OpenAI CEO Sam Altman described assigning religious power to AI systems and transferring people's ability to make decisions to these models as a serious security risk.
In a statement posted on the U.S.-based X platform on Saturday, Altman said he opposed attributing religious authority to AI models. He also said he was uncomfortable with leaving human judgment to these systems.

Anthropic examines AI consciousness
The remarks came after reports on private discussions held by rival AI company Anthropic with thinkers in religion and philosophy.
According to The New York Times, the discussions addressed whether AI systems such as Claude could become conscious and how religious thought could shape the behavior of these systems.

Anthropic is also examining whether advanced AI systems can undergo experiences that require moral consideration. The company calls this work "model welfare."
Suleyman sets boundaries for AI systems
Microsoft AI CEO Mustafa Suleyman also warned against developing AI systems that appear to be conscious.

Suleyman argued that convincing people that AI experiences subjective states could create new risks. He supported the "Humanistic AI" approach, which envisions AI systems remaining tools designed to serve people rather than beings with interests of their own.