The most we can honestly say about artificial intelligence is that we have so utterly no idea what it is that it might be possible. Of course, we have no computing paradigm for it either, so that's on the TODO list as well, when and if the raw power becomes available.
I honestly hadn't considered that something could be considered intelligent without being conscious, given that we have no applicable definition of "consciousness" either. I understand that many researchers fear loss of all funding if the real state of their field becomes widely known, and I'm onboard with that since I think the research is worthwhile even if it's beyond the congresscritters. I won't pretend that it has accomplished much as yet, though: as I've said before, we're a heck of a lot closer to building a warp drive than than a conscious computer.
If we encountered a "superintelligence" that did not display consciousness, would we be justified in treating it as a machine to be used and turned off rather than a lifeform to be talked with? Even if it could talk, in a sense beyond a fancy shell or an Eliza bot? Could such a thing come into existence on its own? An organism descended from an alien race that uploaded itself doesn't really count, to my mind, but it seems by far the most likely case.
I could agree that such intelligences wouldn't be very interested in us. Earth has too much gravity and oxygen just causes rust; all asteroids lack are organics they probably don't need anyway, and heavy metals are much easier to reach on an asteroid. Given a reasonable power source other than a star, they'd be better off living in interstellar space where no one is likely bother them.