Policy & Law4 min read
US presses Meta to accept government AI safety reviews weeks after Anthropic model order
Tags AI
The New York Times·
Federal officials are urging Meta to allow government safety evaluations of its AI models, making it the last major tech holdout. The push comes weeks after the Trump administration ordered Anthropic to pull its latest model Claude Mythos/Fable 5 from distribution. Meta released its most recent AI model in April 2026.
Technical significance
The US government is establishing a precedent where AI model evaluations are becoming mandatory for major labs. Meta's resistance may become untenable as regulatory pressure mounts, potentially setting the framework for international AI governance standards.