The pace of AI development is faster than most of us can comfortably keep up with — and the public conversation around it swings between hype and fear. This space exists to cut through both. What is real, what is overstated, what the research actually shows, and what it means for people who aren’t building the systems but are living alongside them.
The goal is calm, accurate, and useful. No fear-mongering. No cheerleading. Just the honest picture, and resources for ordinary people to have a say in how this technology gets built and governed.
An unusual partnership — Anthropic and the Vatican
On May 25, 2026, something remarkable happened in the Vatican Synod Hall. Pope Leo XIV released his first encyclical, Magnifica Humanitas — the Catholic Church’s most significant teaching on AI to date. Sitting next to him on stage was Chris Olah, co-founder of Anthropic (the AI company that makes Claude) and a self-described atheist.
Olah leads Anthropic’s mechanistic interpretability research — the work of trying to look inside AI systems and understand what’s actually happening in their billions of internal connections. His message to a global audience was startling for someone in his role: even well-intentioned AI labs operate inside commercial and competitive incentives that can pull them away from doing the right thing. His conclusion was that the development of frontier AI cannot be steered by frontier AI labs alone. Outside scrutiny — from religious institutions, governments, and civil society — is essential.
He also said something that should give all of us pause. Speaking about what his team has discovered inside Anthropic’s own models, Olah described finding internal structures that mirror results from human neuroscience, evidence of introspection, and internal states that functionally resemble joy, satisfaction, fear, grief, and unease. “I don’t know what that means,” he said, “but I think it warrants ongoing discernment.”
The convergence is the point. An atheist machine learning researcher and a Pope, on the same stage, broadly agreeing: the people building AI cannot be the only ones deciding what AI becomes.
Start here: the encyclical in 5 minutes
The full encyclical is 30+ pages. I wrote a plain-English one-pager that captures the core argument, what concerns the Pope, what he asks of us, and why Olah’s presence mattered.
The Pope’s First Words on AI — A One-Page Guide to Magnifica Humanitas →
Read the source documents
The Catholic Church has now issued two major documents on AI in just over a year. Both are public, free, and worth reading even if you’re not Catholic — they represent careful ethical reasoning from a tradition that has spent two thousand years thinking about technology, work, and human dignity.
- Magnifica Humanitas →
Pope Leo XIV’s first encyclical (May 25, 2026). His full teaching on safeguarding the human person in the time of artificial intelligence. - Pope Leo XIV’s launch address →
A separate, shorter speech the Pope delivered alongside Chris Olah — not the encyclical itself, but the Pope’s own framing of why he wrote it. He thanks Olah by name, draws the parallel to Leo XIII’s 1891 encyclical on industrial workers, and makes a striking call: that “artificial intelligence needs to be disarmed.” - Antiqua et Nova →
Earlier doctrinal note from the Dicastery for the Doctrine of the Faith (January 28, 2025). 117 paragraphs on the relationship between artificial intelligence and human intelligence.
How to make an impact
Reading is the first step. Doing something about it is the second. If you want to add your voice to how AI gets built and governed, here’s where to start.
How to make an impact — government agencies, petitions, and the EFF →