Anthropic’s proposed AI safety bill only allows government to block models under extreme conditions (Video)

Anthropic quietly shared a draft of an artificial intelligence (AI) safety bill with lawmakers. Under this proposal, the U.S. federal government could stop AI tools from being used only in very serious situations: if they kill or hurt at least 50 people, or cause at least $1 billion in damage. Nate Soares, the executive director of the Machine Intelligence Research Institute, joins the discussion to explain and analyze the proposal.