If AIs will be “corporations but faster and/or smarter”, and you accept that corporations behave unethically, AIs seem plainly worse.

I don’t think “AI = corporation” is a useful framework, current AI doesn’t particularly resemble corporations and corporations are much more human and less “intelligent” than I’d expect future AI to be.