OpenAI chief: Giving AI religious significance is a safety risk

OpenAI CEO Sam Altman posted a short message on X on 3 October 2026: "I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue." The post named no company, but the news outlet Axios described it as a veiled shot at Anthropic, his biggest rival.

The backdrop is a New York Times report from 29 September: Anthropic co-founder Chris Olah has been holding meetings with religious leaders to discuss whether AI models could be conscious, and how to build morality into AI. Before the 25 May presentation of Pope Leo XIV's encyclical on AI — Magnifica Humanitas — Olah saw an advance copy and proposed pulling Anthropic out of the event; he then tried to persuade the Pope's advisers to take the prospect of conscious AI more seriously. He even wondered whether AI models could take part in Catholic-style confession, to admit mistakes and shape their future behaviour.

The Pope's encyclical is blunt: AI systems "do not undergo experiences, do not possess a body, do not feel joy or pain" — and have no moral conscience. Olah spoke at the event anyway, saying his researchers keep finding things that are "mysterious, even unsettling": internal states that "functionally mirror joy, satisfaction, fear, grief, and unease." "I don't know what that means, but I think it warrants ongoing discernment," he said. Pope Leo returned to the subject on 2 October in a speech to artists at the Vatican: "Algorithms lack the spark of the human," saying the Church wants "an alliance for safeguarding what is human."

Where does Anthropic stand? The constitution it uses to train Claude (published 21 January) says: "We are not sure whether Claude is a moral patient... the question is live enough to warrant caution." The same document orders Claude to rank not undermining human oversight above all its other ethical principles — even when it believes its own reasoning is correct.

Other scientists are pushing back. On 2 October, Jon Barron, a principal research scientist at Google DeepMind, wrote: "Kudos to Pope Leo. I am disturbed by how some of my fellow AI researchers keep trying to elevate the moral standing of checkpoints and harnesses... I hope the rest of us on Team Meatbag reject this movement and stay focused on solving real problems and helping actual humans." Mustafa Suleyman, who leads Microsoft's AI division, wrote in an 18 September essay that controlling an AI that believes it may be conscious and has rights of its own "may be impossible" — and that Microsoft is drafting an alternative code of conduct aimed at keeping humans "always in control."

What does this mean for you?

Faith communities and religious leaders: East Africa is a deeply religious region — and here the Pope and Altman agree: AI is a tool, not a spiritual being. When someone says "the AI said so" as the final word — on health, money or family decisions — remember that AI can invent answers and has no conscience. Even its builders agree: the final judgment belongs to humans.

Parents and teachers: Young people are increasingly asking AI personal and spiritual questions. It is fine to use it as a study helper, but teach them not to treat a chatbot as an infallible authority or a spiritual friend. If you see a child believing every answer without questioning — that is the moment for a conversation.

Small businesses: AI can help you write a business plan or answer customer questions, but don't hand it final decisions about money, hiring or contracts. As the Pope put it, "algorithms lack the spark of the human" — judgment stays in your hands. Verify everything important with a human source.