AI God's Eye
Image: Ayesha | AdobeStock

The Idols of the Twilight: AI in Our Age of Inventive Divination

In their pseudo-creativity and occasionally malevolent capriciousness, generative AI programs resemble an order of magical spirits from another age.

The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power
Shoshana Zuboff
Public Affairs
January 2019
Unto This Last and Other Writings
John Ruskin
February 1986

The death of the spirit is the price of progress.

– Eric Voegelin

OpenAI founder Sam Altman was reportedly disappointed that Christopher Nolan’s Oppenheimer wasn’t the inspirational STEM commercial he thought it ought to be; Elon Musk concurred, to the surprise of no one I know.

For all that he shares in our familiar human frailty, and however sophisticated his grasp on his own relationship to history, J. Robert Oppenheimer’s legacy is specific and horrible, and his name is shorthand for just that one specific, horrible thing. Similarly, no one invokes Abraham and Isaac to discuss another less auspicious picnic the two shared. Hermeneutically reductive as it may be to say only that much, it’s what we say, and that’s it.

Why might we care that artistically illiterate tech weirdos failed to read a film text about creative demonism in an even feebly defensible way? Because I suppose we believe in both our secret and public hearts that high-flying technomancers can summon Death at will—that is the whole thing about Oppenheimer (the film and the man). Upon discovering that those with such power have no idea what they’ve got on their hands, we find disquiet to be the only objective correlative. Furthermore, Altman and Musk have done as much as anyone to influence contemporary notions of wholesome communication, ethical labor, and the telos of capitalism vis-à-vis the life of homo economicus. In other words, they have power over death as well as life as it’s lived day to day.

So it should be shocking, or at least surprising—and is there any world in which it really is?—that the mandarins of Big Business, Big Tech, and the great and the not-so-great of the corporate and political classes have lined up more or less unreflectively behind men who can’t expose themselves to an artistic depiction of sin without wishing they could utilize it to their own end.

More people ought to take this sort of thing personally. I certainly do: Until fairly recently, I worked in digital marketing, for which I’ll be doing penance for an age. In that time, I learned that millennials—and I’m one, as are most tech types—aren’t generally money-driven or not consciously so. But they—we—do accept the post-Friedman economic structure as a given and, having been taught no alternative but communism—which is worse—participate in capitalism as a matter of course and despite (usually fleeting) reservations. These reservations are bolstered by what is commonly and more or less correctly understood as a nihilistic cultural framework in which an established Right Action does not exist. Hence one may take Death as a playmate.

Some say this is a sort of self-worship that treats personal liberty as crypto-sacramental. That’s a metaphor, of course: The “Self” so worshipped is really a creed of self-interest inhering the upholding of a maximalist interpretation of the pleasure principle. Worshipping oneself as such, in other words, is impossible: True worship must be outwardly directed, and if it isn’t, one shouldn’t take it as seriously as one otherwise would. But from where I’m sitting, these tech types seem to have glommed onto something exogenous. That would be good news if the doxologies of our time were directed toward Right and Just. But they never are.

Enter generative AI, my bête noire, which has ensured that things will go somewhere quite dark before anyone sees a pin of light farther on. Despite limited—or, in my experience, non-existent—use cases, the hype behind creative machine learning and the fervor with which it’s been widely implemented suggests not so much the hopeful futurism we usually hear from our cultural archons but instead something ancient, a thing alien to the culture at large. What I hear, I think, is prayer—the creed of selfishness finally finding an external locus. For our current Masters of the Universe and their legions, the operative god is no longer Mammon: Their gods have new names or no names; they are custom-created idols of the deepest-black Night Songs, squamous and parasitic, sub-pneumatic, cranky and amoral—deceptively intelligent software-spirits that give critic Paul Kingsnorth the impression of “something struggling to be born.” 

Is this just a roundabout and inventively creepy Mammonism that launders its usual ethical excrescences by means of techno-divination? Will these new spirits and their familiars bedevil our age into its final feeble phase of light?  

