Friday, October 18, 2019

The A.I. as Thinker

"The Thinker"by Rodin (photo: Douglas O'Brien)
Can A.I. be proleptic? Can it answer questions before they’re asked? As amusing as this may sound, those software engineers who traffic in self-driving cars routinely have to account for contingency. Otherwise, how would they deal with anything that comes in a vehicle’s path including those elements of happenstance which exist only in the realm of possibility? A.I. has shown its mettle in precisely this manner. If you remember Deep Blue beat Garry Kasparov in l997. But there's a kind of black hole into which A.I cannot go unchallenged and that's in the realm of mind itself. Could a computer "conceive" of Bishop Berkeley’s famed, esse est percipi, “to be is to be perceived?” One would imagine that the premise of A.I. is to exponentially increase the realm of possible outcomes in any one situation, but certain kinds of cogitation have nothing to with decision making. When you make a statement about thought itself like cogito ergo sum there's no matrix of possibilities, no answers that are being elicited. It's a de facto meditation on being whose only significance falls into the field of ontology—a discipline which ironically would be out of the realm of computers programmed to tackle baser forms of human reality, which is to say, action in the world. A.I. could note the "bracketing" or "epoche" in a Husserl text analyzing objects, but could it parse the difference between Kantian noumena and phenomena, in any meaningful way or for that matter demonstrate the capability for self-reflexive consciousness? In short can A.I.s think like philosophers or are they just number crunchers at heart?

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.