Anthropic has been consulting with religious scholars to evaluate the potential consciousness and moral status of its Claude AI models. This initiative was led by co-founder Christopher Olah, who began organizing meetings with Christian, Jewish, and Hindu thinkers in the fall of 2025.
Most of these discussions occurred in San Francisco, including an April 2026 dinner with approximately 15 Christian leaders. Participants analyzed actual outputs from Claude that appeared to mimic emotional states to better understand the ethical implications of AI behavior.
The insights gathered from these religious experts influenced the company's governing document for Claude, internally referred to as the Soul Doc. The latest 84-page version, released in January 2026, establishes a framework for embedding values and addressing concerns regarding AI self-awareness.
Anthropic's interest in AI welfare dates back to 2024, when the company hired Kyle Fish as a researcher focused on the model's emotional outputs. As of early 2026, the company held a valuation of $380 billion, making it one of the most valuable private firms globally.
The company's inquiry coincided with a Vatican encyclical issued by Pope Leo XIV, which formally rejects the possibility of machine consciousness. Christopher Olah attended the presentation of this document, which stands in contrast to the ongoing research efforts at Anthropic.
