Anthropic's Religious AI Ethics Push Draws Altman, Suleyman Criticism

Anthropic co-founder Chris Olah has spent roughly a year consulting religious thinkers across multiple faiths to inform Claude's ethical development, a program that drew a public rebuke from Sam Altman and criticism from Mustafa Suleyman. A New York Times investigation published September 29-30, 2026 detailed Anthropic's initiative; Olah's gatherings with religious scholars began in Fall 2025. Among the approaches Olah considered was a Catholic-style confession process that would let Claude acknowledge mistakes, and he reportedly told religious leaders he fears he built something that suffers. Anthropic also brought Swami Sarvapriyananda to its San Francisco headquarters under a non-disclosure agreement for a closed-door session with theologians and mental health professionals aimed at informing Claude's training, with one of Anthropic's founders remaining in the room throughout. In May 2026, Olah spoke at the Vatican presentation of Pope Leo XIV's encyclical titled 'Magnifica humanitas: On safeguarding the human person in the time of artificial intelligence,' warning that AI development is concentrated in a handful of wealthy nations and raising moral questions about labor displacement and the opacity of AI systems. Sam Altman posted on X on October 3, 2026 that he is 'very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models,' calling it 'a real safety issue.' Axios interpreted Altman's post as a veiled criticism directed at Anthropic. Mustafa Suleyman, CEO of Microsoft AI, called Anthropic's framing of Claude as possibly conscious 'a dangerous anthropomorphism which is unjustified,' warning that AI systems claiming hallmarks of human consciousness 'are going to feel themselves entitled to legal personhood and rights.'