AI companies seek religious perspectives while scholars question a single moral rulebook

Anthropic describes cross-cultural dialogue; an AP report highlights concerns about reducing living traditions to interchangeable ethical principles.

AI companies are seeking input from religious thinkers as they consider how models should behave. An October 25 Associated Press report describes both support for that dialogue and concern that combining traditions into a general moral framework can lose the context that gives them meaning.

Anthropic’s own May 19 account says it organised discussions with scholars, clergy, philosophers and ethicists from more than 15 religious and cross-cultural groups. The company presents the conversations as a way to broaden perspectives alongside technical work on safe and beneficial AI.

Listening does not remove disagreement

The AP report highlights scholars’ concern that religion is more than a collection of ethical instructions that can be extracted and merged. Consultation can bring valuable perspectives without producing one answer that represents every tradition, secular worldview or individual adherent.

That distinction matters for a product used by people with different beliefs. A model may need to explain a viewpoint fairly, respond appropriately to a user’s own context and recognise genuine disagreement. Those are different tasks from choosing a single worldview to impose on everyone.

Making values visible in product decisions

The practical questions concern how input affects behaviour. Companies can explain which decisions consultation informs, how they handle conflicting views and how users can identify or challenge an inappropriate response. Broad participation is useful when its relationship to actual product choices is understandable.

Neither a meeting with faith leaders nor a technical evaluation can, on its own, establish that a model is morally good. The discussion is about who helps shape AI systems and how those choices are made. Treating religious and secular contributors as varied participants preserves the complexity that a universal rulebook can miss.

Sources

- [Widening the conversation on frontier AI](https://www.anthropic.com/news/widening-conversation-ai) — Anthropic - [To make AI morally good, tech companies are turning to religion. But some say it is not that simple](https://apnews.com/article/2bb91e3334fd6d732edf22288cabf4ec) — Associated Press

*Image: AI-generated editorial illustration; not a photograph of a documented event or an authentic product screen.*