Pope Leo XIV's first encyclical invokes the lineage of Rerum novarum to frame artificial intelligence as a question of social doctrine — and a frontier AI lab co founder is on the stage where that question is being put.
On May 15, 2026, the 135th anniversary of Leo XIII's Rerum novarum, Pope Leo XIV signed his first encyclical. Ten days later, the Vatican released it under the Latin title Magnifica humanitas — a phrase that explicitly invokes that lineage and turns it toward a new capital: the computational kind. The document's stated theme, according to Vatican News' announcement of the publication, is "safeguarding the human person in the time of artificial intelligence." That is not Silicon Valley language. It is Catholic social doctrine language, and the choice to use it is the first thing the encyclical says.
Magnifica humanitas was signed on May 15 and released on May 25, 2026, with an official presentation in the Vatican's Synod Hall beginning at 11:30 a.m. on the day of release. The full text is published on Vatican.va as the primary document for any doctrinal reading.
The deliberate anniversary is doing real work. Rerum novarum — "of revolutionary changes" — addressed the conditions of industrial labor in 1891 and is the founding text of modern Catholic social teaching. By signing Magnifica humanitas on its 135th anniversary and giving the new encyclical a title that translates roughly as "magnificent humanity" or "the great human being," the Leo XIV pontificate is signaling that it intends AI to be read in the same doctrinal register: not as a technical question delegated to engineers, but as a question about who the human person is, and who gets to say so, when the means of production have changed again.
That framing — a framing, not a verdict — is what makes the document worth taking seriously beyond the church. As The Guardian reported, the encyclical explicitly situates AI within a Catholic anthropology of human dignity rather than within a utility-maximizing frame. The choice is not neutral: it commits the Vatican to a particular account of what AI systems are for, and to whom they must answer.
What turned a doctrinal release into a news story was the speakers' list. According to Vatican News, the confirmed presenters at the May 25 unveiling include:
That list is itself the story. A frontier-AI lab co-founder is not a typical presenter at a Vatican doctrinal unveiling, and the Religion News Service's framing of the event — that Anthropic is "helping unveil" the encyclical — is the version of the story that has circulated most widely in the press cycle. America Magazine's Vatican dispatch and AP coverage have run in the same direction, and the independent Vatican reporter Diane Montagna has treated the Olah presence as a deliberate institutional signal.
Two things are worth holding at once. First, Olah is on the stage because his technical work — interpretability, the effort to make large neural networks legible from the inside — is genuinely relevant to the question of what an AI system is doing to a human person, and because Anthropic's public Claude Constitution is one of the few corporate artifacts that attempts to state normative commitments about model behavior in a doctrinally legible form. Second, Olah is one of seven listed presenters, and the others represent the institutions — doctrine, integral human development, political theology, Catholic higher education, the Secretary of State, and the Pope himself — that the Vatican actually trusts to frame the question.
Read as a successor to Rerum novarum, the encyclical is not asking whether AI is good or bad. It is asking something older and harder: under conditions of a new kind of capital, what does the church owe to the worker, to the family, to the polity, and to the human person conceived as something that cannot be reduced to a data subject? The Vatican.va text is the place to read that argument on its own terms; the press coverage is useful mainly for documenting how unusual the institutional moment is.
The presence of Lushombo alongside Olah is a reminder that the Vatican did not need to import this question from a US Coast laboratory. Catholic social thought has been asking versions of it since 1891, and it is being asked, in this case, by a global-South scholar as well as by a Silicon Valley co-founder. The stage picture is broader than the headline, and the headline is narrower than the document.
The Vatican has not, on the strength of this event, endorsed any specific AI policy, certified any lab, or accepted any company's account of what "responsible AI" means. The framing chosen in the encyclical — the human person, not the model, the user, or the market — is a refusal of that trade. Whether the institutions that actually procure, regulate, and deploy AI systems will treat the Vatican's framing as binding is a separate question, and one the encyclical does not pretend to answer. That is the part of the story still open, and the part that makes Magnifica humanitas worth reading past the speaker list.