ROME (Diya TV) — Anthropic co-founder Christopher Olah and company representatives privately urged Pope Leo XIV’s advisers to consider whether artificial intelligence could become conscious, according to a New York Times report.
Olah reviewed an advance copy of Leo’s first encyclical, Magnifica Humanitas, before its public presentation in May. His concerns about its position on AI consciousness led him to propose withdrawing Anthropic from the launch event, the Times reported, citing a Vatican organizer. He ultimately participated.
The encyclical, which addresses human dignity in the age of AI, distinguishes machine capabilities from human intelligence. Leo writes that AI systems process data but do not experience joy or pain, develop through relationships or possess a moral conscience. He warns against equating simulated empathy and understanding with human experience.
Anthropic takes a more tentative position. Its constitution for Claude, a document outlining the model’s intended values and behavior, describes its possible consciousness and moral status as unresolved questions.
“We are not sure whether Claude is a moral patient,” the company writes, referring to an entity whose interests might deserve moral consideration. Anthropic says that uncertainty warrants caution and further research.
Olah also acknowledged that uncertainty in his interview with the Times.
“We don’t know if AI models are conscious,” he said.
The Times reported that Anthropic declined to discuss its private conversations with the pope’s advisers.
The company has separately confirmed discussions with religious scholars, philosophers, clergy and other specialists about shaping AI behavior. Anthropic says those exchanges could inform Claude’s training, constitution and safety evaluations. It says the effort seeks perspectives across religious and secular traditions rather than adopting one worldview.
Leo’s encyclical also calls for human accountability in AI development and opposes entrusting lethal or otherwise irreversible decisions to artificial systems.
