Anthropic Employees Accuse Trump Administration of Targeting Them
TL;DR
Anthropic employees are reportedly unsettled by internal discussions around the Trump administration’s move. Some staff see the order less as routine oversight and more as targeted pressure on the company. The trigger is an export-control directive covering Anthropic’s newest models, Fable 5 and Mythos 5. Anthropic took them fully offline because it could not cleanly separate foreign-national access.
Nauti's Take
Security concerns around powerful models are real. Pulling the emergency lever without clear public criteria is still bad AI governance.
Anthropic has spent years arguing for strict guardrails, which makes process even more important here. Turning models off through ad hoc directives does not build trust; it turns every AI roadmap into a Washington weather report.
Briefingshow
This is bigger than one company’s fight with Washington. If cloud access to frontier models can be treated like an export-controlled good overnight, AI infrastructure becomes a political dependency. Companies, governments and developers now have to judge not only model quality, but whether access can survive the next policy shock.
Sources