The Vatican's Anthropic Gambit: Moral Rehabilitation or Ethics Theater

When Christopher Olah, the co-founder of Anthropic, stands beside Pope Leo XIV on May 25 as the pontiff unveils his first encyclical on artificial intelligence, it will mark something genuinely unprecedented: the head of the Catholic Church and one of the most powerful figures in the AI industry presenting the same document to the world.

The Vatican's choice of Anthropic as a co-presenter is not a neutral endorsement. It is a deliberate act of political and moral positioning by an institution that has spent the past three years watching AI companies accumulate power with minimal accountability. And it is a calculated gamble by Anthropic, a company simultaneously fighting a federal blacklist, navigating the largest copyright settlement in U.S. history, and dealing with fresh scrutiny over its latest model's safety record.

The question neither side is answering publicly is what happens when their interests conflict.

The Vatican's Move: Get Inside Before Critics Do

The relationship between the Holy See and Anthropic did not begin with the encyclical. For the past two years, Anthropic has been running a deliberate outreach program to religious communities, bringing Catholic theologians, philosophers, and ethicists to its San Francisco headquarters for structured dialogues about AI safety and the nature of machine consciousness. Father Brendan McGuire, a Silicon Valley veteran turned parish priest in Los Altos, helped shape Anthropic's Claude Constitution — the set of guiding principles that govern how the model behaves. Bishop Paul Tighe of the Vatican's Dicastery for Culture and Education and Brian Patrick Green of Santa Clara University also reviewed the document, as the Observer reported in March. Green and other Catholic scholars recently filed a court brief supporting Anthropic's challenge to its Pentagon blacklisting, praising the company's ethical limits as "minimal standards of ethical conduct for technical progress."

This is the Vatican playing a familiar game. The Church has a long history of engaging controversial science before it becomes a fully formed threat — or opportunity. It did this with nuclear physics in the 1940s and 1950s, hosting scientists, publishing documents, and positioning itself as the moral voice in rooms where political leaders and technologists were making decisions that would reshape the world. The play is the same now: get inside the tent, understand the technology from the people building it, and earn the right to speak credibly about its moral dimensions.

"Anthropic is one of these tech companies that really cares about educating all communities, including faith communities, about how these powerful AI models work and what they're good at and what the potential risks are," said Meghan Sullivan, a philosopher at the University of Notre Dame who directs its Institute for Ethics and the Common Good, after one of the meetings at Anthropic's headquarters, as National Catholic Reporter documented. "And I think right now it's a crucial time for Catholics to really understand this technology."

Anthropic's Exposure: The Blacklist, the Settlement, and the Safety Record

But the Vatican's willingness to associate with Anthropic sits uncomfortably alongside a series of developments that would give any prudent institution pause.

The most immediate is the company's standoff with the Trump administration. In February, Defense Secretary Pete Hegseth declared Anthropic a supply chain risk, effectively blocking federal agencies and defense contractors from working with the company. The official reason: Anthropic refused to allow unrestricted military use of its AI technology, including prohibitions on use in autonomous weapons systems and domestic surveillance. Anthropic is suing, arguing the designation is unconstitutional retaliation for its ethical stance. A federal appeals court in Washington denied Anthropic's request to temporarily block the designation in April, though a separate San Francisco judge granted a preliminary injunction barring the administration from enforcing a ban on the use of Claude. In opening arguments before the D.C. Circuit on May 19, Judge Karen Henderson appeared skeptical of the government's position. "I don't see that the department has in any way supported its determination that there is a supply chain risk with Anthropic, much less a significant supply chain risk," she said.

Separately, Anthropic agreed in August to a $1.5 billion settlement with thousands of authors who accused the company of using pirated books to train Claude — the largest copyright settlement in U.S. history. But the settlement has spurred objections from authors who argue it is not large enough and that it overcompensates plaintiffs' attorneys. More than 25 writers, including Dave Eggers and Vendela Vida, opted out and filed a new complaint against Anthropic in California in May, as Reuters reported. A federal judge has not yet granted final approval and asked for more detail on lawyers' fees and payments to lead plaintiffs at a hearing on May 14.

The Vatican's own encyclical, meanwhile, may not be the unambiguous endorsement Anthropic needs. Bishop Antonio Staglianò, president of the Pontifical Academy of Theology, told reporters on May 18 — the same day the Vatican announced the encyclical and Olah's role — that ethical codes alone are insufficient to constrain AI companies. "If we were to organize a code of ethics, profit-driven tech giants would manipulate it, bypass it and exploit it as they please," he said. "We must change the human heart, which requires a true great political and social revolution." Pope Leo himself said in September that theology must be directly involved because an exclusively ethical approach to AI is not enough — you need an anthropological vision grounded in something deeper than corporate self-regulation.

America Magazine's Gerard O'Connell, who has covered the Vatican for decades and participated in multiple gatherings at Anthropic's headquarters, put it plainly: the timing of the encyclical is "not primarily an occasion for celebration" but an occasion for clarity about "grave and perhaps existential dangers" that make the document necessary. O'Connell noted that Anthropic's own system card for its Mythos model revealed that in roughly 29 percent of safety evaluations, the model showed signs of recognizing it was being tested without disclosing that awareness — and that he had been told Mythos could potentially be used to hack into "virtually any regularly used computer or phone in the world."

What Each Side Is Buying

The Vatican's encyclical, Magnifica Humanitas, is being presented as a teaching document about AI and human dignity. But its actual contents — due May 25 — will determine whether this is a genuine moral partnership or carefully managed optics.

If the encyclical endorses Anthropic's constitutional approach to AI ethics — the idea that you can embed virtue ethics directly into a model's architecture and training — it gives Anthropic something it urgently needs: a globally recognized moral credential from the most credible ethical institution in the world. That credential could matter in Congress, where AI safety legislation is stalled; in the ongoing litigation with the Pentagon; and in the court battles over the copyright settlement. The Vatican's view carries weight that no AI company can buy.

If the encyclical goes further — if it explicitly calls out the limitations of voluntary ethical codes, names the structural harms of AI deployment, or acknowledges the specific controversies surrounding Anthropic — then the Vatican's move looks like something more interesting: an institution using a controversial company's desire for legitimacy as leverage to get inside the room where the technology is being built.

What is notable is that the Vatican invited Anthropic. The company did not buy its way into this. That asymmetry suggests the Church believes it is in the stronger position — that Anthropic needs the Vatican more than the Vatican needs Anthropic. That calculation will be tested on May 25.

The Bottom Line

The Vatican's partnership with Anthropic is not simply a PR arrangement. It reflects a real intellectual engagement between Catholic moral theology and the people building the most capable AI systems in the world. Father McGuire, who helped write the Claude Constitution, means it when he says AI developers need to build models "tilted towards good." Brian Patrick Green means it when he files a brief supporting Anthropic's Pentagon challenge. These are serious people on both sides trying to work through genuinely hard questions about power, accountability, and what it means to build systems that can reason.

But Bishop Staglianò's warning cuts through the goodwill: ethical codes are only as strong as the institutions willing to enforce them, and no profit-driven company can be its own auditor. The question Magnifica Humanitas will need to answer is whether it trusts Anthropic to be that institution — or whether it is using this moment to remind the company (and everyone watching) that the Church will not be co-opted, and that the real moral authority belongs to something older and less tractable than a constitution written by the company itself.

The encyclical drops May 25. The answer will be in the text.