The goal of AIs is to create something substantially smarter than a
human in all domains, from science to philosophy, so that there's no
question of who judges the AI to be intelligent; the AI is a better
judge than we are.
The purpose of AI is to create something substantially smarter than
human, bringing about the next step into the future - the first truly
substantial step since the rise of the Cro-Magnons - and ending human
history before it gets ugly (nanotechnological warfare, et cetera). We
don't really know what comes after that, because the AIs are smarter
than we are; if we knew what they'd do, we'd be that smart ourselves.
But it's probably better than sticking with human minds until we manage
to blow ourselves up.