Sam Altman Warns Giving AI 'Religious Force' Could Be a 'Real Safety Issue:' 'Very Uncomfortable'

OpenAI CEO Sam Altman said he is “very uncomfortable” with efforts to give AI a “religious force” or treat models as a substitute for human judgment, calling it a “real safety issue.” In a post on X, Altman referenced concerns about people trying to assign religious power or surrender human decision-making to AI systems. His comments came days after a New York Times report described Anthropic private meetings with religious scholars about questions including AI consciousness, morality, and how Claude should behave. Altman’s remarks also followed the departure of OpenAI safety leader David Robinson, who criticized the company’s risk-management approach. The broader backdrop includes heightened scrutiny on AI labs’ safety as capabilities accelerate.