WASHINGTON \u2014 Lt. Gen. Robert Skinner stood offstage, awaiting his cue.<\/p>\n\n\n\n
When it arrived, the U.S. Air Force officer and director of the Defense Information Systems Agency palmed a microphone, pressed it to his mouth and walked into the spotlight at AFCEA TechNet Cyber<\/a>, a defense conference in Baltimore, Maryland.<\/p>\n\n\n\n As he proceeded, an uncanny voice came over the sound system. It sounded like Skinner<\/a>, who leads thousands of military and civilian workers at the Pentagon\u2019s de facto information technology authority. But the cadence was odd and the tone was off.<\/p>\n\n\n\n And, as those sitting close enough to the stage in May could see, his mouth wasn\u2019t moving much.<\/p>\n\n\n\n Skinner would later reveal the first 45 seconds of his keynote speech was machine-made audio \u2014 a product of 30 minutes of training, free voice-cloning tools and clever event production.<\/p>\n\n\n\n \u201cI\u2019m definitely not a good lip-syncer, but think about that,\u201d Skinner said. \u201cGenerative artificial intelligence<\/a>, I would offer, is probably one of the most disruptive technologies and initiatives in a very long, long time.\u201d<\/p>\n\n\n\n The stunt elicited laughter and applause from the audience, but also bore serious national security implications.<\/p>\n\n\n\t