Editor’s note: This article first appeared at The American Spectator.
There’s an apocryphal story of automation in the 1500s that runs as follows. The Reverend William Lee, an English inventor, studies the movements involved in knitting and builds a machine to imitate them. It works; with his “stocking frame,” laborers can sew socks with a fraction of the time and skill required before. Excited, Lee brings his device to Queen Elizabeth and asks for the Crown’s support. But Elizabeth denies him: “Consider thou what the invention could do to my poor subjects,” she says. “It would assuredly bring them to ruin by denying them employment.”
Or perhaps she doesn’t say that at all; records of the encounter are hard to come by, and their veracity is questionable. But the sentiment — that technological progress will destroy human society — is evergreen. We hear it in the nineteenth-century Medical Record, which warned that the telegraph would lead to an “increase in nervous disease … made by the conditions of modern life upon the brain.” An earlier version appears in Plato’s Phaedrus — learning to read, he says, will:
create forgetfulness in the learners’ souls, because they will not use their memories … you give your disciples not truth, but only the semblance of truth; they will be hearers of many things and will have learned nothing.
The webcomic Dresden Codak imagines the objection a step farther: an innovative caveman demonstrates a flaming torch or sharp rock, while his fellows disapprovingly grunt, “You am play gods!”
Today, generative artificial intelligence — genAI — is the most popular guess for the doom of civilization. Predictions for its future use range from the merely disastrous (the New York Times: “Silicon Valley is bracing for a permanent underclass”) to the outright apocalyptic (Eliezer Yudkowsky’s book If Anyone Builds It, Everyone Dies). Even Dario Amodei, CEO of Claude-creator Anthropic, predicted in an Axios interview that we might face a future where “cancer is cured … and 20% of people don’t have jobs.” (Of course, for a tech giant that’s as much a marketing pitch as it is a warning; as computational pioneer John von Neumann said, “Some people confess guilt to take credit for the sin.”)
Available now from Faith & Press: Born of Ideas. Click here to order.
Claims of disaster may be ever-present; that does not make them all wrong. Technology disrupts — by its very nature, it breaks things. It’s probably true that the 1590s were not a great time to be a veteran stocking maker, nor the early 1900s to be a horse and buggy supplier. The practitioners of these fields could justly complain that no one asked their permission before making their professions obsolete. And Plato was right, too: it’s harder to make the case for oral memory when everyone can read.
So suppose we accept all the modern critiques of technology. Many of them are clearly true: yes, obesity is rampant, thanks in part to the move from hands-on labor to sedentary office jobs. Our children are screen-obsessed, and our teenagers struggle with neuroses born of social media. Technological advances contribute to illnesses both physical and psychological, political fragmentation, and personal disconnection. And AI seems poised to unleash new problems — deepfake pornography, say, or threats to arts and literature. Isn’t that the end of civilization?
Well — compared to what?
The tricky part in declaring civilizational decline is to identify the point we’re declining from — and then to consider all the changes, good and bad. Tech proponents famously neglect the latter; “learn to code” was a catchy slogan in the 2010s, but it was a callous response to unemployment and genuine suffering. But it’s easy to ignore the positives as well. Queen Elizabeth (or her fictional stand-in) might have been right about the stocking frame’s immediate impact on textile workers, but their successors ultimately benefited as well. (How many people in hosiery today would voluntarily return to sixteenth-century wages?) For those of us on the purchasing end, the tradeoffs are simpler; automation is the reason that “buying nice socks” is a Walmart afterthought and not a once-a-year splurge.
The other critiques balance out similarly. Written cultures may not encourage memorization like oral ones do, but they’re vastly superior at communicating and preserving information. Our calorie-rich and exercise-poor lifestyles are terrible for us; are they worse than the rampant starvation they replaced? We all felt how “thin” digital connections could be during the pandemic lockdowns — but would we have been less isolated without Zoom and FaceTime?
There’s a pattern here in which technology replaces mandatory ills with voluntary ones. If you have no food, you will starve; if you have too much, you can overeat. Dying of typhus or smallpox wasn’t a choice, while compulsively scrolling a media timeline is. For that matter, nothing prevents you from memorizing The Iliad or knitting a pair of stockings — technology doesn’t remove these options from us. Rather, it offers us new, more enticing ones, and we too often choose the treat over the work.
