A Tech Skeptic's Apology

In defense of the good of the given

Last week, I published at Commonplace a modest essay entitled “Tech Policy’s Libertarian Hangover.” It was hardly my most ground-breaking or earth-shaking writing in this space, intended rather merely as something of a ground-clearing exercise—to cut away the underbrush of often frankly terrible arguments that get in the way of serious thinking about how to shape and govern the technologies that are currently shaping and governing us. The essay provoked a surprisingly strong reaction among some techno-optimists, however, inviting the characterization that I offered a full-blown Luddite rejection of technology, calling for us to stop AI in its tracks.

Now, I will be the first to concede that perhaps some nuance was lost under the editor’s scalpel, as more than 1,000 words were cut between submission and publication—and I must own any resulting loss of clarity. That said, I will have to warn the techno-optimists up front that this is an apology in the classical sense—of Justin Martyr or G.H. Hardy—not a retraction. On further consideration, I think the claims of the essay are actually quite clear, although it left much more to be said. This essay today represents an attempt to say some of that “much more,” and to wrestle through some of the more interesting responses that the essay provoked.

On the Best Way to Stop Runaway Trains

First, let me address a pretty straightforward misreading of the piece: that in it I called for a Buckleyite stance of “standing athwart history yelling ‘Stop’.” This line appears in the second main section of the piece, responding to the cliche that “You can’t turn back the clock.” In fact, though, it was not my purpose to endorse this stance, which has always seemed to me a grievously truncated view of what it means to be a conservative. Far better is Edmund Burke’s understanding of conservatism, which I quote further on in that same section, as a call to “slow down” rather than “stop”; for conservatives have always understood that when it comes to the health of living things or bodies politic, the pace of change is at least as important as the direction.

Someone objected that I seemed to be endorsing Buckley’s maxim because I emphasized the need for “braking”—what, he demanded, was the difference between “stopping” and “braking”? I don’t know about you, but living in rural Loudoun County with crowded curving two-lane roads, I press the brakes of my car probably several times every minute; however, I have such an antipathy to stopping that I bought my house partly on the basis that there are zero stop-signs or traffic lights between it and the highway. Stopping, in other words, is a subset of braking, and hopefully a rare one at that. Even more importantly, though, it calls for a fundamentally different position. Buckley’s metaphor is of a fearless but foolhardy protester standing on the tracks in front of a runaway train and likely to be run over. “Braking,” though, is something you can only do from the engineer’s cab. In other words, we can only apply the brakes to technology from the inside: by daring to be in the rooms where the technology is being built or where policy is being shaped.

Of course, this is not to say that only such insiders deserve a voice. On the contrary, we urgently need political institutions that can bring outsider voices to bear as well; the engineers must hear from those whose communities are in the path of the runaway train. This was the point I sought to make with my invocation of the much-maligned Luddites, who were objecting above all to being totally cut out of the process when it came to deliberations about how to adopt technology well in the workplace. Their most memorable acts of defiance—actually smashing the machines—were theatrical protest gestures signifying not an opposition to all technological change, but the fact that no other option had been left to them. If you do not consult someone on the front end, do not be surprised if they decide to just scuttle the whole project in order to get a hearing. So it is today that communities across America have been mobilizing to demand moratoria on building data centers—a movement greeted with shock and horror by most in DC policy circles. Indeed, many on the tech right have cited this opposition as proof that the American people simply cannot be trusted to give any input on the AI revolution, when in fact it is simply proof that if you refuse to give people a voice, they will grasp desperately for any point of leverage still available to them.

The essential question of tech policy, then, is one of agency: do we have the will, the imagination, the institutions, and the patience to exercise agency over the technologies transforming our lives—to actually consult, deliberate and decide on what we want to do about them, rather than simply lying down in front of the train or letting ourselves be dragged along behind it? Of course, the answer is not entirely up to us, for there is no simple “us”—we are made up of warring factions and interests, and even if the American people could speak and act with one voice, we are not the only agents in the world. You may notice that among the cliches I sought to debunk in the essay, “But we need to beat China” was not one of them. It is a tired talking point to be sure, sometimes invoked in the most absurd of contexts, but it is also not easy to dismiss: we really do find ourselves in a moment of great power competition, trapped in a prisoner’s dilemma from which there is no easy escape. From this vantage point, “pausing AI,” whatever that might mean, is not really an option, even if we should want it to be. My inversion of Trump’s mantra, “You can just refuse to do things,” was somewhat tongue-in-cheek, for it is never that easy. While I critiqued the cliche of technological inevitability in my essay, I also conceded that there is some real truth to it; horse-drawn buggies and gaslight were at some point doomed to go the way of the dinosaur.

