OpenAI CEO Sam Altman said treating artificial intelligence models as a religious authority and leaving human decision-making to these systems posed a serious security risk.
Altman made the comments on Saturday on the U.S.-based platform X. He said he was uncomfortable with attributing religious power to artificial intelligence and transferring human judgment to these systems.

Anthropic's research into consciousness and model welfare
Altman's comments came after reports about private discussions held by rival artificial intelligence company Anthropic with thinkers in religion and philosophy.
The New York Times reported that the discussions covered whether artificial intelligence systems such as Claude could be conscious and how religious thought could contribute to shaping the systems' behavior.

Anthropic is also researching whether advanced artificial intelligence systems can have experiences that require moral consideration. The company calls this work "model welfare."
Suleyman opposed systems that appear conscious
Microsoft AI CEO Mustafa Suleyman also warned that developing artificial intelligence systems that appear to be conscious could create new risks.

Suleyman argued that convincing people that artificial intelligence experiences subjective states could have dangerous consequences. He also backed the "Humanistic AI" approach, which envisions artificial intelligence systems remaining tools that serve people rather than becoming entities with interests of their own.