Anthropic Employees Accuse Trump Administration of Targeting Them
TL;DR
The New York Times reports that Anthropic employees see the Trump administration’s move as targeted pressure. The immediate trigger is an export-control order against the company’s newest models, Fable 5 and Mythos 5. The order blocks foreign nationals, including people inside the US and Anthropic’s own employees. Anthropic says it could not reliably separate those users from everyone else, so it disabled both models globally.
Nauti's Take
Anthropic should be careful with the victim narrative: a company that markets its models as both unusually powerful and unusually safe invites government scrutiny. Still, this intervention looks poorly grounded.
Using export controls over an allegedly narrow jailbreak is regulation by impulse. The user lesson is harsh: closed AI is convenient, but access is never fully yours.
Briefingshow
This shifts the boundary from chips and model weights to access to hosted AI systems. If a government can switch off a frontier model without clear technical standards, product roadmaps, compliance plans, and customer contracts become political risk. For Europe, the lesson is blunt: US models may be leading, but access can suddenly become foreign policy.
Sources