I have been talking about the knowledge that we—as humans—create, discover, learn, store, build on and extend, manipulate, communicate and share, and use. I have also posited that such knowledge is inevitably based on the axioms of sensory input [1].
But as we stumble into the new world of AI, is it possible that machines will create, discover, learn,... knowledge that is not based on human sensory metaphors?
If so:
If not senses, what would they base it on?
What might it look like? Could such knowledge have different, even radically different, characteristics than "human knowledge"?
Would we—grounded as we are in our sensory metaphoric structures—even understand such knowledge?
If we don't or can't understand this new "machine knowledge", where would that leave us?
If machines do create such knowledge, it might well be quite, um, alien. So, let us—with our oh so limited sensory metaphoric knowledge basis—try to answer these questions...
FOOTNOTES
[1] That is, I have based this hypothesis from my reading of a large body of work by people much smarter than I. Specifically, the work of Lakoff and Nunez. I would not have realized this by myself.