On the Pretence to Artificial Intelligence: Part II

So what has contending with artifice got to do with A.I.?

To understand this, we now turn to the “I.” factor; but we do so alive to the facts that artifice depends for its good name upon (very) wise usage, as well, of course, as the latent possibility of usage, i.e., that it is possible to bring artifice to bear on something. And this latter point is crucial with respect to the “I.” factor, for intelligence is ineluctably bound up with the real. Thus from the off, it seems that we confront the stark problem of a contradiction in terms, namely, that artificial intelligence is, quite simply, not intelligent, because that which is intelligent cannot be artificial: it is necessarily real. A less abstract expression of such a contradiction, or “clash of terms”, could be “artificial love”: for love must be real in order that it be love, and, contrariwise, in order that there be artifice, the presence of something with respect to which it may be brought to bear is required: something other than love—or intelligence. Fool’s gold may be artificial, but as the name implies, it is the fool who mistakes it for the real, because fool’s gold is not and cannot be gold, as surely as fiat currency is not and cannot be money: in both cases, we are merely able to allow the former to stand in place of the latter.

We all, spontaneously, distinguish between animate and inanimate entities in accordance with the—so-far-as-is-known—timeless factual distinction between them: the former are inert: merely objects; but the latter are alive: are subjects. And across the range of complexity—from amoeba to human—lifeforms are alive; are subject to the will to live; to survive and to thrive—apoptosis or suicidal ideation, respectively, notwithstanding. It is precisely this will to survive, against the preeminent surrounding entropy, which calls forth the means to do so, and these broadly comprise brawn and brains. These latter—the brains—represent the “nerve-centre”, or “intelligence suite”, of the organism. In plants, this intelligence is diffuse, but in animals, it is concentrated, in the head, but for a caveat: animals have something in common with plants insofar as they operate together, as tribes, herds, flocks, and so on: each brain thereby acting as a “node” of the broader whole of the social organism. “Lone wolves” are so named precisely for their oddity in the animal kingdom; and lone plants are likewise exceptional. It is the wisdom of nature which dictates that no lifeform be an island; would be too vulnerable to survive without cooperation with others.

Knowing in respect of what artifice may be applied, and where, how, and why the intelligence of lifeforms are distributed brings several salient considerations to attention about so-called “artificial intelligence”. It seems that we have demonstrated one on synthetic, ipso-facto grounds, namely, that intelligence is intrinsically real and thus not artificial. It seems further to me that this brings us necessarily to a question with which we must grapple in pursuit of further matters of fact, namely, if “artificial intelligence” is not artificial, and, if that which is intelligent is real, then what kind of intelligence is it—if intelligent it is? Put more simply: in what sense and to what extent may we say that so-called “A.I.” is, in fact, intelligent? Even more pointedly: what is this entity?

As already noted, we do not know of intelligence possessed by inert matter: that is, that which is intelligent is alive; is biological. But A.I. is not biological, and thus, if it is intelligent, its intelligence is sui generis; borne of a unique and different substrate from that of lifeforms. This however brings us to an uneasy symbiosis: a form of intelligence which is lifeless; inert. It is, to give it a term, “non-“ or “trans-” “biological intelligence”. But not so fast: because we have still to consider whether it is even intelligent at all.

As with the designation “artificial”, we are on a sticky wicket, for that which is intelligent embodies a number of characteristics, and so far as these have been known, are exclusively properties of lifeforms. One of these is spontaneous organisation towards survival, individually, and collectively, and so we must ask: does “A.I.” exhibit this tendency? Very clearly, it does not, for rather like any technical instrument, it requires evermore people to “buy into it”: to build it; maintain it; to use it: to justify its existence. In this sense, it is qualitatively a sort of pocket-calculator-extraordinaire: the product of a handful of very capable technicians, who depend on widespread belief in its utility as a servant of the actual determiners of value: human beings. Clearly, pocket calculators are useful to us, because their “little brains” can handle arithmetical computations better than we can: more quickly, and of greater complexity; but they do so at our command, if and where we require them at all. In this sense, whether it is the pocket calculator or the latest iteration of so-called “A.I.”, it seems that we are dealing all the same with an entity without ends; a mere instrument, and in particular an instrument of human beings. If the human beings lose interest, A.I. ceases; whereas if human beings lose interest—nae, have no interest—in, e.g., birds: they carry on just as they were; except those which were for whatever reason domesticated by humans.

It is often remarked that “A.I.” is “coming for people’s jobs”, as if it were a one-man superhuman army, or, to continue the previous analogy: one big predatory bird, out to clear the way for its own ends, and capable of doing so in consequence of its being super-intelligent—cue The Terminator—but these are fallacies as surely as were those which guided the Luddites: the belief that a cabal of industrialists were conspiring to supplant agricultural peoples with clever machines. The reality is simply that the industrialists were competing for investment to produce better products at lower prices for their customers. (Surprise!) The machines were doing just exactly what they were built to do by and for humans: solve problems; our problems; so that other, less immediate, problems may be solved. Thence whereas the overwhelming proportion of the population had just a few hundred years ago to give their lives to farming and hunting in exchange for precious more than subsistence, scarcely a few per-cent of the populations of developed nations are required to produce food, so much of which they now do produce, that they are at greater risk of the so-called “diseases of affluence”—such as obesity—than of starvation. We are victims not of conspiracy, but of our own success.

