TL;DR
Pope Leo XIV signed Magnifica humanitas on May 15, 2026, and released it May 25, framing AI as a defining test for Catholic social teaching. The source material says Anthropic co-founder Chris Olah was among the AI experts present at the Vatican presentation, while OpenAI, Google DeepMind and xAI were absent.
Pope Leo XIV released his first encyclical, Magnifica humanitas, on May 25, 2026, setting out a Vatican critique of artificial intelligence power, labor disruption and AI-enabled warfare at a launch where Anthropic co-founder Chris Olah was present and major frontier labs including OpenAI, Google DeepMind and xAI were absent, according to the source material.
The document was signed May 15, 2026, the 135th anniversary of Pope Leo XIII’s 1891 encyclical Rerum novarum, and released ten days later. Thorsten Meyer AI’s account says Leo XIV presented the document personally at the Vatican, a choice it described as unusual because popes often delegate such presentations.
In the account, the encyclical is organized into five chapters that place AI inside Catholic social teaching. It warns that AI power may land “in the hands of only a few,” and argues that moral claims about AI are inadequate “if that morality is determined by a few.”
The source material states that the document addresses human dignity, concentrated technical power, workplace monitoring, employment reshaped around machines and AI’s role in warfare. Thorsten Meyer AI’s interpretation is that the guest list became part of the message: Anthropic’s presence was fitting, but the absence of OpenAI, Google DeepMind and xAI made the room narrower than the encyclical’s own logic seemed to call for.
Technology is never neutral — and neither were the empty chairs
Pope Leo XIV’s first encyclical casts AI as this century’s Rerum novarum moment. He presented it personally — with Anthropic’s co-founder in the room. OpenAI, Google DeepMind & xAI were not. For a “broadside against AI companies,” that guest list is itself an argument.
A Rerum novarum for the age of AI
The signing date wasn’t incidental. Leo XIV chose the 135th anniversary of Leo XIII’s 1891 encyclical — and, by taking the Leonine name, cast himself as the pope who answers AI as Leo XIII answered industry.
The same move, 135 years apart
AI ethics books for developers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
AI Ethics (The MIT Press Essential Knowledge series)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Five chapters, one worry: concentration
The recurring anxiety is that AI’s power lands “in the hands of only a few” — and that a more moral AI isn’t enough “if that morality is determined by a few.”
A dynamic doctrine, faithful to the Gospel
Situating AI in the Church’s social teaching — the living tradition from Rerum novarum onward.
Foundations & principles
Human dignity that is “neither acquired nor earned”; the common good; the universal destination of goods — tech must not be held by a few.
Technology & dominance
The “technocratic paradigm.” AI can simulate a person but has no moral conscience or empathy. Calls to “disarm” AI from the logic of competition.
Safeguarding humanity: truth, work, freedom
The “new ways” of working aren’t always better; AI too often makes workers adapt to machines. Warns of an “architecture of visibility.”
The culture of power & the civilization of love
The hardest charge: “no algorithm can make war morally acceptable.” Argues even “just war” theory must now be overcome.
AI regulation and policy guides
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
AI monitoring software for workplaces
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Who was in the room — and who should have been
Leo XIV presented the encyclical personally (popes usually delegate). Among the AI experts: Anthropic’s Chris Olah. The other frontier labs? Empty chairs. Tap each seat.
The presentation · May 25, 2026
A defensible single invite — or a diluted broadside? Press play, then judge.
AI transparency tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
AI warfare simulation tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
A broadside delivered to one delegate
The Washington Post read the encyclical as one that “fires a broadside against AI companies.” A reckoning aimed at an industry is weakened when one member — the most safety-branded one — is present to receive it.
The encyclical’s hardest charge is about AI and war — and it implicates the labs that weren’t there.
Its most uncompromising passages condemn AI-enabled weapons and the lowering of the threshold for violence. But that lands hardest on the defense-entangled players and the leaders most explicit about military & geopolitical ambitions — not the lab that showed up.
Account vs. anoint
One sympathetic guest tilts it from “the Church holding the industry to account” toward “the Church beside its preferred firm.”
