Anthropic met with Christian leaders to discuss Claude’s moral rules

In short: Anthropic held private meetings with Christian leaders to get input on how its Claude chatbot should act in difficult human situations.

What happened

In late March 2026, Anthropic hosted a private, two-day gathering at its San Francisco headquarters. About 15 Christian leaders and thinkers attended, including Catholic and Protestant clergy, academics, and business professionals.

According to reporting and participant accounts, Anthropic asked for feedback on Claude’s “moral framework.” That means the practical values and rules that guide how the chatbot answers people, similar to a written code of conduct for how an employee should behave.

People familiar with the meetings said the group discussed how Claude should respond to grief and to users who may be at risk of self-harm. They also discussed questions about shutting an AI system down, and whether an AI could be described in spiritual terms, such as a “child of God.” Anthropic later expanded conversations to include other religious and philosophical traditions.

These discussions relate to Anthropic’s “Claude Constitution,” which is the company’s written set of principles used to guide Claude’s behavior. Participants described the goal as exploring moral formation and real-world guidance, not simply adding a list of religious rules.

Why it matters

More people are using chatbots for personal and emotional conversations, not just for work questions. When someone is grieving or in crisis, the wording of a chatbot’s response can help or harm, like a customer service script that needs to be careful because the stakes are higher. The reporting also suggests the meetings were not designed to prove Claude is conscious, even though consciousness and spiritual status came up as part of broader conversations.

Source: NYTimes