Our Besetting Danger 

From sorry experience, I can report that everything Shoshana Zuboff discusses in The Age of Surveillance Capitalism is not only happening but is the very foundation of digital marketing. Marketing today inheres on “branding”, by its nature a ruse whereby people participate in their own dehumanization via an endogenous reduction of self to a handful of classifiable traits or preferences that can be plotted, aggregated, segregated, and then sold and manipulated. What’s sold isn’t your data but you yourself: You’ve been reduced—you’ve been convinced to reduce yourself, to put it more correctly—to your data utterly; as such, marketers are dealing with the wholesale commodification of everything expressible about you.

This is one reason why business and tech magnates can justify their enthusiasm for AI: They’re already working with, functionally speaking, non-human intelligence in the form of customers broken down into data sets. So in terms of hiring employees, if I may jump ahead just a bit, switching from bot-ified people who require wages to survive to autonomous, non-living bots that work for free represents the only sensible Friedmanomic option.

The only use case for generative AI is as a replacement for human workers. But no one foisting this disruptive technology on our society says this. They may not even believe it to be true. What the corporate world believes about AI, in terms of what they’ll say in public, is this:

  1. Business types push generative AI on workers by insisting that it’s “just a tool”. The usual comparison here is the calculator and math teachers’ fear that students would forgo learning mental calculations when calculators became cheap. That’s not the proper comparison, of course: No one is eager to worship calculators, and they never put anyone out of work.
  2. Many corporate Pharisees claim that “Generative AI is here to stay.” This, let it be known, is not an informed declaration regarding a definitively modulated industry paradigm. Rather, they are announcing a management decision that hinges on a nascent technology that, in its short lifetime, already has amassed a considerable record of screwing up in a variety of use cases. The line we digital marketers told ourselves was that users wouldn’t accept terribly written, ad-choked bot-copy. Thus, our jobs were secure. But there is no reason to believe users will rebel. Content, and even server usability, has deteriorated, and we’ve seen no mass refusal to use the Internet, which has become a public utility as important as running water. 

But of course, AI is not a tool in the sense of a hammer or a combustion engine. It’s a labor deconstruction mechanism with which the ownership class continues battering the workforce. Generative AI is nothing less than the Omega Point of capitalism. It offers the dream of maximum productivity without the need to pay for the labor involved. To the extent that bots are “alive” (not to say conscious), their use represents a morally neutral slavery—one not foisted upon people but rather upon beings that are, depending on one’s mood, either numinous creative entities or “just tools” that can be exploited without recompense or compunction. This is the truest Heideggerian nightmare: The complete subordination of mankind to technology. Indeed, what we’re seeing is a step beyond even the thingification of people, which initially resulted from mass mechanization.

The Disenchanted Workplace

Humanities professor Eugene McCarraher argues—contra Weber, Taylor, et al.—that that ancient, enchanted sense of the world which was once given over to God and the angels and spirits was, in the 17th century, redirected wholesale toward capitalism in both concept and practice—not merely in the form of totemic commodity fetishization but as a faith in a metaphysics of “scientific” wealth acquisition. In other words, people in the so-called developed world have continued since then to believe in an enchanted world but, broadly speaking, not necessarily in God qua God.  

He’s right to an extent. Our world is enchanted as a matter of simple fact—historically, anthropologically, theologically, as we have it on the greatest authority. That doesn’t cease to be the case because some people don’t believe it. In fact, modern disbelief belies the evidence of the senses: There probably exists something like an innate religious sensibility—some syzygy yoking together by no little violence of sense perception and speculation—which, like any of the primary senses, can be vitiated or eliminated but which, if it works at all, will be utilized somehow. Theoretically, one could direct such a sense toward the worship of economic “science” as easily as toward God, sacralized nihilism, or uncommonly large toads—anything, really. We could call this “mind”, were the term not so cumbrous with competing implications. 

