Anthropic Employees Accuse Trump Administration of Targeting Them
TL;DR
Anthropic employees reportedly accuse the Trump administration of singling out the company after Washington sharply limited access to its newest models, Fable 5 and Mythos 5. The order targets foreign nationals, including Anthropic staff inside the US. Anthropic responded by taking the models offline entirely rather than sorting access by nationality.
Nauti's Take
This looks less like mature AI governance and more like improvised power politics. A model with serious cyber capabilities deserves scrutiny, but an abrupt shutdown without a clear public standard is not safety policy; it is uncertainty dressed up as control.
Anthropic also made itself vulnerable by loudly warning how dangerous frontier models can be. Still, regulation should not depend on which CEO happens to irritate the White House.
Briefingshow
This turns frontier AI from a product dispute into a geopolitical infrastructure fight. If the US government can use export controls to cut off cloud access, the precedent reaches beyond Anthropic to every lab building comparable models. Without clear criteria, political alignment may start to matter more than measurable safety work.
Sources