Yeah, I realize that my comment could be taken as implying that people's "revealed preference" is to lose their autonomy to AI. No, that would be a bad thing. W

Yeah, I realize that my comment could be taken as implying that people's "revealed preference" is to lose their autonomy to AI. No, that would be a bad thing. What I meant was really that people fail to recognise the potential for adversarial behaviour in seemingly impersonal automatons, which is a problem that predates AI. But people can learn.