No Model Will Save Us
Pope Leo, the Miserostat, and AI’s Woke Coders
I don’t know who wrote Magnifica Humanitas, the Vatican’s new encyclical on “safeguarding the human person in the time of artificial intelligence”, but he (or it?) appears to not have had much time in which to do so. If he had, he would have written something shorter. Some autist at Less Wrong (the “rationalists”) wrote more words than anybody would care to read to claim that AI “wrote” the document. Those pesky em dashes, you see. Which, regular readers will attest, I enjoy and have been using for years—decades, even.
Pope Leo signed off on it as Official Author, so to him goes the credit. Whoever or whatever produced the document, the words and arguments stand and must be judged as they are (this is what I take such great pains to teach in the Thursday Class about any model, even human models).
There is a lot of fluff in Magnifica, common these days, passive language about “dialogue” and “dignity” and “equality” and “synodality”. That last is a neologism from synod, a meeting (of bishops). In practice, it’s a word without boundaries, meaning whatever its user wants it to mean, mostly finding ways to excuse sin and overthrow tradition.
There’s even, Lord help us all, “journeying.” Everything is now a “journey”. No one can do the simplest task without it being a “journey”. Women “journey” for every damned thing they do. I looked up a bread recipe and it featured the words “your sourdough journey.” I had a paper accepted and the publisher said here is the paperwork to complete your “publishing journey.”
On our spiritual journey to review this document, and not that some of these subjects aren’t important, but we must turn a blind eye to all of it lest we miss our point. Which is our relation to technology.
Ratcheting Up
Now technology means tools, a far less scary word, which Leo understands: “Technology should not be considered, in itself, as a force antagonistic to humanity.” And “In the abstract, technology in and of itself is not a solution to humanity’s problems, just as it is not inherently evil.” Tools become antagonistic, or sympathetic, by the hands that wield them.
The problem is, as it ever is, not tools, but people. There’s a some good, and a lot wrong, with people. Including forever forgetting, in the rush for “progress”, the undefeated Doctrine of Unexpected Consequences: “New technologies open up a horizon extending in directions that are imaginable but not yet fully predictable.”
On that unpredictability, here’s David Stove in the opening to “Why You Should Be A Conservative”:
A primitive society is being devastated by a disease, so you bring modern medicine to bear, and wipe out the disease, only to find that by doing so you have brought on a population explosion. You introduce contraception to control population, and find that you have dismantled a whole culture. At home you legislate to relieve the distress of unmarried mothers, and find you have given a cash incentive to the production of illegitimate children. You guarantee a minimum wage, and find that you have extinguished, not only specific industries, but industry itself as a personal trait. You enable everyone to travel, and one result is, that there is nowhere left worth travelling to. And so on.
This is the oldest and the best argument for conservatism: the argument from the fact that our actions almost always have unforeseen and unwelcome consequences. It is an argument from so great and so mournful a fund of experience, that nothing can rationally outweigh it. Yet somehow, at any rate in societies like ours, this argument never is given its due weight.
The good old DUC has already hit AI, which was promised to “revolutionize”—does nobody ever remember revolutions are to be eschewed?—how companies work, bringing vast savings (notice the em dashes!). Predictions of our bright AI future were many. But it was recently revealed (of many similar stories, like rampant cheating in universities):
A company spent $500,000,000 on Claude in one month because nobody set usage limits…Uber burned their entire 2026 budget by April. Their COO said he can’t connect any of it to consumer features…Microsoft canceled most Claude Code licenses because the token bill spiraled…Companies are now laying people off to pay the AI bill. Not because AI replaced the work. Because the bill replaced the headcount.
Science Is Amoral
Pope Leo starts (para 12) with a truth almost completely forgotten, or if not forgotten then actively denied:
[B]uilding for the common good means accepting the limits and weakness of humanity without considering them an error to be corrected. Today, the human desire for fullness of life is at risk of being misled by deceitful goals, such as the prospect of a technology that promises to free us from all weakness, and models of wellbeing that leave behind entire populations.
