Thank you for your reply @akdx.
I think AI is not the problem, we are the problem.
I think you got the point. Especially considering that our concerns about AI are always connected with typical human behaviors and approach. However in human's world such situations are not very likely to happen, as society has learned how to prevent them and control extraordinary individuals. The problem begins, when AI takes all of our negative behaviors and becomes more intelligent. We would not be able to control it.
We want to make our life happy but the quest of happiness is a mirage.
That's true. The question is: would AI understand the concept of happiness? If so, would it be its desire?