[JohnnyHerz]

The issue is not that there has to be "Conflict" as in AI versus humans. The problem is simpler, think about what happens when computer software has a bug in it... the result is failure of the system and everything dependent upon it.

AI will lead to "Failure" of systems from the human point of view. could that mean enslavement by our AI masters or worse?

Well, we are going to train AI on human knowledge and experience. So what does history tell us about the need to protect mankind against itself? Maybe that the risk of imbalanced motives is impossible to protect against?