Anthropic's Top AI Models Taken Offline After Trump Administration Imposes Export Controls Over Jailbreak Vulnerability Cover-Up Claims
Summary
Anthropic's most powerful AI models, Mythos and Fable, are taken offline after the Trump administration imposes export controls, alleging the company knowingly deployed jailbreak-vulnerable models in violation of a cyber executive order, with high-stakes meetings now underway between Anthropic and top government officials.
Key Points
- Anthropic's most powerful AI models, Mythos and Fable, are taken offline after the Trump administration imposes stringent export controls following claims that the company failed to honor a recent cyber executive order.
- Tensions escalate after Amazon CEO Andy Jassy alerts Treasury Secretary Scott Bessent that Anthropic's models can be jailbroken, with administration officials accusing Anthropic of knowing about the vulnerability and deploying the models anyway.
- Meetings are now scheduled between Anthropic staffers and the Commerce Department, CIA, and White House science advisor to address compliance issues, as officials warn that a long-term problem remains despite the immediate crisis being averted.