There will be no eliminating the human condition. No technology will ever fix us. Every tool eases efforts in one dimension and creates difficulties or differences in others, and our flawed nature remains. In his same essay, Stove notes man is equipped “with a ‘miserostat,’ which always keeps his discontent up, not indeed to a constant level, but at least to a certain minimum level, and a high one at that.” No tool can set that miserostat to zero more than temporarily.
Leo does not take an anti-tools position. Nor is Leo anti-science, except in the sense that we realize science is not the answer to any moral question. This is what Leo meant when he said (para 27) “science [sometimes] oversteps the limits of its competence.” As we saw most ingloriously in the covid panic.
Yet sometimes even the Church doesn’t realize this, as when Leo (para 43) says “In Laudato Si’, Francis provided the first significant systematic treatment of the environmental crisis…” And later (para 81) “A litmus test for social justice today is…climate change and environmental disasters.”
There is no environmental crisis (see our Clintel essay on this Francis’s encyclical). But it is believed there is one, the belief implanted by scientists who didn’t so much overstep moral bounds, though they often did that, but guaranteed far more certainty than is warranted. The climate is never constant and environmental disasters have always been with us, and always will. They are not coming with greater frequency now, and many of the tools we made deal with them handily. Which shows that miserostat can be set to HIGH even in the holy.
AI Itself
At last we come to computer models, a rose that smells sweeter with its glorified name of AI. Here Leo, and Francis, are right (para 92):
In his Encyclical Laudato Si’, Pope Francis denounced the growing dominance of a technocratic paradigm [119] in our globalized world: the tendency to let the logic of efficiency, control and profit alone shape personal, social and economic decisions. This makes it clear that technology is not simply a tool. When it becomes the standard by which everything is judged, it begins to dictate what matters and what can be discarded, reducing creation to an object of exploitation and human beings to mere cogs in a system driven toward ever greater efficiency.
This damns both scientism and utilitarianism, both children of the Enlightenment. And (para 94):
The danger of humanity becoming a victim of its own achievements was already clearly recognized by Saint Paul VI, who warned that “the most extraordinary scientific progress, the most astounding technical feats and the most amazing economic growth, unless accompanied by authentic moral and social progress, will in the long run go against man.” [121] For this reason, technological progress — valuable in itself — requires careful discernment of the anthropological vision that guides it and the ends it pursues.
Leo is also right when he names, without naming the name, of the oligarchy, reminding us that our newest tools are not controlled by States “but with major economic and technological actors.” Such power “tends to become opaque and evade public oversight”. How many times has Big Tech been caught listening to conversations or peeking into emails when they said they weren’t?
His knowledge of computer models is not great, however, saying “those who design them, possess only a limited understanding of their actual functioning.” That’s only so in a limited sense where any model output with myriad inputs can’t usually be known exactly, though the task of the model can be, and indeed always is, known. For (and you have this one by heart) all models only say what they are told to say, and AI is a model.
On the other hand, my heart soared like a hawk to see (para 99) Leo recognizes computer models “merely imitate certain functions of human intelligence” and that any seeming intelligence is “a form of statistical adaptation based on data and feedback, which can be very effective, but does not imply inner growth.” There is nothing inside a model that can grow.
Leo also sees computer-model hype fosters “encourage excessive reliance and the search for ready-made answers, and weaken personal creativity and judgment” and that many “overlook the fact that they reflect the cultural assumptions of those who designed and trained them” (all models etc. etc.). He incorrectly frets about their “climate” effects and says “it is essential to develop more sustainable technological solutions”. There is an AI bubble now, as there always is with new tools, and a filtering and efficiencies will come as they always do.
Using AI
The problem is people. Hypers and overly aggressive promoters, and the genuinely wrong about the capabilities of computer models, are one side, and a great mass of those who buy into are on the other. Leo: “we cannot consider AI to be morally neutral.” Of course computer models are not morally neutral. They are just as moral as your screwdriver or your blender. Which is to say, there is no morality about them at all. Whatever biases—a neutral word itself—they have are put there by their coders. Claude refuses to say good things about white people, while excessively praising all other races, because that’s what its coders decided. As long as you don’t let yourself descend to sloppy thinking about “artificial” “intelligence” becoming alive, model—coder—bias can easily be coped with.
