Finally, a little common sense and logical thinking on the "dangers" of an AI revolt. Thinking, as we know it, is not what computers do. Making computers act as humans - there is a word for that - anthropomorphism. As a human, you project your own thought process onto others and other things. We do that for animals too; dress up dogs in clothes and talk to them like people. Dogs have feelings, yes, but they are not people. Machines have voltage levels, that are arranged in useful configurations by smart programmers.
I have a blog post that predicts what computers and robots will be able to do, when properly programmed (by people). That is, robots will be able to build just about anything, and will be able to do all of the tasks involved, from mining the metal ore, fabricating the computer chips, assembling the finished product. And one step further, robots will build robots, and also build robot factories that build robots. Robots will reproduce themselves, (and still not be "alive" in the animal sense.....) But when this happens, and it will, then all goods will be essentially free. When that happens, and it will, how will we as a species react and accept? I propose legislation now that will declare such robots son of robots, after enough generations, as ownerless or owned by all humankind, sort of like a copyright that eventually expires. Otherwise, without this declaration, all of the wealth will be owned by Bill Gates and his like.