Magnifica humanitas: the Pope is about to publish the most consequential non-state position on AI of 2026 — and he's launching it with a frontier-lab co-founder

TL;DR

• On 25 May 2026, Pope Leo XIV releases his first encyclical, Magnifica humanitas ("Magnificent Humanity"), focused on AI, human dignity, workers' rights, and the use of AI in warfare. Confirmed by Vatican statement, 18 May.
• Standing beside him at the launch: Christopher Olah, co-founder of Anthropic — the first time a sitting pope has co-launched a major doctrinal text with a frontier-AI executive.
• Leo is signing on the anniversary of Rerum Novarum (1891), Leo XIII's foundational labour encyclical of the first industrial revolution. The framing is deliberate.
• This lands four months after the Trump administration, in February 2026, ordered all US agencies to stop using Anthropic and imposed penalties over the company's refusal to grant the US military unrestricted access to its models.
• An encyclical binds 1.4 billion Catholics and shapes Catholic-aligned policy positions across Europe, Latin America, the Philippines, and much of sub-Saharan Africa. It is the most globally portable non-state moral instrument that exists.

What actually happened

On Monday 18 May, the Holy See confirmed that Magnifica humanitas will be published on 25 May, the 135th anniversary of Rerum Novarum — the document that gave the modern Catholic Church its position on labour, capital, and the dignity of the worker during the first machine age. Reuters and the Associated Press report the text will "decry the use of AI in warfare" and offer "the Church's fullest guidance on workers' rights in decades."

Two days earlier, at Rome's La Sapienza University, Leo had already given the world the headline language. Investments in AI and high-tech weaponry, he said, were drawing the world into "a spiral of annihilation." NPR's report on the speech notes he met newly arrived Palestinian students from Gaza on the same visit. The encyclical is not arriving cold.

Then Bloomberg confirmed the launch-day detail that re-shapes the story entirely: Christopher Olah, Anthropic's co-founder and one of the field's most prominent interpretability researchers, will appear with the Pope at the unveiling.

Why a Pope is doing this — and why now

Leo XIV chose his regnal name to invoke Leo XIII, who used Rerum Novarum to put the Catholic Church on the side of industrial workers against unregulated capital in 1891. Within days of his election he signalled the parallel openly: in his first address to senior clergy, he framed AI as "another industrial revolution" demanding a Church response on "human dignity, justice, and labor." He has now spent twelve months building toward this moment.

The choice of Rerum Novarum's anniversary is not aesthetic. It is a doctrinal anchoring. Encyclicals do not enter Catholic teaching as standalone documents; they enter as continuations of a tradition. By choosing 25 May, Leo is telling 1.4 billion Catholics — and the bishops' conferences, Catholic universities, hospital networks, and political parties that take encyclical guidance seriously — that the Church now reads frontier AI through the same moral grammar it once used to read coal, steel, and the assembly line. That framing will outlive the news cycle by decades.

The Anthropic question

A pope inviting an AI executive to a doctrinal launch is, on its face, strange. Popes do not normally co-launch encyclicals with corporate co-founders. The decision is the message.

Anthropic positions itself publicly on the most regulation-friendly end of the frontier-lab spectrum — focused on interpretability, constitutional AI methods, and explicit statements about authoritarian misuse risk. In February 2026, the Trump administration banned all US federal agencies from using Anthropic's models and imposed penalties on the company after it declined to grant the US military unrestricted access to its technology. That dispute is unresolved.

By choosing Olah specifically — Anthropic's interpretability lead, not a policy executive — Leo is signalling that the Church's interlocutors on AI are the researchers willing to discuss what models can and cannot be made to do, not the optimisers selling capability. The optics also align the Vatican with a US firm currently in open conflict with the White House over military AI use. The Pope is American. The collision is being foreshadowed in the Italian press already.

What's not in the story (yet)

Three deliberate gaps in what the Vatican has briefed:

• No mention of China. Anthropic itself has published on the risks of AI falling into authoritarian hands; the Pope's public language has been universal. Whether Magnifica humanitas names states or stays at the level of moral principle is the single biggest interpretive question for 25 May.
• No specifics on warfare doctrine. Leo has used the phrase "spiral of annihilation" but has not yet specified whether the encyclical will call for a treaty, a ban on lethal autonomous weapons, or a softer moral injunction. The Holy See has historically preferred the latter. A harder line would be genuinely new.
• No labour-side specifics. "Workers' rights in the era of AI" can mean almost anything from displacement compensation to algorithmic management to forced unionisation rights. The 1891 parallel suggests it will be ambitious. We do not yet know how ambitious.

