The world lacks an answer if machines begin to think independently, UC Berkeley professor Stuart Russell said AI company CEOs privately admit risks but public warnings remain rare, he said Russell said a Chernobyl-scale disaster might prompt government AI regulations