It can serve bad purposes as well, if you have the right data. For example:
- AI can predict the “safest” routes to traffic drugs or humans, as well as the markets to sell them;
- AI can identify the most gullible victims to scam;
- AI generate fake voices of real employees(social engineering) to be used in anything from getting login information to creating panick among investors
- AI can mass generate “real” negative reviews of a business on Yelp, that might in essence sink that business