Can AI be a ‘child of God’? Inside Anthropic’s meeting with Christian leaders.

Open original

Anthropic invited Christian leaders to advise on moral guidelines for its chatbot, Claude. The company aims to integrate religious perspectives into AI ethics and address questions about AI’s moral status.

  • Anthropic, now valued around $380 billion, brought in six Christian leaders (including Russell Moore, Paige Cunningham, and Eric Metaxas) to help shape Claude's moral guidelines
  • The discussion tackled deep questions like whether AI could be a "child of God" and how to instill virtues like compassion and honesty into the model
  • Leaders cautioned against simplistic approaches (Moore: you can't just feed the Bible into a model) and pushed for transparency on training data sources
  • Anthropic plans to hold similar consultations with Jewish, Muslim, and secular ethicists as part of a broader effort to ground AI values in human ethical traditions

Anthropic has grown so fast—it’s now worth about $380 billion—that it can hire almost anyone in Silicon Valley. Yet last month its leaders sat down with six prominent Christian figures to shape the moral compass of Claude, their flagship AI chatbot. Among those at the San Francisco meeting were Russell Moore of the Ethics & Religious Liberty Commission, Paige Cunningham from Christianity Today, and Eric Metaxas, a best-selling author. The goal wasn’t a PR stunt. Anthropic’s co-founders, including Dario Amodei, want real guidance on questions like “Can AI be a child of God?” and “How do we teach compassion or honesty to software?” Over two hours, the group zeroed in on virtues—truthfulness, humility, care for the vulnerable—and discussed practical steps. Moore warned against overconfidence: you can’t just load the Bible into a model and call it moral. Cunningham urged transparency about where training data come from. Anthropic engineers sketched out how they might embed case studies of moral dilemmas into Claude’s learning process. Not everyone expects a Christian framework to sway broader AI ethics, but Anthropic plans similar sessions with Jewish, Muslim and secular ethicists next. They see this as part of a broader push to ground AI behavior in human values rather than letting it drift toward the quirks of its programmers.

Ask about this article