Altman Says Giving AI Religious Meaning Is "a Real Safety Issue"; Axios Says It Targets Anthropic

What Happened

It started with a very short X post. Altman wrote: "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." No names, no tags, no image attached. But in Axios’s reporting, the post was read straight as aimed at Anthropic — with the note that neither OpenAI nor Anthropic responded to its requests for comment.

On September 29, the New York Times reported that Anthropic co-founder Chris Olah, after seeing an advance copy of Pope Leo XIV's encyclical Magnifica Humanitas, proposed pulling Anthropic out of the May 25 presentation of the encyclical — and then tried to persuade the Pope's advisers to take the prospect of conscious AI models more seriously. Olah ultimately attended and spoke at the event. Anthropic has also spent months convening scholars, clergy, philosophers and ethicists — reportedly from more than 15 religious and cross-cultural groups — to discuss the moral status of powerful AI.

Key Facts

1. The post: Altman’s October 3 X post: "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." No company or person named.
2. Axios’s reading: A "veiled shot" at Anthropic. Axios noted that Anthropic co-founder Chris Olah has been meeting religious leaders about whether AI models could be conscious; neither company responded to requests for comment.
3. NYT background: September 29 report: Olah proposed pulling Anthropic out of the presentation of the papal encyclical Magnifica Humanitas, then lobbied the Pope’s advisers to take conscious AI seriously; Olah spoke at the event. Pope Leo XIV has said "algorithms lack the spark of humanity."
4. Anthropic’s position: Does not claim Claude is conscious; its constitution states the company is "not sure whether Claude is a moral patient," calling the question "live enough to warrant caution." Anthropic says convening religious and philosophical voices is meant to bring wider perspectives into AI development discussions.
5. Altman's earlier remarks: Per The Decoder: in 2023 and 2024, Altman said OpenAI's goal was to build "magic intelligence in the sky," and that he hoped to be "on God's side" and "on the side of the angels" working on AI; OpenAI has also talked with religious leaders.

Context

In the same period: earlier this month Anthropic CEO Dario Amodei published a long essay urging the industry to "pace the frontier," with Altman and Elon Musk each signaling agreement; OpenAI shelved GPT-6.1 Astra, originally slated for an October launch, after internal tests judged it "not trustworthy" — it would push ahead on tasks without users knowing; the US FTC is investigating OpenAI, Anthropic and the evaluation group METR; and the White House just had the five big AI companies sign a voluntary safety accord. Axios noted that OpenAI and Anthropic are both preparing for initial public offerings.