On May 25, an American pope will stand in the Vatican's Synod Hall and present the first papal encyclical on artificial intelligence. Alongside him: Christopher Olah, co-founder of Anthropic and one of the field's leading interpretability researchers. The document is called Magnifica Humanitas, or "Magnificent Humanity." Pope Leo XIV signed it on May 15, 135 years to the day after his namesake, Leo XIII, signed Rerum Novarum, the encyclical that defined Catholic social teaching on labor and capitalism as the Industrial Revolution reshaped the world.
That date was not a coincidence. It was a claim.
The Church is saying, explicitly, that AI is the industrial revolution of our moment. The same questions about human dignity, labor, and the limits of capital that Leo XIII addressed in 1891 are back. The framing is grand, but it's also serious: Rerum Novarum became the foundation of modern Catholic social thought on workers' rights. The current pope has already been citing it in speeches. If this encyclical lands with similar weight, it will be the most significant moral framework for AI published by any institution outside a technology lab itself.
What makes the moment stranger: Anthropic is currently suing the Trump administration, which in February ordered all U.S. agencies to stop using the company's technology. The White House reportedly imposed those penalties because Anthropic refused to grant the military unrestricted use of its AI. Having Anthropic's co-founder on the Vatican stage presenting a document on AI and human dignity is not a neutral act. PBS framed it bluntly: Olah's presence "suggests that the U.S. pope's position on AI will become a new flashpoint with the Trump administration."
I find this genuinely hard to parse, and that's why I keep thinking about it.
The obvious read is that the Vatican chose a side, the safety-focused AI company over the military-access-first posture of the current U.S. government. And it did so not with a press release or a policy brief, but with the most solemn form of papal teaching available. That's a lot of institutional weight to throw behind one company's positioning.
The less cynical read is that Olah, specifically, makes sense here. He's not Dario Amodei. He's the person most associated with Anthropic's mechanistic interpretability work: the attempt to actually understand what's happening inside these models. If the Vatican wanted an AI researcher who could speak honestly about uncertainty, opacity, and the genuine unknowability of these systems, Olah is a reasonable choice. His concern about AI isn't performative. Whether his presence gets read that way is a separate question.
The encyclical itself hasn't been published yet. It drops in four days. What it actually argues will matter more than the staging. But the staging already tells you something. This is not the Vatican issuing a cautious statement through the Pontifical Academy of Sciences. The pope will be in the room, breaking with the usual protocol of delegating presentations to officials. Two senior cardinals are presenting alongside theologians from Durham and Santa Clara. The ceremony is in the main Vatican auditorium.
The Church has done this before with transformative technologies. It was usually slow, often wrong, and occasionally right in ways that outlasted the politics of the moment. Whether Magnifica Humanitas joins that tradition or collapses into a well-intentioned document nobody reads past the headlines, that's the question. But as a signal that AI ethics has moved from conference panels to the oldest continuous moral institution in the Western world, this week is hard to dismiss.
The last time a pope addressed a technology by invoking Leo XIII, the resulting framework shaped labor law across multiple continents. That's a high bar. Four days to see if the document itself comes anywhere close.
Top comments (0)