And so back to AI. My students struggle with digital temptation on a daily basis; more honestly, I struggle alongside them. It’s hard to motivate mastering a skill — writing, painting, programming — when the computer can effortlessly outperform a beginner. We wrestle with formative tasks, while a button marked “Do this for me” is persistently just a few clicks away.
That’s the real danger, I think, to our picture of civilization. The threat is not that AI will eliminate all our jobs and leave us to sit with folded hands, unable to find any fruitful labor. Automation has been a growing presence for at least 400 years, and for all the work we’ve assigned to the machine, we have yet to run short of things that need to be done. Instead, we’ve discovered new things to do, and it’s become affordable to mass-produce what was once the provenance of kings.
We can see that progression in my own discipline of computer science. Since its inception, the trend has been for computer programming to become faster, easier, and more accessible: from hand-coded binary, to punch cards, to modern English-like programming languages. With each transition, programmers have written less code to accomplish more. If automation means the demise of work, the number of practitioners in the field should have dwindled with every step. Instead, it’s grown: easier software means more calls for software, which means more jobs. If generative AI reverses this trend, it’ll be the first time.
The danger, then, isn’t that there will be no work; the danger is that we won’t want it. Call it unhealthy dopamine reward structures, call it corrupted human nature, but the problem is the same — if our culture slides into decline, it won’t be because AI forced that on us, but because we considered the options and decided on sloth. We choose, too often, not to do the hard things.
But we could do them.
One of the great gifts of technological progress has been time. A peasant might have the genius of Einstein or the talent of Shakespeare, but if the wheat fields demand attention from sunup to sundown, it doesn’t much matter; there’s no room left to utilize those skills. Overwhelmingly, the discoveries and cultural achievements that come down to us are the product of a privileged few — the people who could spend moments on more than mere survival.
With automation, we put boundaries on work — these days for those hours, and the rest of our time freed for rest or education or simple thought. We democratized creativity, or at least the opportunity for it. The result was a century of unprecedented innovation, creating new forms of media, scaling heights that were literally astronomical. We have more art today than any other time in history; if there’s more dross than ever before, there’s more gold as well.
AI could extend these possibilities — this leveling of opportunity — once again. In its simplest mode, generative AI serves as an expert on tap. It can be a tutor for the student in poverty, a translator for the harried employee, an idea board for the entrepreneur struggling to brainstorm. In many cases, it can do so for free.
But these are the surface; the options now are well beyond “clever chatbot.” With agentic AI, we have an increasing ability to set the computer to real work: make this video, organize these files, build me the shell of that application. That’s a shorter path between “idea” and “existence” than we’ve ever known. We will discover new talents this way — new thinkers, new directors, new innovators in artistic expression — that would have otherwise gone unnoticed for simple lack of opportunity.
And beyond that, AI opens opportunities for better life. The Christian theologian Abraham Kuyper described a category of “common grace” — gifts like fire or honest government, that provide for human welfare or constrain destructive forces. Surely genAI fits this category. In the last five years, we’ve seen it used to detect early-onset cancer, fold proteins for custom-tailored medication, and read scrolls so fragile they can never be unrolled. AI-driven cars reportedly see 90 percent fewer accidents. AI-assisted tax returns discover thousands of dollars in missed benefits. These are the first fruits; there’s potential here to promote human flourishing that would be as shocking now as our relative longevity might have been to Elizabethan England.
The catch, once again, is that every common grace has its abuse: fire burns houses as easily as kindling, governments established in justice succumb to tyranny. AI could mean the end of conveying real skill in education. It could lead to generations lost in indolence and virtual pleasures. It could be the last real innovation; human progress could end, not under the skeletal heel of a Terminator, but in the soporific comfort of a gaming chair.
To avoid that path will require work, from those of us laboring now as much as from the generations that come after. It’ll mean rebuilding educational strategies to avoid (or incorporate) computerized shortcuts. It will demand we rethink hiring approaches: how do we train up new human experts, knowing that the computer outperforms our novices for their first few years? In many cases, it will necessitate discovering reflective human work that sits above the AI, answering not merely “can you make these things?” but “how ought they to be made?”
Creating a culture that pursues these ends will be difficult, but we can choose to encourage it — in our families, our workplaces, our communities. If we don’t, if our civilization slides into indolence and ruin, it won’t be something the machine did to us; it’ll be our choice to abandon the hard and good things in favor of the quick hit. The fault, dear friends, is not in our chatbots, but in ourselves. And perhaps it’s there — in forcing us to confront that truth — that AI offers both the greatest terror and the most important gift.