Anthropic scrambles after Trump administration freezes its top AI models

1. 1. The Trump administration forced Anthropic to suspend its advanced AI models Fable and Mythos via export controls, barring foreign nationals from using them.
2. 2. The freeze was triggered by an Amazon researcher report identifying a potential jailbreak in Fable that could bypass guardrails and extract software vulnerability information.
3. 3. Anthropic disputes the targeting, arguing the jailbreak capability is not unique and can be demonstrated in rival systems like OpenAI's.
4. 4. The government's freeze does not apply to leading models from OpenAI or Google, despite those having demonstrated comparable capabilities.
5. 5. The Pentagon has designated Anthropic as a supply-chain risk to national security, and the two sides are in litigation over the designation.
6. 6. The European Commission is assessing the implications of the directive on its citizens, calling the challenge shared and not discriminatory against partners.
7. 7. The White House previously signaled a less interventionist approach, and an executive order earlier this month lacked a mechanism to stop model releases.