If you had told me at the turn of this century that the two most hotly anticipated cultural events of summer 2026 would be a screen adaptation of Homer and a papal encyclical, I’d have blessed your heart and called you crazy. Yet here we are, and the subjects everyone told me were “useless” in school—philosophy, literature, dead languages—are informing several ferocious disputes over the fate of our civilization. The future looks bright for the humanities! 

The controversial screen adaptation is Christopher Nolan’s Odyssey, which will appear in July and either confirm or refute its critics’ worst fears of meddlesome politicking on the part of its director. The papal encyclical, of course, is Pope Leo XIV’s Magnifica Humanitas, on the challenges and opportunities of artificial intelligence. As it happens, that’s also the subject the editors at Law & Liberty asked me to address in conversation with several other authors, each of whom has made an illuminating contribution to the discussion. I want to thank them for tracing such intriguing lines of argument, and to round off our exchange with a few thoughts inspired by reading their essays alongside the first American pontiff’s first encyclical.

Pope Leo invokes the Tower of Babel described in Genesis 11 as a cautionary image of human enterprise ungoverned by transcendent purpose—sheer technological power with no directing principle but that of dominion. He brings up a point about this story that has often struck me as significant, too: “When a city is built on pride and the claim to self-sufficiency, communication breaks down, languages are confused and people no longer understand each other.” 

Of course, we all know that the punishment—or consequence—of building Babel is the confusion of tongues. The most obvious way of picturing this crisis is to imagine the builders getting hit suddenly with a miraculous bamboozlement, forgetting how to speak their native Akkadian and variously speaking Phoenician, Sumerian, Hittite, etc. But there is another, less sensational interpretation, which is to picture the builders continuing to use the same words but gradually coming to mean different things by them. 

Young people are immersed in a torrent of words written by nobody about nothing and pictures painted by no human hand.

Participants in American politics over the last several decades know that this state of affairs can arise quite easily under conditions of extreme social dysfunction. “What is a woman?” What is “our democracy”? Who is a socialist, a fascist, a Communist, or a Marxist? The answer depends very much on who’s saying the words, even if we’re all technically speaking English. Thucydides wrote that during a particularly dreadful civil war in the island city of Corcyra, “Words had to change their ordinary meaning” as formerly agreed-upon principles of law and justice disintegrated.

In his meditation On Common Eloquence, Dante Alighieri imagined the builders of Babel undergoing a similar process, forming specialized task forces and immersing themselves in different parts of the construction effort. Some groups focused on troweling mortar, others on baking bricks, others on laying foundations, as each group developed its own set of concerns and ways of expressing them—until, “As many varieties of work were involved in the project, that was the number of languages into which the human race was now divided.” In other words the confusion of tongues was not a transformation of the words people used, at least initially, but a fragmentation of the mutual assumptions behind the words
that once made them make sense. Over time, different visions of the world would have prompted the invention of differing argots, new dialects, and finally a proliferation of altogether mutually unintelligible languages. The first essential change, though, would have been a loss not of shared syllables, but of common purpose. 

So, when he dreamed up his vision of hell years later, Dante made his Inferno into the ultimate Tower of Babel and filled it from the threshold to the depths with incomprehensible gibberish—uttered, as Virgil tells the Pilgrim early on, by those “who have lost the good of the intellect.” Those who share no common good may speak the same words, but can never understand each other.

Is this, perhaps, what makes our debates about AI so fraught? They are filled with words that mean different things to different people: Intelligence. Artificial. Human. Mind. The controversy about Nolan’s Odyssey involves translation, too: viewers and readers are fighting over how to port concepts like “hero” and “beautiful” from ancient Greek into English, and from the language of poetry to that of cinema. But the controversy over artificial intelligence poses linguistic problems of an even deeper sort. 

For example, Kevin Frazier, our most enthusiastic AI advocate, accuses me of “a failure to appreciate the exponential character of AI improvement.” But my essay was about the categorical differences between real human intelligence and all its mechanized approximations. This leaves me feeling that Frazier and I are talking past each other. You can exponentially increase your velocity in the direction of the moon, but it won’t get you a hair closer to making gold out of lead. The “improvement” involved is simply not that sort of improvement. 

My argument was that certain things may remain beyond the proper remit of computer programs by their natures. Those things, I suggested, include real artistic creativity and responsible moral decision-making. In my view, this judgment applies across the board to Large Language Models, Vision-Language Models, Segment Anything Models, Latent Consistency Models, and any of the other “entirely different types of AI” that Frazier alludes to. I don’t consider all AI researchers to be “ultra-elite, out-of-touch, West Coast snobs.” But it does seem to me that when I raise the issue of categorical differences between man and machine, AI maximalists often respond that I can’t possibly understand the coming revolution until I spend half the cost of my mortgage on an upgrade to the pro model. In other words, I can’t imagine how much better the machines will get any day now at doing what they are already doing, but more. This strikes me as an inability or a refusal to address the actual questions being posed. 

Those questions include which sorts of things machines can really do well, which sorts of things they can only simulate, and most importantly, which sorts of things are inherently such that only we can do or should do. When Jeff Bilbro worries that “incorporating AI technologies into one’s workflow and personal life tends to gradually turn agents into victims,” I think he is identifying one major symptom of the confusion between human and digital capacity. I, too, have noticed, and research seems to confirm, that people who let AI write their essays or ad copy degrade their own ability to think through what they mean and find words to say it. How could it be otherwise? Letting a machine write your sentences for you is not the same as letting a machine make your toast for you. You’re not freeing up time for the real business of life by outsourcing menial labor. You’re relegating yourself to menial labor by outsourcing the real business of life.

This is why I appreciate Brent Orrell and Kathleen O’Toole, whose essays I read as a pair, reminding us that one of the things machines can’t do effectively—and probably shouldn’t do on their own at all—is train human beings. “The more a task requires judgment, creativity, or human connection, the more difficult it will be for an AI-powered android to take it over,” writes O’Toole. “Yet there are few activities that require more of all three than shaping a human mind.” Her emphasis is on education; Orrell’s is on the institutions and traditions of apprenticeship that equip people to exercise good professional judgment after their years at school: “serious career and technical education that puts learners in real work environments under real stakes. … Rigorous sector-based training. … Apprenticeships, residencies, and the older professional formation traditions.” 

In my own classroom, I have found that the looming presence of AI makes the old mental disciplines of the liberal arts vastly more necessary, not less. It makes their value more apparent to students as well. It’s actually quite easy to show a phone-sick generation that deep reading and handwritten essays will hone their minds into something no machine can replace. If you believe it yourself, that is. Maybe this is why, in a tech-riddled era, papal encyclicals and epic poetry have suddenly come to seem so burningly important. Maybe this is also why commencement speakers at college graduation ceremonies this year were surprised to find their praise of the AI economy met with boos and jeers.

Young people are immersed in a torrent of words written by nobody about nothing and pictures painted by no human hand. They seem to be getting a bit fed up with piling brick after brick up to heaven. What they need to make the babel of confusion into a world of meaning is not more stuff but more sense. What we all need to steady us going forward is “the good of the intellect”—the unity of vision that only living insight, and human wisdom traditions, can deliver. None of this means we shouldn’t build state-of-the-art data centers or equip people to use the latest tools. It just means that humanists like me are going to keep insisting—annoyingly, I realize—that the tools aren’t the people, and the people are the point.

Source link lawliberty.org