The Tragic Side of Technology

But if that is so, my next question is, “Are we allowed to be sad about it?” Is there a tragic dimension to technology, or must we stand on the sidelines like sorority girls insipidly smiling and reciting our cheerleader chants? I myself am actually a comparatively early and enthusiastic adopter of AI, to a degree that sometimes scandalizes my friends and family, but I harbor real sympathy for those who can see it only as an engine of devilry. If I can accomplish nothing else in my work, I hope at least to create space for humanistic laments such as those of Anton Barba-Kay’s majestic A Web of Our Own Making or Greg Conti’s extraordinary essay published last week at Compact: “The AI Apocalypse Is Already Here.” If you haven’t yet read it, I strongly urge you to set aside an hour sometime this week for it. In it, he contends that, even if the AI models advanced not another step, the technology represents already a “moral and cosmological cataclysm,” one that seems doomed to decimate education, literacy, and human creativity as we know it. Whatever else one might want to say in defense of AI or its extraordinary potential for good, it is difficult to deny the truth of Conti’s laments, or to dismiss him as just a troglodytic crank. Certainly, one comes away from the essay with the sense that no one should be forced, kicking and screaming, into this particular technological revolution.

In another fine recent essay, this one in The New Atlantis, Brad East argues for “The Bartleby Option”—which is to say simply, that society, through its laws, institutions, and above all social norms, should preserve space for those who say “I would prefer not to.” East lobs no grenades at the techno-enthusiast or early adopter; he asks simply that they treat with dignity their peers who dare to cling to the maxim, “if it ain’t broke, don’t fix it.” To be sure, society cannot always continue to create space for the full opt-out; if you decide to literally live “off the grid,” don’t complain of missing out. That said, we can and should at least preserve space for “I would prefer to decide when to.” We may perhaps distinguish here between the internet and the smartphone: it is one thing to say, “if you want to participate in modern life, you need to be able to get online,” another to say, “you must carry the internet on your person.” As Byung Chul Han writes, “There is no fundamental difference between being reachable at all times and being enslaved. The smartphone is a mobile labor camp in which we voluntarily intern ourselves” (Non-Things, 24). Except now it is scarcely voluntary, as I have elsewhere argued. There are contexts within which it will make sense going forward to say, “From now on you will need to use AI to do this”; there are other contexts in which it will not, and in which our capitulation to AI hegemony represents mere laziness and abject lack of imagination.

The Need for a Positive Vision

Against all of this, however, there stands a rhetorical power move that is ubiquitous these days in tech policy circles: “You don’t have a right to oppose what you don’t like about this technology until you can articulate a positive vision for where we should go instead.” This statement is so common, and so important, that I will spend the rest of this essay grappling with it.

Let it first be said that there is a strong and valid instinct here: mere gestures of negation can be lazy, unimaginative, and certainly unpersuasive. As I tell my grumpy child when he grumbles about a series of suggested Saturday afternoon outings, “Well, you’re gonna have to tell me what you want to do instead.” Humanity is meant to keep moving forward, and it is incumbent on all of us to articulate a preferred destination. That said, there is also something question-begging about this demand, for it rules out from the outset the sort of hard “No”s to technologies that humanity has occasionally been forced to articulate (in my previous essay, I mention chemical weapons and human cloning as examples). It puts us all in the position of the experts flown in to Isla Nublar at the beginning of Jurassic Park, ostensibly there to weigh in on the wisdom of the dinosaur-cloning operation proceeding on the island, but actually expected to accept it as a fait accompli. Do we have to articulate a “positive vision” for where our dinosaur-cloning efforts should go next, or are we allowed to simply say, with Professor Malcolm, that the whole thing is a bad idea?

However, I do not think that this is the right approach with AI, because AI is not just one thing. We are talking about the most general-purpose of general-purpose technologies, and although this poses its own grave moral dangers (see a piece I have forthcoming at Mere Orthodoxy on Wednesday), it does at least leave us ample room to say what we do and don’t want the technology to become, what we do and don’t want it to be used for. From this standpoint, the demand to articulate a positive vision must be answered.

If we’re trying to steer this emerging technology (a metaphor I used in the Commonplace essay), then surely we must have some idea of what we are steering toward? If we’re hoping to channel it, the way the Army Corps of Engineers channeled the Mississippi River (my favorite metaphor in this context), then surely we must have some concept of what fields we want irrigated along the way, what hydro-electric dams powered, rather than merely wringing our hands over what cities we don’t want it to flood. And if we are to build—for build we must, as human beings—how can we build without a form or a blueprint that we are building towards? So it is with artificial intelligence: it really is incumbent on social conservatives to say not merely what they want to be conserved and protected, but what horizons they would like to explore, what problems they would like to solve, what challenges they would like to overcome. It is possible to drive innovation toward conservative ends; witness for instance the remarkable developments in zero-knowledge-proof age verification technologies or AI-driven porn detectors that are at last putting parents back in charge of their children’s digital lives. We need to spend more time celebrating such success stories and imagining new ones, such as attribution-based control for AI that could help re-empower creators and communities with meaningful digital sovereignty.

