The Future Belongs to Those Who Resist It — The New Atlantis

Online Exclusive

May 29, 2026

The Future Belongs to Those Who Resist It

Magnifica Humanitas is an inspiring invitation. But its focus on war, unemployment, and oligarchy misses the more insidious threat: that AI will turn the human experience itself into slop.

Clare Coffey

Subscriber Only

Sign in<br>or Subscribe Now for audio version

The main text of Magnifica Humanitas, Pope Leo XIV’s encyclical on humanity in the age of artificial intelligence, opens with two biblical images of building. The first is ambitious, gleaming, rationalizing, and maximizing. It is the Tower of Babel. The other is humbled, limping, uncertain, and diffuse. It is a scene from the book of Nehemiah, where the prophet assigns each family of Jerusalem a section of the wall to rebuild. The same choice is before us, Pope Leo believes: to build Babel or Jerusalem, a “culture of power” or a “civilization of love.”

There is, of course, no one that Silicon Valley loves more than a “builder” and nothing, ever since the word first escaped containment in its cramped wet market of ideas, that it loves more than the builder’s agency. This has not prevented techno-optimists from adopting a posture of anxious passivity with regard to artificial intelligence. AI is a train, and you can get out of its way or let it mow you down. AI is a ship, and you can scuttle up the gangplank before it sails or be abandoned to the gnashing and wailing of the permanent underclass. You need to incorporate agents into your workflows, not just in cases where they demonstrably solve some problem for you, but so you can keep up, so you can engage, so you can compete, so you can speak the new language of the new world, so you can scramble for some advantage on the margins: so you don’t get left behind.

Pope Leo ranges himself against this kind of disciplinary pseudorealism in political questions of all types. But when it comes to AI specifically, the encyclical whiffs an obvious opportunity.

In a section on education, Pope Leo nods to a striking feature of LLMs: that the major ethical struggle they provoke does not revolve around how they are designed (with all due respect to every company currently posting open positions for a neuroscientist who reads Marcus Aurelius) or what we will do with the powers they bestow, but whether society will be able to successfully identify, stake out, and defend the territory of human life where their use is inappropriate:

This is a fundamental issue because every technology shapes those who use it. Educating people about the use of AI, then, involves teaching them to decide when and for what purpose it ought not to be used.

How strange, then, and how frustrating, that a treatment of AI chatbots can muster only the most timid and qualified of corrections:

The artificial imitation of positive human communication — words of advice, empathy, friendship and even love — can be engaging and at times genuinely helpful. However, for less discerning users, it can also be misleading, creating the illusion of a relationship with a real personal subject. When words are simulated, they do not build genuine relationships, but only their appearance. The artificial imitation of care or support can become particularly risky when it enters contexts where real relationships and emotional bonds are lacking. Here, the danger is not so much that a person may believe they are communicating with another person, but rather that they may gradually lose the very desire to form genuine human connections.

Artificial friends aren’t verboten, exactly — in fact, they can be “at times genuinely helpful.” The problem is not fundamental, not ineradicably present in the choice to socialize with a machine, but occurs only in the “danger” of a “risky” activity — especially for “less discerning users.”

Our natural affections require proper objects, and even our lowest forms of animal sociality are scaffolding for an encounter with another immortal human soul. To seek the pleasures and comforts that attend human connection — or even the unequal but genuine companionship of a dog — in the manufactured simulacra of a human mind is grotesquely disordered. It is grotesquely disordered no matter how sophisticated an understanding of neural networks accompanies it.

The more lonely and vulnerable someone is, the less “discerning” they are, the more they will benefit from clear, concrete, unambiguous guardrails protecting the domains where AI agents must not be welcomed. But this is exactly what Magnifica Humanitas had the opportunity, and failed, to provide.

Pope Leo is not in any way indifferent to the protection of these domains. He insists that a statistical model cannot assume moral responsibility and therefore cannot judge in a true sense; it must never authorize lethal actions in war. But he is most interested in what might be called macro concerns: war, displaced workers, the opacity of algorithms,...