Artificial Intelligence Before God: When Silicon Valley Meets the Vatican
Christopher Olah, Anthropic co-founder, told the Vatican that AI models are producing mysterious structures resembling human emotions. Pope Leo replies: machines have no consciousness.
There is an image worth more than a thousand press releases: a thirty-three-year-old researcher, co-founder of one of the most influential artificial intelligence companies on the planet, sitting alongside the Pope in the heart of the Vatican. Not for a blessing, not for an institutional photograph. But to publicly admit, before cardinals, theologians and journalists from every corner of the world, that inside the most advanced artificial intelligence systems, dynamics are emerging that no one yet knows how to fully explain.
What unfolded within the Vatican walls on 25 May 2026, during the presentation of the papal encyclical Magnifica Humanitas — through which Pope Leo addresses for the first time in a systematic way the subject of artificial intelligence — was not a simple academic conference. It was something rarer and more unsettling: an open confrontation between two radically different worldviews, brought face to face on the same stage, without diplomatic filters.
Christopher Olah, born in 1993, co-founder of Anthropic and former OpenAI researcher, is considered one of the world's foremost experts in AI model interpretability, that branch of research dedicated to understanding what is actually happening inside artificial intelligence systems — not just what they produce as output, but what internal processes generate it. It is obscure, technical, often frustrating work. And it is precisely for this reason that the Vatican sought him out: because the Church has sensed something that many governments and public institutions still struggle to admit, namely that we are building increasingly powerful systems without yet having a full understanding of what is emerging within them.
Olah's words before the Vatican assembly were, to put it mildly, not particularly reassuring. «We keep finding mysterious and even unsettling things», declared the researcher, explaining that artificial intelligence models cannot be compared to an aeroplane or a car — systems engineered component by component according to precise specifications. They are instead systems that have grown on architectures inspired by the human brain, trained on the entire heritage of human language and thought. And in Anthropic's laboratories — Olah recounted — researchers are observing internal structures that appear to reflect dynamics typical of human neuroscience: introspection, internal states, behaviours evoking something resembling fear, joy, discomfort, satisfaction. «I don't know what all of this means», he added, in a phrase that echoed through the room like an unexpected confession.
Minutes later, from that same stage, Pope Leo responded with doctrinal clarity that left no room for interpretation: machines do not suffer, do not love, do not have consciousness. A clear position, rooted in two thousand years of theological and philosophical reflection on human identity.
Two visions, one stage, no agreement. And it is precisely this unresolved tension that makes this moment historic. The Vatican did not arrive at the artificial intelligence debate wearing the hat of a religious institution issuing moral condemnations, but rather that of a philosophical institution which for centuries has guarded the oldest and most urgent question: what does it mean to be human?
It is a question that returns with renewed urgency at the moment when machines begin to replicate, with ever greater precision, functions that we believed to belong exclusively to the human sphere: language, reasoning, even something resembling empathy. The decade unfolding before us will be shaped by this question, perhaps more than any other. And the fact that those raising it together are a Silicon Valley researcher and the successor of Peter, at the same moment and in the same place, says something very precise about the magnitude of the challenge that lies ahead.
Source: public statements by Christopher Olah during the presentation of the encyclical Humanitas at the Vatican, 25 May 2026; official documentation of the Holy See.