Now this is an interesting topic, and I am optimistic on this since for starters, we are far from getting to the strong AI (I have seen machine learning and how it works, and it is absolutely fascinating) to work as we want them to. So AI is not moral or ethical because it is entirely driven by facts and data for the goal of improving whatever it is made to improve. Now since it is not emotionally connected to the lives of human beings it means that it is indeed going to have a good idea on how to improve things. I just don't see how we would ever be stupid enough to give AS-I the power to do what it wants since we don't understand it, because it is obviously going to be able to outsmart us all. I don't understand too much about what is needed for them to function that well, but if we can control it's ability to do what it wants and limit it with information and power(connection to the internet for example) then we shouldn't have that problem. Also we are going to learn much more about them as we go, this is definitely a question very much worth worrying about, but I think we are cautious enough, also the rules of robotics look like they don't have loopholes that a machine might use except for the part where it would think it's in our best interest to hurt someone, but then the other rule forbids it, so I think we might make it so that they don't take that kind of ideas into consideration at all, but rather work with what they can. Anyway Great post!!!