White House Grants Access to Anthropic’s Mythos Model to 100+ U.S. Institutions; Fable Still Shut Down
The Trump administration granted over 100 U.S. institutions access to Anthropic's Mythos AI model, de-escalating a dispute after export controls were imposed. The weaker Fable 5 model remains shut down, and critics warn the ad hoc policy of White House control over frontier AI creates a flawed regulatory framework.
Background
- **Anthropic** is the AI company behind the Claude chatbot; it's one of the most valuable private tech companies in the world, competing with OpenAI and Google.
- **Mythos** and **Fable 5** (formerly "Fable") are Anthropic's most advanced AI models. Fable 5 was briefly the most capable consumer-facing model available before it was taken offline.
- Two weeks before this article, the Trump administration suddenly imposed export controls on Mythos after Amazon and others warned the model could be "jailbroken" (tricked into ignoring its safety guardrails). This forced Anthropic to shut down both Mythos and Fable 5.
- Now the administration has partially reversed course, granting 100+ US institutions access to Mythos via a letter from Commerce Secretary Howard Lutnick. Fable 5 remains blocked, though talks to release it too are ongoing.
- This is controversial because it amounts to ad hoc White House control over frontier AI releases — no clear legal framework, just executive discretion. Critics argue this is especially dangerous under an administration they characterize as transactional and hostile to scientific expertise.