Third camp: Friendly vs Unfriendly AI is completely irrelevant, for the simple reason that self-willed AI of either kind isn't going to happen, at least not in the foreseeable future, not until long after it's irrelevant for other reasons. I posted a longish argument for that a little while ago, I think it was to the AGI list; let me know if you want me to dig it up.
<br>