This intuition represents not just the reflexive positivity of a can-do civilization, but a deeper metaphysical truth: the priority of the good. Western philosophy and ethics, especially in its Christian form, have always insisted that good is ontologically prior to evil, truth to falsehood, beauty to ugliness, virtue to vice: that the concept of bent or broken exists only in relation to that of straight or whole. Thus, in a healthy morality, it is not enough to condemn the evils of lust, betrayal, and abuse, and a man fixated on merely avoiding such sins will have a mediocre marriage at best. He must instead learn to pursue the good of marriage as a high and lofty positive ideal in its own right, a good that we will never finally achieve, but which should guide all of his strivings and which must be the basis for his rejection of vice. Does this not mean then that when it comes to forging technology, we should not get hung up on mere harm-avoidance or risk-mitigation, but should set a vision before us of the good society and build boldly toward it?

In Defense of the Given

Sed contra. Isn’t it curious, given the “priority of the good,” that nearly all the Ten Commandments come to us framed as prohibitions—the famous “thou shalt nots”? Only two of the ten (the command to remember the Sabbath and to honor parents) are framed as positive commands; the other eight all tell us first what to avoid. To be sure, Christian moral theologians have rightly insisted that it is our task to discern the implied positive command underlying all these prohibitions—“thou shalt not kill” means “thou shalt respect life.” Still, if what we should really be concerned with is articulating a positive vision for Christian marriage, why is the seventh commandment couched in such crudely negative terms: “Thou shalt not commit adultery”? Instead of just fixating on what we should avoid, shouldn’t God be propounding for us an inspiring positive image of the kinds of marriages he wants us to build? Well, spoiler alert, he does, but he still thinks we need to start from negations—and this, I think, has important implications for conservative tech policy.

The good is that at which we aim, to be sure, and yet the good that we seek can only ever be asymptotically approached by avoiding the evil, which is always nearer and clearer. Even as I climb toward a distant mountain peak, most of the time my eyes should stay anchored to the trail immediately ahead of me, if I don’t want to slip on a treacherous stone or walk off a cliff. When steering my car, yes in some sense I steer toward my final destination, but I’ll never reach it if I’m not constantly adjusting my course to avoid ditches and oncoming traffic. When channeling a raging river, I may have certain ideas about fields I would love to irrigate and hydro-electric plants I would love to power, but I must begin with constructing levees to protect the precious human lives currently in danger. Even when building, there is danger in having too fixed a blueprint in mind, one that will blind you to the contours of the site on which you have to build; a good city-builder will, like Jane Jacobs, prioritize the protection of the rich ecology of the streets and markets already in existence, rather than proceeding, like Le Corbusier, to demolish all in pursuit of an architectural ideal.

Indeed, it is not just the case that there is a kind of heightened moral urgency to protecting existing goods before seeking to bring new goods into the world, there is also a greater epistemological clarity. While it is true that we should be able to “articulate a positive vision for where we should go,” we should also expect that this vision will be quite blurry most of the time. Distant horizons come only slowly into focus, and we will often find that, as we approach them, we have to considerably revise our ideas about what we were hoping to achieve or how we might best arrive there. We can and should dream up possible new futures to pursue, but they will generally remain quite vague compared to the goods that sit right before us today, summoning us to preserve and protect them against the evils by which they are threatened today.

After all, I think there is often something of a category error in the techno-optimist’s demand that conservatives “articulate a positive vision for where we should go instead.” For while the good is ontologically prior to the evil, the already-existing good is ontologically prior to that yet to come; act precedes potency. In order to have a positive vision for technology, in other words, I do not have to dream up as-yet-unrealized states of humanity, sci-fi fantasies of man/machine mind merges or people beaming across the stars; I need only look in my child’s face. The good that technology is meant to promote is the good of humanity as it is given to us, and it is against this yardstick that we can and should judge new innovations.

What does this mean in concrete terms? It means that we must refuse to be cowed by the condescension that today rains down on conservatives who insist on first naming the wounds being visited on our common humanity before dreaming of new capacities waiting to be unlocked. I have seen many a wagging of the head and a rolling of the eyes at the concern over “chatbots” and “kids’ safety”—small potatoes, it seems, compared to the extraordinary new promises (and perils) opening up to us in a world of autonomous AI coding agents and recursive self-improvement. Sure, it’s bad that more than half of American children are increasingly outsourcing love, friendship, and rational thought to artificial companions, but we have no right to sound the alarm “until we can articulate a positive vision for where we want this technology to go instead.” Really?

Again, don’t get me wrong. I want to see bold, imaginative, forward-thinking conservatives dreaming big about how we can channel and steer these extraordinary technologies along different paths in order to make them genuine tools of human dominion. I am proud to know a few of them, and we need many more such people. But this is hard work indeed, requiring quite a unique blend of skills, and will take time to come into focus—time that is increasingly denied us. “Shall there be then in the meanwhile no doings?” we might ask, with Richard Hooker. No, for “there are the weightier matters of the law: judgment, and mercy, and fidelity. These things we ought to do; and these things, while we contend about less, we leave undone.“ In other words, the reflexive self-protection that we are seeing right now across the political spectrum in defense of the minds and souls of our children, in defense of the right for meaningful and dignified work, and in defense of the ability of communities to govern themselves—each of these are right and healthy instincts, not signs of weakness. It is only by demonstrating our willingness to fight for these known goods that we can earn the credibility, and develop the wisdom, to scale as-yet-unknown peaks and slay as-yet-unknown dragons.