I think the manipulation of the three laws of robotics in Isaac Asimov's "I, Robot" is a simple but powerful demonstration of the problems we will face when trying to program AI.

I just think there are so many variables in this world and the interpretation of a set of logical arguments that are programmed by humans into a machine that has achieved artificial general intelligence cannot end well for humankind.

We're simply not smart enough to predict all the outcomes. I mean, the richest man on earth couldn't even deliver on a crash-proof operating system (nor can he count to 10 for that matter).

Elon Musk is a great man and I believe there are many things he will continue to do to improve humankind, up until the very day the singularity arrives and, with it, our robot overlords. :P

It's an unfortunate set of circumstances with robots being programmable and humans programming them in a human manner that some robots are going to be programmed to kill.