Friday, March 21, 2014

AI: It's Role in Fiction (Person of Interest)

What is its the "referent" of AI in science fiction? By that I mean, what is it actually saying something about? One possibility, the "literalist" (rather than the literary), is that it claims that AI is possible in the real world, that it is possible for a machine's logical program to evolve into a living being capable of full human thought. This would be an interesting concept for which to use the word "evolution," because it is the same basic movement as is meant by the extreme form of the usual meaning of the term, only from the opposite direction. The usual usage of the term most often refers to the evolution of the physical, organic human body from lower animals. The more extreme form would say that everything about humans, including all aspects of psychology (even rational thought and "free will"), developed solely from animal psychology. According to the teaching of the Catholic Church, this would he distinctly hetrodox, as it denies a spiritual component and nature in humans (spirit being a whole other order of reality than the physical and biological). AI, on the other hand would be the evolution of pure logic (unfettered by animal body) into life on the same level as humanity. And it would be equally heterodox (replacing the mystery of the Incarnation with evolution, denying the former altogether, at least at some inevitable point).

But the question is whether or not this is what AI fiction, or any fiction, is doing (making metaphysical assertions). I would offer an alternative, which is that it can be doing analogy. All good literature deals with issues of human persons. Here the computer represents the rational side of humanity. The question is what the rational side of humanity does with the larger reality that is human nature.

Pure rationality can do cost analysis. This ultimately ends in a Jeremy Bentham utilitarianism. The thing that I like about the Person of Interest consideration of AI is what I mentioned in a previous post, Finch's statement that it was only when he taught the machine to value human life (sympathy) that it was able to understand human behavior enough to predict threats. The purely rational cannot, on its own, comprehend human nature without valuing it. Reason needs pathos in order even to do its own job of understanding human life and action. I am not sure if I would go all the way to saying that this pathos arises directly from the animal psychological drive for self preservation (I'm not sure, but that might violate my principles of avoiding any solely evolutionary explanations), but I do think that in the human condition (being both animal and spiritual) that animal self-preservation does play a role in the development of the necessary pathos.

An even more speculative move would be to include the other role that I usually assign to AI in good science fiction. Artificial Intelligence often, I believe, stands in for Artificial Identity. I would argue this is most often seen in the group identity, the whole that becomes more than the sum of its parts through human construction of the concept of that identity (sometimes conscious, but often collectively unconscious). The machine taking on a life of its own beyond the original program would be an analogy of the group identity taking on a life of its own, now defining the members, rather than the individual members defining the group.

I am not entirely sure how I would tie this together with the first concept of AI I started this post with; these are more just raw thoughts.  But it does seem to me like the conjunction of (1) AI being reason contending with human nature as including the biological and (2) AI being collective constructed identity provides a more interesting starting point for the question of the "greater good" vs the individual good (the tension between PoI's relevant list and irrelevant list) than does a purely rationalistic utilitarianism that gives no consideration at all to the question of valuing human life in the first place, and then simply tries to perform calculus on the value.


No comments: