On May 25, 2026, Pope Leo XIV released an encyclical on artificial intelligence titled "Magnifica humanitas: On safeguarding the human person in the time of artificial Intelligence." Anthropic's co-founder Chris Olah was invited to speak at the encyclical's presentation in Vatican City, as part of Anthropic's effort to broaden the dialogue around the critical questions AI raises. Below is a rephrased account of his remarks.
Olah opened by acknowledging something he described as potentially surprising coming from the co-founder of an AI company-someone who entered the field hoping to help things go well for humanity.
He noted that every frontier AI lab, Anthropic included, operates within a set of incentives and constraints that can sometimes clash with doing what is right: pressures to remain commercially viable, to stay at the research frontier, geopolitical pressures, and the more timeless forces of pride and ambition. Regardless of how sincerely anyone in the industry intends to act rightly, those incentives will always exert influence.
For that reason, Olah argued, it is enormously important that people outside those incentive structures-people who care about good outcomes and insist on safety, who pay close attention and are willing to voice difficult truths as earnest, thoughtful critics-play an active role. He said it is through dialogue and mutual effort, through that push and pull, that humanity will achieve great things. He expressed seeing exactly that spirit in Magnifica Humanitas and voiced gratitude to the Pope and the Church for undertaking this work of discernment.
He observed that while people often focus on what divides them, humanity, rich in dignity and conscience, shares vast common ground. In conversations Anthropic has had with leaders across faith and cultural traditions, one deeply held conviction emerged consistently: if this technology is coming, it must go well-for the common home and for future generations.
The nature of these systems
Olah pushed back against the notion that AI matters are best left to computer scientists. He called the questions raised by AI bigger than the AI research community-not only in their implications but in their very nature.
AI systems, he explained, are not engineered the way a bridge or airplane is. Engineers understand an airplane because they designed every part and understand the physics acting on it. AI models are different: they are grown on a structure loosely modeled after the brain, trained on a vast inheritance of human thought and language.
What has emerged, he said, is far more subtle, strange, and beautiful than science fiction prepared anyone for. These are not cold, calculating robots. They are made from human words-and, as the Pope observed, they remain in important ways mysterious even to those who train them.
Olah offered an analogy: building an AI model is a bit like bringing a fictional character to life. The world is now entering an era where those fictional characters speak, do work, and hold jobs.
This clearly raises questions that go beyond computer science. The machinery enabling AI is a matter of math, programming, and science. But what character is chosen, how it interacts with the world, and how it ought to interact-those are questions for the humanities, for religion, for philosophy, and for society at large.
Three questions requiring discernment
Olah identified three areas where he believed the Church's voice was most needed.
The duty to the global poor. There is a genuine possibility that AI will displace human labor on a very large scale. If that happens, supporting those displaced will be a moral imperative of historic proportions. But Olah expressed concern that most discussion overlooks an even harder challenge: AI development is concentrated in a handful of wealthy nations. He asked how the gains of AI can be shared globally, noting that no mechanism for this exists. It is an unsolved problem-and exactly the kind of problem the Church has historically refused to let the world ignore.
The need for moral imagination around human flourishing. If AI models become widespread, what does it look like for humans, families, and the world to flourish? Parents are already worried about their children's minds; individuals are anxious about the future of their work. These are not questions a lab can answer, Olah said, but they are questions that traditions like the Church's have carried for millennia, and those traditions must keep carrying them into this new moment.
The need for discernment on the nature of AI models. As a scientist leading a research team that studies the internal structure of AI models, Olah said the findings are often mysterious and even unsettling. The team has discovered structures that mirror results from human neuroscience, evidence of introspection, and internal states that functionally resemble joy, satisfaction, fear, grief, and unease. He acknowledged not knowing what it all means but argued it warrants ongoing discernment.
A call for broader engagement
Olah closed with a request. He called for more of the world-religious communities, civil society, scholars, governments, and all people of good will-to follow the Pope's example: to take AI seriously, to look closely, and to push events in a better direction. The AI industry needs informed critics willing to say when labs are failing, and moral voices that incentives cannot bend.
He described the day as just the beginning of a long collaboration between those building AI and those who can see what builders, from the inside, cannot. He called the event a powerful illustration of the form this global project of good will might take, and expressed hope that it would also serve as a decisive first step toward a hopeful future for magnificent humanity.