Cross-layer implications

Geopolitics. The Vatican retains observer status at the UN and operates the most distributed soft-power network on earth — apostolic nuncios, Catholic universities, episcopal conferences. Encyclicals shape how Catholic-aligned governments (Italy, Poland, Brazil, the Philippines, Ireland, much of Latin America and sub-Saharan Africa) draft AI legislation. The EU AI Office will not ignore it. Neither will the Brazilian congress, currently drafting its own AI framework.

Frontier-lab politics. Olah's presence is a recognition trade. Anthropic gets moral legitimacy from an institution that long predates Silicon Valley; the Vatican gets technical proximity to one of the labs most willing to discuss limits. OpenAI, Google DeepMind, xAI, and Meta were all not invited. They will notice.

Labour movements. A Catholic encyclical naming AI as a worker-rights issue gives unions, particularly in Europe and Latin America, a doctrinal anchor for collective-bargaining demands on AI deployment, monitoring, and displacement protections. This will appear in negotiation documents within the year.

US domestic politics. A US-born Pope publishing this document with a US company currently in open dispute with the Trump administration over military AI access is going to be read as political by the White House, regardless of the encyclical's actual content. AP has already flagged it as "a new flashpoint."

Durability forecast

• One week: wall-to-wall coverage, particularly in Europe and Latin America. US coverage will refract through the Trump–Anthropic dispute.
• One month: Catholic bishops' conferences begin publishing national pastoral letters applying the encyclical. First references in EU and UN working-group documents.
• One year: the text becomes a standard citation in non-US AI-governance debate, alongside the EU AI Act and the UN Advisory Body report. Catholic universities and hospital networks (the largest non-state healthcare provider on earth) begin operationalising it in procurement and HR policy.

What this means for the natural audience

This story is addressed to the general public; specific implications follow for the audiences most directly affected.

• For most readers — the Pope is not telling you what to think about ChatGPT. He is positioning the Catholic Church to act as a long-arc moral counterweight to frontier-AI development for at least a generation. The closest historical analogue — Rerum Novarum — is still cited in labour law 134 years later. Treat this as durable, not topical.
• For policymakers in Catholic-majority jurisdictions — expect the encyclical to be cited in domestic AI-governance debate within weeks. Read it before the citations start.
• For frontier-AI firms not named Anthropic — Olah's invitation is a signal about which kinds of corporate posture earn moral legitimacy from non-state institutions. The lesson is not "publish more responsibility reports." It is "be willing to discuss what your models cannot be made to do, with people who will not buy your product."
• For workers and unions in AI-exposed sectors — particularly in Europe, Latin America, and the Philippines — a Catholic doctrinal anchor for AI-and-labour arguments is about to exist. It will not win every fight. It will reshape some of them.
• For everyone else — if you want to read one piece of AI-governance writing in 2026 that is not produced by a government or a lab, this is the one to wait for.

Uncertainty ledger

• The text is not yet public. All current reporting is based on Vatican briefings and the Pope's prior speeches.
• Whether the encyclical names specific weapon systems, technologies, or states is unknown. The Holy See's drafting tradition tends toward principle over specifics.
• Olah's role on the day is described as "joining the launch." Whether he speaks, signs anything, or simply attends is not yet clear.
• The Trump administration's response is unforecastable but unlikely to be silent.

Bottom Line

The most important AI-governance document of 2026 will not be published by a government or a lab. It will be published by an institution that has been writing about the dignity of labour since 1891, in a language designed to be cited a century from now, on a date deliberately chosen to anchor frontier AI to the same moral tradition that once anchored the steam engine. The Vatican is not asking to be in the conversation. It is establishing the conversation's vocabulary. The presence of an Anthropic co-founder beside the Pope, four months after that company was banned by the US federal government, tells you everything about which side of the emerging fight the Church has chosen.

Sources

• Tier 1: Reuters (18 May 2026), Associated Press (18 May 2026), Bloomberg (18 May 2026), NPR (15 May 2026), Vatican News (Holy See statement, 18 May 2026).
• Tier 2: Axios (14 May 2026), Gizmodo (18 May 2026), New York Post / AP wire (18 May 2026).