GOVERNANCEAnthropic
Pope Leo XIV publishes "Magnifica Humanitas", the first Catholic encyclical entirely on AI. Chris Olah (Anthropic co-founder and head of the lab's interpretability research) delivers remarks at the Vatican.
AI governance just got a stakeholder labs can't lobby or out-fund. Religious institutions move on decades, not quarters. Different game.
Olah's "internal states that functionally mirror joy, fear, grief" quote is now on the doctrinal record. That's the most candid public statement on model welfare from any frontier lab. The Catholic Church frames data as a common good. That language will travel to EU AI Act rulemaking.
Get your comms team a one-page brief. This will surface in customer questions. Watch for similar partnerships with Anglican, Jewish, Islamic, Buddhist voices by Q4.
⚡ Why this matters
- The Vatican now has a seat at the AI table. Labs can't lobby or out-fund it.
- Olah said on the doctrinal record that Claude has "internal states that functionally mirror joy, fear, grief, unease." That quote is now permanent.
- AI ethics curricula and regulatory hearings will cite this within 12 months.
🔍 What happened
- May 25, 2026. Vatican publishes Magnifica Humanitas. 42,300 words.
- First Catholic encyclical entirely on AI.
- Leo XIV picked the name to echo Leo XIII's 1891 Rerum Novarum (labor and capital after the industrial revolution).
- Chris Olah, Anthropic co-founder, seated among cardinals. Delivered formal remarks.
- Encyclical positions: data as a "common good"; algorithmic decisions on jobs and credit lack "compassion, mercy, forgiveness"; AI amplifies the powerful; AI energy use is a moral concern.
💬 Smart takes
- Olah at the Vatican: "Every frontier AI lab operates inside incentives that can conflict with doing the right thing. We need critics the incentives cannot bend."
- Pope Leo XIV: "AI systems are more 'cultivated' than 'built.'"
- Simon Willison: "Some of the clearest writing I've seen on the ethics of integrating AI into modern society."
- Skeptic: Encyclicals shape long-arc culture, not next-quarter strategy. Operational impact on lab behavior is unproven.
🧭 Where this goes
- Anthropic launches similar dialogues with Anglican, Jewish, Islamic, Buddhist institutions before year-end.
- EU AI Act and US state bills cite Magnifica Humanitas within 6 months.
- Catholic universities (Notre Dame, Georgetown, BC) adopt it in 2026-2027 syllabi.
- Olah's "internal states" quote becomes the standard model-welfare citation in academic papers.
- OpenAI, DeepMind, Mistral face pressure to send senior leaders to similar dialogues.
🎯 Implication
- For PMs and execs in public-facing roles: get a one-page brief for your comms team. This will land in customer questions and regulator submissions within 90 days.
- For interpretability and safety teams: Olah just made "labs need critics we cannot influence" a publicly stated principle. Engage with it.