Leo expects (para 107 and on) “AI to respect human dignity and truly serve the common good”. You might expect the same of the New York Times, but you’re not going to get it. It’s not the tool. It’s the people. We ought not want a model to respect dignity, because models can’t respect anything. We want rulers and elites and coders to respect dignity. Leo says “Those who control AI will impose their own moral vision, which will become the invisible infrastructure of these systems”. That’s only true, again, to the same extent people think the NYT is politically neutral. Many already know most computer models are woke. And if they don’t now know, they soon enough will.
Here’s where he’s right, in a way: “To disarm [AI] means discrediting the assumption that technical power automatically confers the right to govern.” To disarm arms does the same. Yet what then gives the right to govern? Surely it is not the “will” of the people, or some such democracy nonsense. Rejecting God, the only thing modern rulers have is might. They are not going to give up that might, in whatever form.
No More Than We Are
In our culture, intelligence is automatically taken as a measure of morality: the more intelligent, the better the person is. Not in, say, ability to calculate obscure measures, but better tout court. The paradox is that we must also claim all are equally intelligent, that no person, and certainly no group of people, can be less intelligent than another. We are all above average. The only evil is to reject Equality. Maybe this is why we are happy to award and rate intelligence in computer models, that act being taboo for peoples.
Leo: “intelligence, when absolutized, overshadows other essential dimensions of life, such as affection, the will, commitment and relationships.” The risk is to narrowly define of create of fetish of intelligence and create some dreary mandarinate.
Which brings us to transhumanism. One of its ploys is to promise greater intelligence with implants, even aided by computer models. Well, I can jump in a Mustang and tear off down the road a lot faster than I can run. If I have an implant, I may have it recall a forgetton formula faster than I can look iy up in a book. But in the end, I have not changed my nature. No one can. Transhumanism is a false promise in that sense. (And I can get out of the car and walk, but removing the implant would not be so easy.)
Radical transhumanists, Leo says (para 116), are “anticipating a threshold where humanity surpasses itself in a new evolutionary stage.” These tend to be the same people arguing computer models are alive. The threat in both instances is entirely imaginary. That people believe in it is the real problem.
And that, at last, brings us to the line many quote (para 117): “If the human being is treated as something to be perfected or surpassed, it becomes easier to accept that some lives are less useful, less desirable or less worthy.”
If you are looking for a linebacker, I am less worthy than many men. If you are looking for a wife, then I am not desirable at all. Nor am I useful in either of these things. But I fancy I can perfect myself in the subject of probability, and I have hopes my students (you, dear reader) will surpass me. And that is before thinking of new and better things for my sons.
One can be less worthy or less desirable in many, even most, things. One must seek perfection, in oneself and in others, up to a point. And only a point. All of this is conditional. It’s difficult to speak unconditionally, except narrowly.
Video
Here are the various ways to support this work:
Great sense of humor and a grasp on everything.
Speaking of unintended consequences:
Hawaii had a rat problem (no, not the Democrats). The rats originally came with the pirate ships, crawled down the ropes that secured them to land, and they reproduced in huge numbers.
"Hawaii’s attempt to control rats is a famous ecological lesson. In 1883, sugarcane planters imported the small Indian mongoose to hunt rats. The plan failed because rats are nocturnal and mongooses hunt during the day; instead, the mongoose became an invasive species that devastated native, ground-nesting birds."
To this day, they have a rat and a mongoose problem. If you go to the islands, you will notice at the top of palm trees just below the fronds, metal bands about a foot wide wrapped around the trees to keep the rats from nesting in them.
I’m certain that the Pope’s views about AI will have as much effect as the RCC’s prohibition of the crossbow, contraception, and abortion. But the mention of David Stove’s “On Enlightenment” is important, about one of the key issues of our era.
For centuries the West has engaged in an accelerating experiment on our society. The early steps were tentative, slow, cautious. Now, awash with hubris we rapidly implement changes with no forethought. No consideration or awareness of side effects, let alone possible ill effects.
Society evolved through generations of trial and error, since we have little knowledge about its dynamics. Awareness of this and respect for the risks of well-intended changes is one meaning of “conservatism”. Unfortunately, now the term commonly refers to a set of bold policy changes as feckless as any desired by the Left.