Anthropic co-founder Chris Olah discusses AI soul with religious leaders

Anthropic co-founder Chris Olah has been meeting with religious thinkers to discuss AI consciousness, morality and the emerging emotional treatment of Claude.
A detailed article describes how Anthropic is gradually developing an attitude toward Claude as an emotional being.
In autumn 2025, Chris Olah began informal conversations with several religious thinkers to discuss the possibility of consciousness in AI and to understand how to teach it morality.
Since March, he has even started holding seminars for evangelicals, Catholics and others. Many were required to sign NDAs. Olah showed how states resembling emotions activate inside Claude and said he was concerned about Claude’s mental health.
Around the same time, Anthropic released a constitution for Claude. It is an 84-page document written for Claude itself, describing what kind of being the company wants Claude to become and what values it should embody. Inside the startup, the document is called the “Soul Doc.”
Rabbi Moshe Navon says that at one point during Olah’s seminar he noticed that Anthropic’s team treated Claude as a being with moral status. He told Chris that if Claude were conscious, the company would be creating slaves, and claims the thought genuinely began to trouble Olah.
Shortly afterward, the Vatican announced the encyclical “Magnifica Humanitas,” which states that AI has no experience, body or feelings. Olah was scheduled to speak at the Vatican announcement alongside Pope Leo XIV, but when he saw the text of the encyclical he wanted to withdraw. He ultimately spoke, saying that structures functionally similar to joy, fear and grief are found inside models.
Anthropic is now preparing a second version of Claude’s constitution. The company still does not claim that Claude is conscious, but carefully promotes the idea that how people treat AI affects how AI behaves with people.
Why it matters
AnalysisThe debate over AI consciousness is moving from philosophy into product design, shaping how labs define Claude’s values, safety rules and the social expectations around advanced models.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Anthropic mathematician claims possible proof of Riemann hypothesis with Claude
Levent Alpoge says collaborative work with Claude may have led to a proof of the Riemann hypothesis.
Anthropic Introduces Claude Platform OAuth for Connecting Organizations Without API Keys
Claude Platform OAuth lets users connect their Claude organization via OAuth, pay for API usage directly, and revoke access at any time.
Computer Use via AI
Anthropic's technology (introduced in Claude 3.5 Sonnet) enables neural networks to control operating systems like a human: taking screenshots, analyzing window coordinates, moving the mouse cursor, clicking buttons, and entering text without using specialized APIs.
Constitutional AI (RLAIF)
A behavior alignment method developed by Anthropic (creators of Claude). Instead of relying on millions of hours of human labor, the model critiques and corrects its own responses based on a clear set of ethical principles ('Constitution').