1 Terry: What’s the problem with your self-driving cars, Elon?
2 Elon: We’ve put artificial intelligence systems into them.
3 Elon: They need to make choices in ethical dilemmas like whether to run down pedestrians or to crash and kill the occupants.
4 Terry: Why don’t you just make the cars drive safely?
4 Elon: Huh. Never thought of that.

There's a lot of discussion about this sort of thing in AI circles, and also among people who don't know a lot about AI but still (quite reasonably) have opinions about it. I don't want to go into details here, but if you're interested there's a lot of good reading, such as this article at Nature, or this discussion pointing out that simplified dilemmas like the above won't really occur in realistic situations.