The Pope Calls to Disarm AI — Then Partnered with Anthropic

Pope Leo XIV's opening act as pontiff crystallised a tension that has shadowed institutions of moral authority since the industrial age: how do you rail against a technology, then shake hands with its most powerful purveyor? On 25 May 2026, the new pope issued twin communications — a stark public warning against AI's capacity to usher in "new digital slaveries," and simultaneously announced a formal partnership with Anthropic, the AI safety-focused laboratory backed by Amazon and Google. The gap between the two gestures is not incidental. It defines the Vatican's predicament.

The encyclical — his first — frames AI not primarily as a technical problem but as a structural one. According to TechCrunch's reading of the document, Pope Leo XIV uses artificial intelligence as a lens to diagnose older pathologies: concentrated power, the erosion of democratic processes, and a technology elite that increasingly shapes the world to its own commercial and political advantage. The diagnosis is not radical. It tracks what a generation of civil-society advocates, antitrust regulators, and academic critics have documented independently about platform capitalism. That a religious institution arrives at the same inflection point does not make the argument wrong. It does, however, raise a question about what institutional leverage the Catholic Church actually commands in response.

The partnership complicates the critique

Anthropic is not a typical technology company in the Vatican's eyes — or so the framing suggests. The lab's stated mission centres on AI safety and constitutional AI, a research orientation that aligns, at least rhetorically, with the Vatican's stated concern that machines should not dominate humanity. Pope Leo XIV's call on 25 May 2026 to "disarm" AI, with technology prohibited from dominating human affairs, reads as an invitation to exactly the kind of self-limitation that Anthropic's governance model claims to pursue. A partnership with the Church would give Anthropic precisely the institutional legitimacy that a company navigating intense regulatory scrutiny in Europe and Washington requires.

For the Vatican, the calculus is presumably different but not uncomplicated. A formal relationship with Anthropic buys access — to technical briefings, to influence over development norms, to a seat at a table where the terms of AI governance are being written. It does not buy the ability to reverse those terms. Institutional讓子 has historically operated through soft power: moral suasion, diplomatic networks, the gravitational pull of two millennia of institutional permanence. Those tools have real effect in geopolitics, in mediation, in the framing of humanitarian norms. Whether they register at all in a boardroom where the agenda item is model capability and market share is a different question entirely.

The history that haunts this moment

The Church has navigated technological disruption before. It survived Gutenberg, industrialisation, nuclear weapons, the原子时代 — adapting, condemning, endorsing, and occasionally co-opting the technologies of each era. On each occasion, the institutional critique arrived after the technology had matured enough to have backers, stakehold利益, and a structural position in the global economy that made outright antagonism self-defeating. AI is following the same script. By the time Pope Leo XIV's encyclical landed on 25 May 2026, Anthropic, OpenAI, Google, and Microsoft had already shaped the regulatory environment in Washington and Brussels through lobbying, sandbox partnerships, and the quiet exportation of their own governance framings into legislative text. The Vatican's entry into that environment, five years into a technology's mainstream deployment, is late but not entirely without precedent.

The harder question is what "disarm" actually means in this context. Pope Leo XIV was not calling for a moratorium on AI development — no such moratorium has genuine political support in the jurisdictions where AI is being built. He was, rather, making a moral framing available to political actors who might otherwise lack one. That is not nothing. Norms do work. The framing of slavery — applied historically to forced labour, to chattel trade, to colonial extraction — carries enough historical weight to complicate any technology company's effort to paint its products as inherently liberatory. But norms work slowly, and the harms that might justify invoking them are already present.

What the Vatican can and cannot do

The sources do not specify the formal substance of the Anthropic partnership — whether it includes direct investment, advisory roles for Vatican officials, or shared governance standards. Without that specificity, the deal's operational weight remains unclear. What is clear is the optics: a pope who warns against digital domination, and a technology lab whose primary competitive advantage depends on the argument that it alone can be trusted to build AI responsibly. Anthropic needs the Vatican's blessing more than the Vatican needs Anthropic's technical access. That asymmetry may, in time, prove to be the Vatican's actual leverage.

Pope Leo XIV's encyclical names a real problem. Concentrated power in the hands of a small class of technology executives, operating across jurisdictions with minimal democratic accountability, is not a hypothetical — it is the documented architecture of the current AI moment. That the Vatican's first act is to name it, and then immediately to formalise a relationship with one of the system's primary constituents, tells us something about the limits of moral authority in an era when moral authority has been substantially displaced by platform authority. The pope can warn against slaveries old and new. Whether anyone building the new infrastructure is required to listen is a separate matter entirely — and one the sources decline to resolve.

This piece reflects Monexus's coverage of AI governance alongside developments in multilateral institutional power. The Vatican's engagement with a leading AI laboratory, juxtaposed with its simultaneous critique, represents a pattern of institutional co-optation this desk will continue to monitor.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://x.com/polymarket/status/1924688392799441089
• https://x.com/polymarket/status/1924672876045148416
• https://x.com/polymarket/status/1924652959890743386