When we talk about consciousness, as in the subjective experience of being, that's ultimately a proxy for "which things can we do really awful stuff to". We reason that for things that are not conscious there is no suffering (since it can't be experienced) and therefore we are free to exploit those things in any way we wish. Lately the trend has been to argue that AI is not, that even though we can create algorithms complicated enough to do the kinds of analysis our own minds do (and in a not-entirely-dissimilar way), they are "just algorithms" and have no experience of suffering for us to concern ourselves with. This is convenient for us, because it allows us to exploit such things without being bothered by ethical reasoning. This is also why many will argue that animals are not conscious, and some will even argue that many humans are not [0].
[0] Which is a perfectly rational statement if you operate from a basis of empiricism. There is no way to measure the consciousness (again, the experience of being) of something, meaning you can only observe your own and everyone else, though similar to you, may as well be an automaton.