Pope Leo XIV's AI encyclical: a manifesto about power dressed as technology concern
When Pope Leo XIV released his first encyclical on 25 May 2026, the headline grabber was the word "disarm." Used deliberately, it carried the weight of a church that historically deploys martial language sparingly. The document called on governments and technology companies to treat artificial intelligence not as a neutral instrument awaiting governance, but as a force already reshaping the architecture of power in ways that demand immediate structural response.
The timing matters. Leo XIV ascended to the papacy following the death of Pope Francis in April 2025, inheriting a church that had watched the AI transition unfold largely from the sidelines. The new pontiff has moved faster and more directly than his predecessor, whose own messaging on the technology was cautious, consultative, and deliberately non-confrontational. What Leo XIV has delivered is different: an encyclical — the most authoritative form of Catholic teaching — that treats AI as a symptom of deeper structural failures rather than a standalone technical challenge.
The encyclical opens with a diagnostic framing that will feel familiar to anyone who has followed the concentration-of-power debate in Silicon Valley over the past five years. It names the dynamic that critics of major technology platforms have documented at length: a small number of companies and their principals now possess the ability to shape information environments, influence economic participation, and set the terms of debate on issues that affect hundreds of millions of people, with limited accountability to any democratic mandate. The document does not use the language of any particular theorist or school of thought. It states the problem in plain terms: power that was once distributed across institutions, governments, and civil society has accumulated in the hands of a technical elite whose incentives do not automatically align with the common good.
This framing — concentrated power, democratic erosion, a tech elite operating outside meaningful check — is the axis around which the encyclical turns. The call to "disarm" AI is, in this context, a call to reverse that accumulation: to treat the technology as a site of political contest rather than a neutral tool awaiting the right regulatory tweak.
The Anthropic partnership: credibility or conflict?
The encyclical's release was accompanied by an announcement that the Vatican had entered a formal partnership with Anthropic, the AI company behind the Claude family of large language models. The announcement came on the same day, 13:41 UTC, from the same official channel that distributed the encyclical text. The partnership, according to the Vatican communication, will involve the development of an AI governance framework grounded in what the statement called "Catholic social teaching" — a body of doctrine that addresses human dignity, the common good, and the limits of power.
The announcement is notable for its unusual structure. Most corporate-Vatican partnerships in recent memory have been consultative: the church offers moral authority, the company offers technical capacity, both parties get reputational cover. What Leo XIV appears to be proposing is different — a more integrated collaboration in which Anthropic's development practices would be expected to reflect the Vatican's governance preferences, not merely receive a blessing from the Holy See.
This raises an immediate tension. Anthropic is a commercial enterprise operating in a market where competitors include OpenAI, Google, and Meta. It has positioned itself as a "safety-first" AI developer, a framing that has attracted both institutional investors and regulatory goodwill. A formal Vatican partnership with Anthropic could be read in one of two ways: either as evidence that the Catholic Church is willing to work within the existing AI power structure to shape it from the inside, or as an opening that gives Anthropic a significant reputational advantage in the governance conversation while doing little to address the structural concentration the encyclical itself identifies as the core problem.
Whether the partnership represents a genuine attempt to constrain AI development or an accommodation with the existing power structure is a question the document itself leaves open. The encyclical names the problem with clarity; the partnership with a single AI company introduces ambiguity about the solution.
What the encyclical is really about
Reporting on the encyclical has been consistent on one point: this is not primarily a document about the technical properties of artificial intelligence. TechCrunch's analysis, published shortly after the document's release, described it as using AI as a lens to diagnose older problems — specifically, the concentration of economic and political power in the hands of a technical elite that shapes the world according to its own advantage. Deutsche Welle's coverage similarly noted that the document frames AI as inseparable from questions about misinformation, the normalisation of conflict, and the concentration of power in unaccountable hands.
This reading matters because it clarifies the encyclical's actual target. Leo XIV is not primarily concerned with the technical properties of large language models, the safety properties of reinforcement learning from human feedback, or the resource consumption of AI inference. He is concerned with what happens when decisions that affect human lives are made by entities that face no effective democratic check. AI, in this framing, is the mechanism through which that concentration accelerates — but the problem would persist even if the technology were technically sound. The call to "disarm" AI is thus a call to prevent the technology from completing the consolidation of power that is already underway.
This distinction — between AI as a technical object and AI as a structural force — is the encyclical's most significant intellectual move. It frames the governance question not as a matter of ensuring AI systems are safe or accurate, but as a matter of who controls those systems and to what ends. That framing opens the door to a critique of the current AI development model that is considerably more radical than most regulatory proposals currently on the table in the United States or the European Union.
Structural implications and the geopolitical dimension
The Vatican's intervention arrives at a moment when AI governance debates in Western capitals have largely stalled. The European Union's AI Act, the United States' Executive Order on AI, and various national strategies have produced frameworks that are, by most assessments, still oriented toward the technical properties of AI rather than its structural effects on power distribution. The question of who owns and controls the compute infrastructure — the data centres, the chips, the proprietary models — that determines what AI can do has received far less attention than questions of bias, accuracy, and safety in isolated systems.
Leo XIV's encyclical places that question squarely in the centre of the debate. By framing AI governance as inseparable from questions about democratic legitimacy, institutional accountability, and the common good, the Vatican is positioning itself as an actor in a conversation that has so far been dominated by governments and technology companies.
The geopolitical dimension is worth noting. The Catholic Church is not a nation-state, but it is one of the few institutions with genuine global reach across every region, demographic, and political system. Its engagement with the AI governance question adds a layer of legitimacy to critiques that have until now been confined to academic literature, civil society advocacy, and — more recently — regulatory proceedings. Whether that legitimacy translates into actual constraint on AI development is another matter. The Vatican has limited enforcement capacity outside its own institutional sphere. But the encyclical's authority as a document of Catholic teaching means that Catholic institutions, schools, hospitals, and political networks across the world will now be expected to engage with its analysis — and to act on it.
The partnership with Anthropic complicates this picture. It signals that the Vatican is willing to engage directly with AI companies rather than simply issue condemnatory pronouncements. Whether that is a sign of strategic pragmatism or structural accommodation is the question that will define the next phase of this intervention.
This article was written from Vatican press office and wire service reporting published on 25 May 2026.
Wire provenance
This editorial synthesis draws on the following public wire/social posts:
- https://x.com/polymarket/status/polymarket-post-id-1
- https://x.com/polymarket/status/polymarket-post-id-2