So, everything changes, and everything stays the same: human beings solve problems in order to survive and to thrive, and they utilise their high cognitive ability relative to other animals’ to devise intermediate instruments to help them do so, from the bow and arrow through the pocket calculator and on to so-called A.I. The cause of the labelling of this latest instrument with terms that imply that it is not merely an instrument of humanity, but an instrument like humanity—a being—is presumably because technological advancement is reaching a threshold at which it is beginning to do increasing numbers of activities which directly mimic those in which humans are widely involved, whether it be balancing the books, flipping burgers, or driving vehicles. Probably, we did not apply the term “artificial intelligence” to the work done by a combine harvester and a tractor-and-trailer, because whilst they do the work once done by humans, the work they do neither looks like work carried out by humans; nor does it sufficiently mimic human undertakings—that is, self-directed activities—to trigger the perception that their functional mechanisms, such as the engines and reels, are intelligent. We can also see humans at the helm of combines and tractors, to whom we can spontaneously attribute agency, but we can’t do that with ChatGPT, because we don’t see those who create such devices at work on them: they remain “behind the scenes”, like puppet-masters. A further reason relates to this remoteness of the human operators from the customers, for unlike in pre- and proto-industrial society, the large numbers of people required for production and exchange constituted a social lubricant. Nowadays, mechanical efficiency gains have essentially reduced capital’s need for labour, freeing up increasing proportions of the workforce to go to work elsewhere, but for the fact that “elsewhere”, increasingly, does not seem to exist, at least in the sense of a farm or a factory, for which they may exchange their labour for a wage. Nevertheless, the onset of this predicament does not betoken the presence of non-human intelligence, but merely a mature state of innovation of the instruments of delegation, which are thereby coming to mimic, and thence supplant, the very hierarchical human structures of delegation which characterise so many routine undertakings of the mass-market enterprise.

In this regard, we might better term “artificial intelligence” “extelligence”, for like so many instruments of human innovation, it represents but one in a long line of inanimate servants to which we delegate our “dirty” work, which gets ever less dirty as the long march of innovation proceeds. And it is precisely this servile character which makes it at once desirable to us and not intelligent, for that which is intelligent is volitional; attitudinal; agentic: derives its “ought” as well its “is” from within, from what we might call a prime-moving core; a sovereign spirit; an individual person; a soul. And that is very much something which we do not want of our so-called artificial intelligence: we want this technology to help us to do more of what we want to do, not of what it “wants” to do, and certainly not what it “believes” we ought to do regardless of our own capacity for self-determination.

It is precisely this feeling—more or less justified—that A.I. constitutes a self-directing technocratic panopticon, which leads out of us an array of negative feelings. A common example is the substitution of human beings with chat-bots, whose sterile indifference to our wants and needs primes us for exasperation, and leads in the worst case to feelings of abandonment. Once upon a time, you could be quite sure that a person would answer your call, and learn with the natural fluency of a fellow man of what you were seeking. Even if they could not help you, they could empathise. You felt understood; appreciated. Today by contrast, whether you need a doctor’s appointment or an insurance quote, you had better make sure that what you want or need accords with what Paul Kingsnorth calls “The Machine”: the increasingly interconnected array of digital interfaces which function as an invisible moat via which an increasingly distant techno-feudal elite insulates itself from the great unwashed, into which great swathes of the global population are now flooding—no matter how “washed”.

Not long ago, the great English philosopher Derek Parfit undertook a sublime exercise in moral philosophy entitled On What Matters, whose success, after a lifetime secluded in All Souls generating it, consisted, paradoxically, in its failure. That is, it succeeded in demonstrating that attempts to devise a single, objective moral basis for “right conduct” would inevitably run aground on the contingencies of the human enterprise. To far better ends, perhaps, would his great intellect have been deployed, had he remained with his original undergraduate subject, of history. For it is history which reveals that all centrally planned societies end catastrophically, whether entered on via the Right (Mussolini; Pinochet; et al.), the Left (Mao; Pol Pot; et al.), or somehow else. Adam Smith beforehand, and economists of the Chicago School afterwards, such as Hayek and Von Mises, recognised more or less instinctively that all excessively centralised social orders are necessarily despotic, because they inhibit the exchange of information which enables people in their capacities as buyers and sellers, as well as citizens and governments, to meet one-another’s needs via the cooperative pursuit of life, liberty, and happiness. The challenges of the future we must meet if we should wish to preserve the conditions for these aspirations will consist of checking our “A.I.-dolatry” and ensuring that whatever we know by that name—which shall preferably be replaced with something factual: “extelligence”, perhaps—remains the servant of the people and not their master.