Trickery

I was fooled the other day by the first paragraph of a recent piece by Sam Kriss in The New York Times Magazine, “Why Does A. I. Write Like…That?” The opening was apparently the work, or an imitation of the work, of artificial intelligence, and the author’s point was to illustrate how passable and ubiquitous this prose without a mind has become today.

In hearing so much about artificial intelligence and machine writing, and like everyone else encountering its language and presence at every turn, I am fast coming to the conclusion that a large measure of the discussion about A. I. is missing the point. Whether the language today’s machines produce is lifelike or not, whether it is convincing and indistinguishable from natural language or not, is merely a necessary question in the quest to better the efficiency of the technology. The linguistic engineers will work and worry to smooth out the phraseology and contemporize the syntax, but their efforts will remain objective and their object will always be a usable and profitable product. Their attempt to replicate the human serves that end in view.

Producing objects, after all, is what the applied sciences and their industries do, by definition and rightly so. When we do something, we want to do it well because we are making something to accomplish an external purpose both pragmatic and practicable. So the question about machine writing, and all the more about artificial intelligence, should not be how well it is working; all concerned better be asking that question. The real query is, what is it trying to do? And if the quick answer to that latter question is to make a machine that acts and appears as human as possible, personality and all, then a third question should obtain: why? Is not producing the information—an airline reservation, a bank balance, some medical advice—enough without the pretense of a human relationship?

Implicit in every question of purpose is an assumption about the nature of the agent. How we answer why the imitatively human manner of artificial intelligence should preoccupy us so will help us articulate just what we believe real human nature, the agent creating the technology, is in essence. With our belief about that in hand, then, we will be able to decide more rightly whether we should limit the reach of A. I., or let it roam wild and free to enter into every interest its creators have in their own hearts, including matters once classed as the artful and more distinctively human. Industry is after something the humanities are not. Both the applied sciences and the arts are productive, but the former produces a pragmatic object of prescribed end, while the latter produces an object not pragmatic but noetic. It is a question of fabrication versus realization, a transitive versus copulative movement of the mind.

We come to be able to accomplish something in the world objectively through technology, but we come to understand ourselves intuitively through an art as something more than an imitable personality. If we are prepared (and we might very well be) to assert that our machines too can imagine extralinguistic realities, or if we persist in confusing an image with signage, evocation with indication, then we have our answer: we are assuming that human nature itself is nothing more than a contraption, something put together by nature, and therefore ultimately bounded by itself, self-enclosed and self-referential. There is no such thing as an extralinguistic reality, this view might continue, because we can have all we could ever wish to have through the manipulation of metalinguistic elements. To suggest that we comprehend meaning subsisting through form and not produced by the form, linguistic or otherwise, is to this way of thinking hopelessly naïve and scientifically unfounded.

That, however, is not the traditional, long-standing, transcultural understanding of the meaning and purpose of art and language. Complementary and not opposed to the outward looking sciences, the humanities are understood to look inward, past the cogitating mind, first to its reflecting conscious surface and then more deeply still, more contemplatively to the mind’s awareness, watching for what is arising to its insight as the source of meaning looks for the next shape to take through the hands of the artist. To understand this as we read or write or paint or sculpt or dance, we must be able to give ourselves over to the artist whose integrity we trust; otherwise, we naturally hold ourselves aloof from what we are trying to understand and remain forever at a conceptual distance. But to give our trust to what we know in our heart is a machine demeans us. The human being writing is not arranging syntax to manufacture meaning, but to symbolize what has been inwardly seen and felt, what concurs with being human. The work of the arts is from this inward ideal, from a vision of meaning prior to the medium in which the artist will give it form. Only this interior depth can give what is meant by the resonance or timbre or maturity of a poem or paragraph or building.

But we seem instead bent on simulating this deeper human presence, and the result, like that first paragraph of The New York Times piece, is a now normalized duplicity we conspire on one another. We accept trickery as part and parcel of our being together. After long years of self-promotion and competitive self-advancement, we take it now as the regular course of business that of course our politicians lie, of course what is advertised is not true, of course the way things really are must more often than not be dusted with sugar. Let’s not be quite so idealistic, we tell ourselves, and with that sophistical self-defense we shape our own future, needing no greater cause, absent turning inward in the direction the humanities bid us, than the wish to make the world just what we, each of us individually, assertively, competitively, wants it to be.

The wish to deliberately create a false appearance of the real and true is tantamount to the wish to deceive. And if that is true, that intention takes us onto moral ground, by which term we need mean no more than its original intent: what pertains to human character, the source not only of our thoughts, but of our actions as well. Why we wish to fool one another, it seems to me, might prove a better question to consider about machine writing and the fortunes of artificial intelligence. Why do we not instead prefer the truth?

***

A Word About Last Week

Last week’s post went out as usual Tuesday morning, but because of a technical problem I still don’t fully understand, a number of subscribers did not receive it. I believe the problem has now been fixed, and for those who missed it, you may read last week’s post here: Ante and Post. My apologies for the mishap.

***