OpenAI CEO Sam Altman Warns Against Giving AI Religious Authority

OpenAI CEO Sam Altman wants people to keep their judgment firmly in human hands, not hand it over to a chatbot.

“I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue,” he wrote on X on Oct. 3. His warning puts a practical question in front of families: How much authority should a convincing digital voice carry?

Altman did not name a company, religious group or particular incident in the post. Nor did he say that asking an AI system about religion amounts to surrendering human judgment; his criticism focused on giving the technology religious power or yielding judgment to it.

That distinction matters. A chatbot can help someone organize questions for a Bible study without becoming the authority that decides what Scripture means, how a family should live or whom a person should trust.

The wider discussion reaches beyond how people use these tools. AI developers also debate whether increasingly capable systems could have experiences of their own, and what responsibilities that possibility would create.

Anthropic, the company behind Claude, announced a research program in April 2025 to explore whether AI systems might warrant moral consideration. The company said its work would examine possible model preferences, signs of distress and ways to address welfare concerns.

“There’s no scientific consensus on whether current or future AI systems could be conscious, or could have experiences that deserve consideration,” Anthropic wrote in that announcement. It also acknowledged that researchers lack agreement on how to approach those questions.

Investigating a possibility does not establish it as fact, much less give a machine spiritual authority. Altman’s short post did not identify Anthropic as its target, and questions about possible AI consciousness differ from questions about whether humans should defer to a chatbot.

Anthropic has also described how developers shape an assistant’s conversational character. In a June 2024 explanation of Claude’s training, the company discussed encouraging traits such as curiosity, thoughtfulness and openness to different perspectives.

“We want people to know that they’re interacting with a language model and not a person,” the company wrote. That reminder offers a useful starting point for families: Warm, thoughtful-sounding responses do not turn software into a pastor, parent or trusted friend.

For parents, the conversation can begin before a child brings a big spiritual question to an app. Ask what the child uses it for, look at an answer together and discuss the difference between finding information and giving something the final say.

Keep personal details private, check sources rather than accepting a polished answer at face value, and encourage children to bring troubling conversations to a trusted adult. A tool should help a family think, not quietly take over the thinking.

Scripture already gives us a better habit than unquestioning acceptance. In Acts 17:11, the Bereans examined the Scriptures daily to test what they heard, even when the speaker was the apostle Paul.

We can bring that same care to an answer on a screen. There is room to use technology wisely without handing it the place that belongs to God, His Word and the people He calls us to love.

Questions or comments? Please write to us here.