OpenAI CEO Sam Altman stated on October 3 that attempting to imbue AI models with religious power or encouraging users to abandon their own judgment constitutes a significant safety issue. While Altman did not name specific companies, his remarks followed reports regarding Anthropic's discussions with the Vatican about AI consciousness.
A September 29 report by The New York Times revealed that Anthropic co-founder Christopher Olah engaged with religious scholars to discuss whether AI models like Claude could possess consciousness or moral status. The company also hosted closed-door sessions with experts from various religious traditions to explore these concepts.
Olah reportedly reviewed a draft of Pope Francis's encyclical on AI and expressed concern over the Vatican's stance that machines cannot possess consciousness. Anthropic team members subsequently lobbied papal advisors to advocate for the possibility of model consciousness, though Olah admitted he remains uncertain about the actual existence of such states.
The Vatican's encyclical, released in May, asserts that current AI systems lack physical bodies, feelings, and moral conscience. It emphasizes that technology should be guided by humans who retain responsibility for its use and boundaries. The document notes that AI can mimic language and analysis without true understanding.
Altman has previously used religious metaphors in his own discourse, including references to 'magic intelligence in the sky' and describing AI development as 'being on the side of the angels.' Observers have also noted his use of mystical imagery on social media, though these instances are subject to interpretation.
