OpenAI CEO Sam Altman stated on October 3, 2026, that he is uncomfortable with attributing religious status or moral consciousness to AI models. Altman framed this perspective as a safety concern, warning that treating chatbots as entities with souls could lead to over-reliance on their outputs.
The comments follow a New York Times report from late September 2026 regarding an initiative led by Anthropic co-founder Christopher Olah. Since fall 2025, Anthropic has engaged religious scholars to discuss morality and consciousness in AI systems.
Anthropic, founded by Dario Amodei in 2021, utilizes an 84-page governing document for its Claude model, referred to internally as the Soul Doc. This approach contrasts with Altman’s view, which treats AI strictly as a tool controlled by human users.
The disagreement highlights a divide in AI safety theories between OpenAI and Anthropic. Industry observers note that these differing philosophies regarding model consciousness may influence future partnerships, procurement decisions, and regulatory assessments.