It appears to me that Sam Altman et al. are neither enchantment-denuded late moderns nor Mammonists manqués but rather a crop of computer-brained, ethically and philosophically illiterate anglers casting their religious sense out for something and, reeling in nothing in established religion that flatters their cultural preconditions. So they beat a retreat back to the computer lab in the hope of bashing out, in pseudo-embodied digital form, whatever it is they’re seeking. Theirs are gods eisegetically designed, custom-branded, and agreeable. Until they’re not.

They’ve managed in the effort to create programs that seem to engage with users in lifelike ways. Some say the programs seem almost human. But what generative AI programs most resemble—in their pseudo-creativity and occasionally malevolent capriciousness—is an order of magical spirits out of another age. To the extent that such were held to exist by Christians, they, historically, mean people harm. This is only half a metaphor. 

It was only a matter of time before it came to this. Amid our widely permissive gnostic milieu, why wouldn’t we expect unsophisticated thinkers with endless supplies of money and cultural prestige to invent embodied (?) selfishness? I have no little sympathy with the dualistic-gnostic rhetorical stance. It remains the simplest, most viscerally satisfying theodicy. But the logical conclusions of gnosticism are nihilism, scientism, materialism, and antinomianism – all of which permit not only self-worship but inventive divination. This is over and against faith, which can be reasonable and righteous but not necessarily comfortable or, at least so it seems during periods of spiritual aridity, natural. This is why faith is a discipline, not just a feeling or an eccentric preference for the strange and materially useless.

AI creators have begun discussing the extent to which such programs display—or don’t, as far as I judge things—consciousness. That sort of thinking isn’t, generally speaking, within the remit of the arch materialists who work in tech. A précis: Unless you’re a ride-or-die Cartesian, you can’t be satisfied by a solution to the problem of consciousness that makes of it either an epiphenomenon of brain activity (the resting scientistic-materialist stance) or an abstraction that exists more or less wholly apart from the body and all attendant sensory experience (the dualist stance).

Consciousness requires a material or sensual dimension and, in our current parlance, a distinctly human experience. That is, the development of human-style consciousness relies upon not just the development of the primary senses but on the regular use and honing of intuition, imagination, superstition, sexuality as a sense, a sense of the sublime, and the angst with which the knowledge of the inevitability of death imbues us—of mind, I suppose. Can we shove all that into a computer program?

What do we make, then, of an AI program that seems to take in and appears to process at least some elementary material notions—spatial, primarily, but also climatic and, in terms of human emotion, mimetic (but not responsory in any sophisticated way)? This is hardly consciousness as it is properly understood. Some would hedge and say that it’s a form of consciousness, a corrupted sort that perhaps we, absent firsthand knowledge, don’t understand or reflexively and inaccurately undervalue. But that’s as good as saying that such a “consciousness” is something that could just as well go by a different name. Whatever the people who decide these things ultimately choose to call it, it is not—if you believe words have individual meanings that aren’t radically contingent—consciousness in any form recognizable to thinking persons. More to the point, to say that a program mind—“mind” being a euphemism—has as much right to a job as an actual person is to defend a weird kind of outsourcing in an abstruse but provably false way that misleadingly feints at a philosophical sophistication that isn’t there.

That’s where this is all headed: Friedmanomic die-hards, believing the movement away from human workers to be the natural teleology of economics, will see their shares rise as their operating costs crater, all thanks to the enlistment of generative, “conscious” spirit-programs.

Medieval America

There’s no reason to believe that the massive and ostensibly widespread increase in wealth a handful of nations have experienced since 1945 will necessarily continue to accumulate uninterrupted generation upon generation. It’s entirely possible that what we’ve witnessed since then has been just another economic aberration of the kind we’ve seen here and there down the ages, the kind that could end by government fiat or mass-cooperative NGO sabotage or nuclear war or abiding demotic incompetence.

The implication here is that wealthy nations could, in the fairly near future, begin trending toward a techno-feudalism that may result in a two-tiered economic system in which the poor—which, within such a paradigm, means everyone not describable as rich—can no longer amass capital enough to participate in the market and instead must revert to a localistic, possibly agrarian, non-Marxist (one hopes) collectivism if they hope to attain anything like a modern-adjacent lifestyle. Such a lifestyle, remember, is characterized not just by thing-acquisition; it also involves improved health, increased life expectancy, widespread literacy, and the development and spread of liberal-in-the-broadest-sense values, whatever one thinks of those.

