U.S. Senator Schatz pushes AI safety regulation legislation, vote blocked by Ted Cruz

U.S. Senator Brian Schatz of Hawaii, along with Democratic colleagues Mark Warner and Andy Kim, attempted to push a bill to a vote that would establish an AI Safety Committee under the U.S. Department of Commerce and require new models to be submitted to the federal government before public release. The vote was ultimately blocked by Republican Senator Ted Cruz, who said the warnings were scary but he was unwilling to give such power to the executive branch, and advocated a cautious bipartisan approach to address threats to critical infrastructure, nuclear, and biosecurity. In an interview with Civil Beat, Schatz said AI risk is urgent, that he cannot judge whether the probability of catastrophe is 10% or 2%, but he is unwilling to tolerate catastrophic risk at that level; he mentioned the Hugging Face incident as a turning point in public understanding of the risk, and advocated introducing third-party independent evaluation from the development stage rather than testing only before release.