Pope Leo Sings from the Constraint Hymnal

In May, Pope Leo XIV published an encyclical about artificial intelligence, and he did not put artificial intelligence in the title.

He called it Magnifica Humanitas. Magnificent Humanity. Two hundred and forty-five paragraphs about the most disruptive technology of our lifetimes, and the word he chose to put on the cover is not the tool. It is us. I don’t think that’s an accident. I think it’s the argument.

The shape of his argument is that AI is not our master. It’s a tool. It sounds almost too simple for a document this long, but I think it’s the hinge the whole thing turns on. A tool is not good or bad. However, he seems to suggest AI is not morally neutral because the design brings with it the version of the human who designs it. I would push back on this view. To me, AI is not a good thing. AI is not a bad thing. It is just a thing. What decides whether it serves you or runs you is not the tool. It’s whether the person using it is still oriented which brings us back to a larger agreement about AI.

The larger point is what the title is naming. Humanness. Not as a nice word, but as the thing that has to be actively preserved or it goes quietly missing, the way it went missing for a driver who stopped checking the road because the GPS was usually right, or an athlete who stopped feeling his own effort because he believed the watch knew more than he did. Pope Leo isn’t writing about AI needing to be humane. He’s writing about humans needing to stay human while something inhuman moves very fast next to them.

And here’s why I think humanness matters as much as it does. When a person loses that, what they lose first is orientation. The sense of where you actually are, what’s yours to decide, what you can still trust from inside your own judgment. I’ve spent the last few years writing about that word, outside of any religious frame at all. Pope Leo gets there through thousands of years of theological doctrine. I got there through markets and endurance sport and a fairly ordinary series of failures. But I think we’re pointing at the same thing. Preserving humanness is how you preserve orientation. That’s the whole case, underneath the Latin.

What surprised me, was how far he goes to build that case structurally rather than just morally. He doesn’t only say be good, be careful, form your conscience, though he says all of that too. He builds something closer to an architecture, and it’s the same architecture I’ve been circling for a long time, even though he never uses my words for it.

He wants us to slow down. Not always, not everywhere, but at the points where the stakes are large enough to earn it. His phrase is close to a slower pace where prudence demands it. He’s explicit that the danger isn’t the technology moving fast. It’s the technology moving faster than our capacity to absorb what it’s doing to us. That’s pacing. It’s the constraint that keeps a system from outrunning the person who’s supposed to be steering it.

He wants us to distrust the machine’s confidence. He points out, almost as an aside, that these systems are grown more than they’re built. That the people who make them don’t fully see what’s happening inside them either. And what looks like objectivity in the answer a model gives you is often just somebody else’s assumptions, delivered cleanly enough that you stop noticing they’re assumptions. That’s humility. It’s the constraint that keeps a person questioning the answer instead of just accepting it because it sounds authoritative.

And he wants the power spread out. Not concentrated in a handful of companies or a handful of individuals, deciding for everyone else what the rules of the digital world will be. He reaches for the old Catholic idea of subsidiarity, where “decisions are made at the closest level possible to the persons involved,” so that people aren’t just handed decisions made by somebody else. Then he points it directly at the small number of actors who currently hold the data and the compute. That’s friction. It’s the constraint that keeps any single builder from owning the whole wall.

Pacing. Humility. Friction. He never calls them that. He doesn’t need my vocabulary to be building with it. It’s there in the structure of the argument whether or not it’s there in the words, the same way you can recognize a load-bearing wall in a house.

I don’t think this is a coincidence, and I don’t think it’s me projecting my own framework onto a document that happens to share a few themes with it. I think it’s closer to two people arriving at the same idea from very different starting points, which is usually a sign that the idea is real rather than invented. He starts from two thousand years of the Church thinking about the human person. I started from decades of watching what happens when people or institutions build or remove the constraints that quietly hold them together. Different roads. Same architecture.

Where it stops being a coincidence and starts being something I want to sit with a while longer is where we may not agree. Pope Leo’s version of these constraints tends to resolve into the heart and the law. Examine your conscience. Form the young well. And where something with teeth is required, let the state or a regulatory body supply it. He’s honest enough to admit, in his own text, that regulation by itself won’t be sufficient. I believe him. What I keep circling is something less dependent on virtue and closer to physics. A design that holds because of the structure. Virtue is desirable, but it is also less reliable.

I don’t have that fully worked out. I’m not sure Pope Leo does either, not all the way down to the specifics. His own conclusion doesn’t end in a policy. It ends in a hymn.

But I keep coming back to the title, because I think it already told me everything else in the document was going to be in service of. He didn’t call it an encyclical on artificial intelligence. He called it the grandeur of humanity, and then spent two hundred and forty-five paragraphs explaining what has to be protected for that grandeur to survive contact with a very fast, very confident tool. Preserve the human. The orientation follows. Accomplished through a constraint architecture. I think that’s the whole argument.

Magnifica humanitas. It’s a beautiful Latin phrase for the same thing I’m trying to say in my own version. He ends two hundred and forty-five paragraphs in a hymn. I think that is the right instinct. Some things can be sung rather than argued.

Thank you for reading Wondering Aloud. Please share it with others if you find it worthwhile.

Steve Barth, CFA

Wondering Macro LLC