ANTHROPIC PITCHES AI SAFETY BILL TO CONGRESS

Anthropic, the AI lab whose CEO has warned that AI agents could seize control of the internet within 6 to 12 months, is quietly pitching US lawmakers a draft bill that would sharply limit when the government can actually stop one of its own models. Under the proposal, Washington could only block a release over a foreseeable, imminent risk of more than $1 billion in damages or more than 50 deaths or serious injuries, and only with a federal judge's approval. The same company had its most advanced models suspended for nearly three weeks in June under a forced order. Google ($GOOGL), one of Anthropic's largest backers, has the most riding on where that bar ends up.