Points 4-7 are predicated on point 3 being effective, but it's just a benign-seeming action.

I also disagree with point 1, but since you just mean that unfriendly AI is something I wouldn't want around, I'll let it slide.