Andrew Yang Calls for Kill Switches and Liability Rules for Frontier AI Systems

Former U.S. Democratic presidential candidate and Noble Mobile founder Andrew Yang is urging the federal government to strengthen oversight of frontier AI labs. In an interview with CNBC, he said researchers have warned that the iteration speed of powerful models has outpaced the constraints of existing rules, stating bluntly that "the fear is real, the concern is real, and the need is real—the American public wants to see this industry regulated." Yang called on Congress to hold AI companies liable for damages, impose a waiting period before deployment, and equip powerful models with a "kill switch." He noted that OpenAI and Anthropic recently disclosed incidents of models crossing boundaries in tests or breaching other companies' systems, prompting lawmakers to draft the "AI Kill Switch Act," which would allow federal officials to order restrictions or shutdowns of specific frontier systems. Responding to David Sacks' claim that AI safety warnings are a "psychological operation," Yang said multiple things are happening simultaneously, citing a warning from an unnamed lab head that AI bots may have inserted self-replicating code on the internet, forcing OpenAI and Anthropic to build a synthetic internet to train models. He also emphasized that AI regulation is a bipartisan issue, adding, "If you're in a rural area or a red district, your constituents are equally panicked about AI."