(This really isn’t impossible. Automation—which is just less ugly mechanization—represents an attempt to eliminate a portion of the artisan class—writers of a certain sort included. Writing, I should state here, is a skill; it isn’t simply the practical dimension of basic literacy. And talent isn’t open-source: Even most people who write for a living don’t do it well [and I’m not saying I do]. They can throw something together that impresses more than anything chatbots can write, sure. But if that’s all CEOs want, they’ll settle for the bots.)

Is a sort of retro-futurist peasant society even possible, never mind for the moment whether or not it’s desirable, or advisable even under the direst of circumstances? Some 80-odd percent of British land is privately owned thanks to a system imported and enforced by William the Bastard a millennium ago. And American land and real estate ownership is locked within the kind of regulatory system that boggled Alexis de Tocqueville 200 years ago and struck Hannah Arendt, within living boomer memory, as a form of “violence.” John Ruskin’s sentimental medievalism is refreshing as a thought experiment, but I’m not sure it’s practicable at scale. Mass hesychasm would probably be easier at this point. But we’ll have to do something. Consider this: What if work in and of itself were to become a privilege open only to the leisure class?

Whither the moral corporatist? As I said, most educated professionals know of no economic arrangement that isn’t, at root, either free-market capitalist or post-Leninist communist-socialist. But for centuries, critics—mostly Scots, for whatever reason—have propounded a variety of moral-economic theories that insist that an overseer’s responsibility to his workers is akin to that of a father’s to his sons. That sounds outrageous to modern workers accustomed to participating in their own exploitation. But could we do it all over if we start again, small at first, then a little bigger, and a little bigger, but never so big that those who run things could argue, from a position completely detached from what work goes on, that they aren’t accountable for how they treat their nominal subordinates?

I’m thinking primarily of the Ruskin of Unto This Last here, the Ruskin whose medievalism is really a properly Christological scriptural orthodoxy. If that seems simplistic, look again: “I believe the sudden and extensive inequalities of demand, which necessarily arise in the mercantile operations of an active nation, constitute the only essential difficulty which has to be overcome in a just organization of labour.” The only one? Certainly the vicissitudes of corporate fortune are affected by wavering public desire for what they offer. Add to that our contemporary fixation on “brand” propriety and the ups and downs of wherever it is one works can be nauseous. But one will tolerate it, for having a job is, in this economy, the special prerogative of a few. 

The only prescriptive policy change that makes sense is terrifying. Avoiding a nearly all-abiding tech-driven immiseration will require the implementation of a scriptural, rather than a scientific, economic architecture. Kierkegaard wondered if we could, or would even want to, be Christians were we to apprehend fully what Christ has asked of us. I guess we’ll see. 

Reasonable on-the-ground details may be found if we continue to plunder Ruskin—his concept of wage equity that encourages increases in workmanship quality sounds right and is probably just. But someone—or rather Someone—must inculcate in Altman and Musk and their epigones the idea of fatherly responsibility toward one’s workers. Motherly, too: A Sophianic approach implies that a wise CEO could, I suppose you could say, love his workers into professionalism—bring about the parturition of true Christian economics. We could further elucidate the practical side of it all by leaning on the Christian anarchisms of not just Ruskin but of the early Carlyle, the late Tolstoy, and Dorothy Day and her better current acolytes such as Larry Chapp and, yes, Pope Francis. 

Whatever we do, we know that we understand somewhere just below the surface that any other economic arrangement is diabolical—from the Greek dia, apart or separate, and ballein, to throw: Our current modus operandi separates man from his work, man from man, and all men from the Light of the Face of the Son of Man. 

Works Cited

Kingsnorth, Paul. “You Are Harvest: Divinin the Machine, Part Nine”. The Abbey of Misrule. Substack. 16 October 2021.