AI without kneecapping its own cyber defense. The industry just told Washington it picked wrong.

So the pressure to reinstate Fable and Mythos is now coming from two directions at once, Anthropic from the inside and the security community from the outside. The models are still dark. Whether this letter speeds up the reversal is the thing to watch this week. When the defenders are asking for the weapon back, that is worth listening to.