A new study finds that, for robots, overlords are less persuasive than peers.