To get almost Heideggarian for a moment, I think when we speak of consciousness in ourselves we are speaking of a different phenomenon to consciousness in anything else, be it our mothers or our toasters, because we simply have no direct experience of it in the same fashion, nor can we ever, because we are not those things.

Consciousness outside ourselves seems to be a compound idea born of a few experienced phenomena. The observed thing responds to stimuli from the outside world, but in so doing, it differs from we've seen other things do -- that is, immediate physical reaction, like a pool ball. In addition, those things we call conscious also show some degree of memory, and self preservation. This can perhaps most easily be observed when you see them do something that causes them harm or pain, and then see them avoid doing that thing again -- though pleasure seeking is the other side and no less evidential. Thirdly, they engage in SOME degree of reciprocity. I think this is why humans are so fond of other mammals more than reptiles, and are more likely to say a squirrel is concious than a snake. But, even reptiles, when faced with harm, will strike back, and when approached calmly, will often respond in kind, unless hungry or otherwise responding to stimuli we may not directly observe at the time.

AI thus far very clearly fails on the self interest front, and, largely reciprocity. The reciprocity it seems to show fails insofar as it is like billiards physics -- visibly pre-programmed.

I'm probably missing some key parts here, but at some point, I do think we get to a situation where for how we actually ever use the term consciousness-outside-ourselves , AI could conceivably get there. It's frankly still a very long way off, and that long way may well be subject to Zeno's paradox. Being able to fool some people though is closer, and for SOME people has already arrived, however temporarily.