“The bipartisan AI bill from Obernolte and Trahan would require leading AI labs to create plans to address catastrophic risks, with compliance verified by third-party auditors.”