Concentration, again
A text whose deepest fear is power “determined by a few” launched by elevating one company as chosen interlocutor.
ethical AI development kits
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
AI moral decision-making models
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Two things are true at once
The criticism is of the exclusivity, not the inclusion. Olah in the room was fitting; Anthropic alone was incomplete.
The most significant AI reckoning yet by a global moral institution
It grounds a critique of concentration, dehumanized work & algorithmic warfare in a tradition stretching back to 1891. Its core insight — technology carries its makers’ values — is exactly the right place to start.
A broadside should be delivered to the industry, not its most palatable face
The choice to present alongside Anthropic alone — defensible, probably well-intentioned — undercut the encyclical’s own insight about whose values get associated with the message.
A beginning, not an endpoint
The same month, Leo XIV approved an Interdicasterial Commission on Artificial Intelligence — a standing body with room for many voices over time. If it brings the whole industry into uncomfortable dialogue, the narrow first launch reads as a first step, not a pattern.
Why It Matters
Why It Matters
The encyclical places AI policy in the same moral lineage as the Church’s response to industrial capitalism in 1891. By linking AI to labor, war, monitoring and concentrated ownership, Leo XIV treats the technology as a social order issue, not only a product safety issue.
The guest list matters because the document’s central claim is that technology reflects the people and institutions that design, finance, regulate and use it. If only one major lab is represented at the launch, critics may read the event as narrowing the field of accountability at the same moment the text warns against decisions made by too few actors.
Background
Context
The source frames Magnifica humanitas as a deliberate echo of Rerum novarum, Leo XIII’s response to the Industrial Revolution. The parallel is tied to both the papal name and the signing date, 135 years after the 1891 document.
The document’s critique is reported to extend beyond commercial AI products. Its strongest language is aimed at AI-enabled weapons and at systems that could lower the threshold for violence. Thorsten Meyer AI argues that this part of the message lands most directly on companies and leaders with military or geopolitical ambitions, including some who were not in the room.
“fires a broadside against AI companies”
— The Washington Post, as cited by Thorsten Meyer AI
What Remains Unclear
What Is Still Unclear
The source material does not establish whether OpenAI, Google DeepMind or xAI were invited and declined, were not invited, or were represented through other channels. It also does not provide a full Vatican attendee list, the official criteria for choosing AI experts, or responses from the absent companies. Those points remain unresolved.
What’s Next
What Happens Next
The next test is whether the Vatican broadens its engagement with AI developers, regulators, labor groups and defense-policy voices after the encyclical’s release. Any response from OpenAI, Google DeepMind, xAI, Anthropic or Vatican officials would help clarify whether the launch was a one-off presentation or the start of a wider public discussion about AI power.
Key Questions
What is Magnifica humanitas?
Magnifica humanitas is Pope Leo XIV’s first encyclical, signed May 15, 2026, and released May 25, 2026. According to the source material, it applies Catholic social teaching to artificial intelligence, with emphasis on dignity, labor, concentrated power and war.
Why did Anthropic’s presence matter?
The source material says Anthropic co-founder Chris Olah was among the AI experts present at the Vatican presentation. That matters because the encyclical argues technology reflects its makers and financiers, making the choice of represented companies part of the public meaning of the launch.
Were OpenAI, Google DeepMind and xAI excluded?
The source material says they were not present, but it does not confirm whether they were excluded, not invited, unavailable or represented in another way. Their absence should be reported as confirmed by the source, while the reason remains unknown.
What does the encyclical say about AI and war?
As quoted in the source material, the encyclical says “no algorithm can make war morally acceptable.” The source reports that Leo XIV criticizes AI-enabled weapons and the possibility that such systems could lower the threshold for violence.
Why is Rerum novarum part of the story?
The signing date links Magnifica humanitas to Rerum novarum, Pope Leo XIII’s 1891 encyclical on labor and industrial society. The source reads that timing as a signal that Leo XIV sees AI as a social rupture comparable to industrialization.
Source: Thorsten Meyer AI
