Why the Tech Press Misread the Vatican’s AI Document

When Pope Leo XIV released Magnifica Humanitas on May 25, 2026, the tech press did what it usually does with anything from Rome: skimmed the headline, noted the word “danger,” and filed it under “religious institutions anxious about modernity.” That framing is lazy, and it caused most outlets to miss the document’s most interesting intellectual contribution — a structural critique of AI development that holds up under scrutiny far better than the coverage suggested.

The 42,000-word encyclical is not a polemic against technology. Leo XIV is explicit about that. Its argument is more specific and more interesting: that the current trajectory of AI development mirrors the Industrial Revolution in ways that created enormous social damage before corrective mechanisms finally caught up — and that waiting for those corrections again would be a predictable and avoidable failure.

That argument deserves a serious read, not a dismissal. Here is what the document actually says, where it holds up, and where it strains.

The Industrial Revolution Analogy Is Historically Accurate

The encyclical’s timing was deliberate. Leo XIV signed it on May 15, 2026 — exactly 135 years after Pope Leo XIII issued Rerum Novarum in 1891, the landmark document on capital and labor that helped provide a moral foundation for European social democracy and, indirectly, the labor reforms of the New Deal era. The parallel Leo XIV is drawing is not decorative. It is the entire argument.

Rerum Novarum was written into a world where industrial capitalism was mechanizing physical labor at scale, concentrating capital in the hands of factory owners, displacing skilled craftsmen, and creating urban poverty on a scope institutions had no vocabulary for yet. Leo XIII provided that vocabulary. It took decades for policy to catch up, but the conceptual framework he articulated — that labor has dignity, that capital concentration is a public problem, that markets require moral constraints — eventually shaped the legislative architecture that governs work to this day.

Magnifica Humanitas makes the analogous claim for AI: just as the Industrial Revolution mechanized muscle at unprecedented scale, AI is beginning to mechanize cognition — judgment, attention, creativity, analysis, and the processes through which meaning gets made. The claim is accurate. A 7B model in 2026 does what required a 70B model in 2023. GPT-5.5 scored 83% on the GDPVal benchmark, placing it at or above human-expert performance on economically valuable tasks. Cognition is being automated in ways that are directly analogous to what steam power did to physical work.

Where journalists wrote “the Pope compares AI to the Industrial Revolution,” they treated it as a metaphor. It is not a metaphor. It is a historical-structural diagnosis — and the diagnosis is correct.

“Technology Is Never Neutral” Is a Systems Argument, Not Mysticism

The most quoted line from the encyclical is: “Technology is never neutral, because it takes on the characteristics of those who devise, finance, regulate, and use it.” Tech Twitter found this quaint. It should not have.

This is not a theological claim. It is a straightforward observation about how technical systems embed the interests and assumptions of the organizations that build them. This is what algorithmic bias researchers have documented empirically for a decade. It is what the DORA 2025 report showed when it found that AI coding tools improved throughput for teams with good DevOps maturity and degraded delivery performance for teams without it — the tools amplified existing organizational dynamics rather than correcting them. The technology was not neutral; it took on the characteristics of the organizations using it.

The encyclical goes further than algorithmic bias, however. It argues that the particular problem with AI is power concentration: the same handful of firms — OpenAI, Google, Anthropic, Meta, Microsoft — are simultaneously building the technology, setting its norms, funding its governance bodies, and deploying it into every domain of economic and social life. This is not a religious objection. It is a structural one. And it is accurate. The Model Context Protocol, now donated to the Linux Foundation, is co-governed by Anthropic, OpenAI, Google, Microsoft, and AWS. The firms writing the governance rules are the same firms whose products the rules govern.

Rerum Novarum raised the same structural alarm about industrial capital. It took the Sherman Antitrust Act, the Wagner Act, and the New Deal to institutionalize a correction. The encyclical is asking whether we want to wait that long again.

The Labor Analysis Maps to What We Actually See

The document’s treatment of labor is where it most directly intersects with data that vortx.ch readers will recognize. Leo XIV warns that AI will displace cognitive workers without a corresponding redistribution of the productivity gains, producing what he frames as a degradation of human dignity through economic marginalization.

That is a strong claim, but the trend line supports it. Snap cut 1,000 jobs while reporting that AI writes 65% of its code. Meta cut 8,000 jobs while committing $145 billion to AI infrastructure. In both cases, productivity gains from AI automation were captured by shareholders and reinvested in AI, not passed to displaced workers. This is not a pathological outcome — it is the default behavior of rational capital allocation. It is exactly what happened in the Industrial Revolution, and it is exactly what Leo XIII’s document was written to address.

The encyclical is not arguing that AI should not automate cognitive work. It is arguing that the gains require governance mechanisms — that left to default market dynamics, the distribution will be as brutal as the textile mills of 1840s Manchester, and the institution with a 135-year track record of caring about that outcome wants to say so before the harm is fully baked in.

Where the Encyclical Overstates Its Case

Being intellectually honest about this document requires acknowledging where it is weaker. The sections on AI in warfare and lethal autonomous weapons are well-argued — the moral hazard of reducing the psychological cost of initiating conflict is real, and the argument that reduced human control makes it easier to begin wars is empirically grounded in military literature. That section deserves more attention than it received.

The sections on “artificial empathy” — the risk that companion robots and AI therapists will produce further social isolation rather than genuine care — are less rigorous. The claim that authentic human contact requires vulnerability “that a machine cannot possess” is a philosophical assertion, not an argument. It may be correct, but the document does not make a case for it beyond assertion. An enterprise audience will correctly note that the evidence on AI-assisted therapy is actually mixed, with some studies showing reduced depression scores in populations where human therapists are inaccessible.

The “Babel vs. Jerusalem” framing — the choice between prideful, remote power and human-centered community — is theologically resonant but analytically vague. The document is strongest when it is specific (power concentration, labor displacement, AI in warfare) and weakest when it retreats into metaphor.

What the Encyclical Actually Asks For

Leo XIV calls for three things in concrete terms: global AI governance that prioritizes the common good over shareholder returns; meaningful worker protections as cognitive labor is automated; and independent oversight of AI systems deployed in high-stakes domains — healthcare, justice, warfare, education. None of these are radical demands. All three map to policy discussions already underway at the EU AI Act level, the UN AI Safety Summit level, and in the U.S. state legislatures that have passed or are passing AI governance bills in 2026.

The encyclical’s contribution is not to advance those specific proposals — it is a moral document, not a policy brief. Its contribution is to give those demands an institutional and historical legitimacy that a Stanford ethics paper or a Brussels consultation process cannot provide. The Catholic Church has 1.4 billion members, a multigenerational institutional memory, and a track record of actually influencing economic governance at scale. That is not nothing.

The tech press covered Magnifica Humanitas as a story about religion and technology. It is actually a story about whether an institution with a demonstrated capacity to shape labor policy at civilizational scale is going to engage with the AI transition seriously. It is. The analogy it is using is accurate. The structural argument it is making is sound. The governance demands it is articulating are reasonable. Dismissing it as religious anxiety is the lazier take — and probably the costlier one.

Further Reading

• Magnifica Humanitas — Full Text (Vatican) — The source document; chapter 3 on labor and chapter 4 on AI governance are the most analytically rigorous sections.
• Pope Leo XIV Compares AI to the Industrial Revolution (TechXplore) — The Conversation’s analysis of the structural parallel with Rerum Novarum, written by researchers who engaged with the document rather than the headline.
• AI Safety Tests Are Failing: The 2026 Report Explains Why — The empirical backdrop for why the encyclical’s governance demands are timely.