Trump administration accuses Anthropic of releasing Fable 5 without government approval, but security experts argue the government's demand for 'unhackable' AI models may be unrealistic.

What happened

Government officials claim Anthropic ignored Trump's recently issued cyber executive order by releasing Fable 5 without waiting for a designated clearinghouse to sign off. An administration official also accused Anthropic of knowing a jailbreak (an unauthorized bypass of the model's safety features) could occur. The Department of Commerce and Anthropic are in talks, with further meetings planned involving the CIA and science advisor Michael Kratsios.

Why it matters

The accusation reflects a fundamental gap in understanding between the government and Anthropic. Security experts note that anyone working closely with AI models knows they can be hacked—OpenAI has warned that prompt injection, a related hacking method, may never be fully solved. If the U.S. government insists frontier AI models must be 'unhackable' before they ship internationally, the conflict signals difficult negotiations ahead. Meanwhile, over 100 security experts and tech industry executives have published an open letter calling for export controls on Fable and Mythos to be lifted, arguing that other models like GPT-5.5, Opus, and Sonnet can perform the same security tasks.

What to watch

The signatories warn that export controls are stripping defenders of the best tools while Chinese open-weight models are only months behind the top U.S. models. The real question, according to the article, is how severe the alleged breach is and how fast countermeasures kick in.