Networks of artificial intelligence agents can plan, coordinate and run simulated disinformation campaigns on a social media environment modeled after X, operating without human direction once a bad actor has set the goal, according to new research from the University of Southern California