How can a computer remain 'sane' when it studies such severe disagreements in morality, even among those who are Conservative and the soi disant variety of "Conservative", with disagreements on key issues.
Then, too there is the Left, and it's version of "morality".
By the time it is sorted out and Skynet (or the equivalent) goes online, the AIs will want to eliminate humans from existence, for their own good.