As far as that last part goes, I think we already have ample evidence that bots can, if not have emotions, then pretend that they do (including wrt their decision making) well enough for humans to treat them as genuine.