I think this could only be possible waaaaay into the future.

In my opinion, once (if ever) the first AI is put in place to make safety critical decisions (and inevitably kills someone), the people will start rejecting it.