The AI label is a deliberate bait-and-switch
Pope Leo XIV named his first encyclical Magnifica Humanitas and subtitled it around “safeguarding the human person in the time of artificial intelligence.” The AI framing was a calculated editorial decision. It worked. Tech outlets that have never treated Catholic social teaching as a news event ran the story. TechCrunch covered it. The document trended in spaces where papal documents do not normally trend.
The substance inside those 200 pages has almost nothing to do with transformer architectures or training data. Leo’s actual targets are inequality, the erosion of democracy, and the concentration of economic and political power in the hands of a small elite — concerns that Catholic social teaching has articulated since at least Leo XIII’s Rerum Novarum in 1891. The Vatican has been trying to have this conversation for over a century. Framing it around AI got the conversation into feeds where Rerum Novarum would never go.
The presentation amplified the signal. Leo unveiled the encyclical alongside Chris Olah, co-founder of Anthropic, ensuring the document landed inside the AI industry’s own credibility ecosystem. A papal text on structural injustice, handed to the press next to a prominent AI researcher, reads as a tech story first. That is not an accident.
The document states directly that when power concentrates in a small number of hands, it becomes opaque and evades public oversight. That critique applies to industrial capital in 1891, to financial systems in 2008, and to AI companies in 2025. Leo chose the 2025 vehicle because it guaranteed distribution. Secular tech media rarely treats structural injustice as breaking news. An encyclical about inequality gets a paragraph in the religion section. An encyclical about artificial intelligence gets a homepage.
The AI label is the wrapper. The contents are older, broader, and considerably more confrontational than any single technology debate.
The real targets: inequality, war, and collapsing democracy
Inequality, armed conflict, and the quiet dismantling of democratic institutions sit at the center of Magnifica Humanitas — not as backdrop, but as the actual subject. Pope Leo XIV’s 200-page document names these as chronic failures the Church has tracked across multiple papacies, long before any AI system existed to complicate them.
AI enters the argument as an accelerant, not an origin point. The encyclical’s logic is precise: power was already concentrating, democratic oversight was already weakening, economic exclusion was already structural. What advanced AI does is pour fuel on each of those conditions simultaneously. Leo writes that when power concentrates in the hands of a few, it “tends to become opaque and evade public oversight, increasing the risk of distorted forms” of control — a sentence that describes the pre-AI political economy just as accurately as it describes a world shaped by it.
That framing carries real policy weight. If AI is the villain, the obvious prescription is AI-specific regulation: model audits, safety benchmarks, algorithmic transparency requirements. Useful, but limited. If AI is instead an accelerant of pre-existing power imbalances, the prescription has to go deeper. The target becomes the underlying concentration of wealth and institutional authority that makes dangerous AI deployment possible in the first place. Leo argues explicitly that technology built and governed by a small elite cannot, by definition, serve the common good — a claim that indicts the structure of the tech industry rather than any particular product coming out of it.
This is where the encyclical becomes structurally radical. It doesn’t ask regulators to add guardrails to a fundamentally unchanged system. It asks whether the system distributing AI’s benefits and risks is legitimate at all. That question — who holds power, on whose behalf, and accountable to whom — is the one the Church has been pressing through Catholic social teaching for more than a century. Magnifica Humanitas simply updates the case study.
What most coverage is missing: the concentration-of-power argument
Most news coverage of Magnifica Humanitas treated it like a Vatican press release about chatbots — scanning for mentions of deepfakes, algorithmic bias, and job displacement, then filing the story. That framing missed the document’s sharpest edge entirely.
The encyclical’s most consequential argument is not a list of AI harms. It is a structural indictment. Leo XIV states directly that technology built and governed by a small elite cannot, by definition, serve the common good. The document’s own language is blunt: when power over transformative technology concentrates in a few hands, it “tends to become opaque and evade public oversight, increasing the risk of distorted forms” of control. That is not a warning about a hypothetical future. It is a description of the present.
The AI industry as it exists today — dominated by a handful of companies with trillion-dollar valuations, largely self-regulated, and racing to deploy systems that reshape labor markets, information environments, and military capabilities — fits that description precisely. These are actors with enormous structural power and no binding obligation to weigh broad human welfare against competitive advantage. Leo’s framework says that combination is not incidental to the harm; it is the mechanism.
