I think that the TV-Show Person of Interest showed a bit about how that tech could be used and misused by government and groups that have other goals than the safety of the people in mind.
I think that AI can be a lot of help on a lot of subjects, but they should be very cautious creating them. But ok... we all know that humans don't make mistakes, so we're totally fine... :D
I am sure that the initial programmers will make sure there are no loopholes a super intelligent computer system can exploit...