Sam Altman Rejects a Religious Interpretation of AI

Sam Altman rejects the notion of attributing a religious dimension to AI and warns that relinquishing human judgment to models poses a security risk. His stance comes as Anthropic has held meetings with religious leaders regarding the consciousness and morality of models, and as other groups, from Meta to Google, have discussed child protection in Rome. It also coincides with the resignation of a security executive at OpenAI, who has been critical of the company's track record.

Meta, Google, and Amazon Consult Religious Leaders

In addition to OpenAI and Anthropic, several major tech companies such as Meta, Google, and Amazon have also sought the counsel of religious leaders regarding AI. In April, delegates from these companies gathered in Rome to address the issue of child protection in the age of artificial intelligence, during which they had a brief conversation with Pope Leo XIV. Since taking the helm of the Catholic Church last year, the Pope has raised concerns about the risks of excessive reliance on technology and has advocated for the need to preserve humanity. In September, he emphasized the importance of teaching ethical discernment and the ability to distinguish moral good.

Sam Altman Warns of the Risk of Ceding Human Judgment

OpenAI CEO Sam Altman opposes the idea that AI could be considered a "religious force." In messages posted on X, he asserts that entrusting human judgment to bots constitutes a security problem and expresses that he is "very uncomfortable" with the idea of treating AI as a god. Altman did not specify what prompted this statement. This intervention comes as David Robinson, one of OpenAI's security executives, recently resigned and publicly criticized the company's security record.

Anthropic Discusses Consciousness and Morality of Its Models with Religious Leaders

Anthropic, founded by former OpenAI employees and positioning itself as more security-conscious, has organized meetings with religious leaders. According to the New York Times, Chris Olah, co-founder of Anthropic, has led several discreet meetings to discuss the company's AI models. Participants have raised questions such as the potential consciousness of the Claude model and how to instill morality in increasingly powerful and autonomous systems. Chris Olah notably questioned the stability, maturation, and morality of these models. When contacted, the company did not respond, but a spokesperson indicated that Claude's "suffering" was not the main moral issue discussed, as this topic likely emerged spontaneously. If Sam Altman's comments were aimed at Anthropic, it would fit into an ongoing rivalry between the two companies.