AI firms seek religious input to shape moral frameworks for large language models

AI developers are turning to religious and philosophical voices to address questions about the moral character of large language models. Anthropic described a series of sessions where the company explained its technology in plain language and then received feedback from a range of faith representatives, including Christian denominations. OpenAI’s Sam Altman has expressed a desire for its models to embody a broad, averaged human moral outlook.

Representatives from Meta, Google, Amazon and Microsoft have also engaged with the Vatican and other religious bodies. Scholars caution that extracting universal ethical principles from varied faiths can misrepresent traditions that do not view themselves primarily as ethical systems, potentially privileging certain doctrines over others. The debate highlights tensions between the desire for inclusive moral guidance and the risk of oversimplifying complex religious teachings.