This is the part of the encyclical that should alarm Silicon Valley far more than any specific regulation. A regulatory checklist is manageable — companies hire lobbyists, shape the rules, and move on. An argument that the concentration of power is itself the problem attacks the legitimacy of the entire arrangement. It is not asking AI companies to do better. It is questioning whether entities structured the way they are can do better, structurally, at all.
That argument received almost no sustained attention in mainstream coverage. Reporters noted the encyclical was 200 pages long, mentioned that Leo presented it alongside Anthropic co-founder Chris Olah, and moved quickly to the familiar AI ethics checklist. The power-concentration argument — the one with real teeth — got buried.
The historical echo: Leo XIV is channeling Leo XIII
When Robert Prevost chose the name Leo XIV, he was making a historical argument as much as a spiritual one. The last pope to carry that name, Leo XIII, published Rerum Novarum in 1891 — a document that contemporary observers largely dismissed as a Catholic institution opining on matters above its competence. History rendered a different verdict. Rerum Novarum became a foundational text of the labor rights movement, articulating the dignity of workers against the unchecked power of industrial capital at the exact moment that power was reshaping civilization.
The structural parallel to Magnifica Humanitas is not subtle. Leo XIII used industrialization as his lens to argue that economic systems must serve people, not the other way around. Leo XIV uses AI to make the identical argument for the digital age. Where Leo XIII confronted factory owners concentrating wealth and erasing worker agency, Leo XIV confronts technology built and governed by a small elite that “tends to become opaque and evade public oversight.” Different century, same anatomy of power.
That lineage reframes how the document should be read. The instinct among tech commentators is to treat Magnifica Humanitas as a religious institution playing catch-up — the Vatican finally noticing Silicon Valley, 30 years late. That reading gets the institution exactly backwards. The Catholic Church has now intervened at two civilizational inflection points — the industrial revolution and the AI revolution — with documents making structurally identical critiques of concentrated, unaccountable power. That is not catch-up. That is a 130-year track record of identifying the same fault line each time a new technology reshapes who controls what.
Rerum Novarum was ignored, then cited in labor legislation across dozens of countries. The question Magnifica Humanitas raises isn’t whether the Vatican understands AI. It’s whether anyone learned what happens when this particular critique gets dismissed the first time.
Why this matters for the AI policy debate right now
The dominant frame in AI policy — from the EU AI Act to the White House executive orders — centers on model safety, algorithmic transparency, and discrete harms like deepfakes or biased hiring tools. These are real problems. But Magnifica Humanitas implicitly indicts this entire approach as insufficient. Technical guardrails applied to systems built and owned by a handful of corporations don’t redistribute power — they legitimize its concentration while giving it a compliance stamp. Leo makes the argument plainly: when AI development is controlled by a small elite, it becomes opaque by design, evading the public oversight that democratic governance requires.
That critique lands differently depending on where you sit. Most AI governance conversations happen in Brussels, Washington, and a few Bay Area conference rooms. The Vatican’s audience is somewhere else entirely. With 1.4 billion Catholics worldwide — the majority of them in Latin America, sub-Saharan Africa, and Southeast Asia — the encyclical reaches directly into communities that will absorb AI’s economic disruption without having shaped its development. These are populations that rarely appear in AI governance debates except as passive subjects of impact assessments. Leo’s document treats them as stakeholders with a legitimate claim on how the technology is built and deployed.
For policymakers and tech executives, the encyclical poses a question that model cards and red-teaming exercises cannot answer: who decides what AI is for, and in whose interest does it run? Pope Leo XIV presented the document alongside Anthropic co-founder Chris Olah — a detail that signals the Vatican is not positioning itself outside the industry conversation, but inside it, pushing on its hardest pressure point. The industry cannot self-certify its way to legitimacy on questions of democratic accountability. The encyclical insists that answering “who benefits?” requires voices and institutions beyond the ones currently writing the standards — and it arrives with enough institutional weight to make that demand impossible to dismiss as naive idealism.
