The measure, known as SB 1047, was one of the nation’s most far-reaching regulations on the booming AI industry. It would have held AI companies legally liable for harms caused by AI and enabled a "kill switch" if systems went rogue.