Anthropic is pitching Congress its own AI safety bill, report says

Anthropic has reportedly been privately lobbying U.S. lawmakers with a draft bill aimed at regulating frontier AI safety. According to a leaked document, the proposal sets specific thresholds for government intervention, allowing federal officials to halt AI models only if they pose an imminent risk of significant financial damage or severe physical harm. The draft suggests designating the Department of Commerce as the primary regulatory authority and includes provisions that would override certain state-level AI laws. This move follows a broader trend of major AI companies attempting to influence federal policy as Congress considers various legislative approaches to catastrophic AI risks. While Anthropic has not publicly endorsed a specific federal bill, the draft aligns with the company's previously stated 'Advanced AI Framework.' The initiative highlights the ongoing tension between industry-led safety standards and the push for comprehensive government oversight in the rapidly evolving artificial intelligence sector.