Anthropic quietly pitches its own AI safety bill to lawmakers

The draft proposal would limit the circumstances in which the government could block the release of an advanced model and preempt some state AI safety laws.