OpenAI CEO Sam Altman warns against treating AI as a religious authority

OpenAI CEO Sam Altman has spoken out about the relationship between AI and religion. According to press reports, he warned that granting AI religious force, or treating models as a religious authority, amounts to a real safety issue.

Altman was reported to say the tendency to treat models as spiritual authorities makes him very uncomfortable. His position is clear: AI can serve as a tool, but it should not be given the standing of an object of faith.

Who is speaking matters here. OpenAI under Altman runs one of the most influential frontier model companies, so when its CEO draws a line around what AI should not become, he is also shaping an industry narrative of self-restraint.

From a safety standpoint, his concern points to a social risk rather than a purely technical one. Once users treat model output as unquestionable doctrine, the model's errors, biases and uncertainty get amplified and correction mechanisms stop working, which is the heart of what he calls a real safety issue.

For OpenAI, such public remarks carry product meaning as well. Repeatedly framing its models as tools is both a safety position and a way of managing expectations and the boundaries of responsibility.

What to watch next is whether these warnings stay at the level of rhetoric or turn into concrete product and governance choices, such as measurable changes to model behavior, disclaimers and user guidance.