On Sunday, Anthropic co-founder Christopher Olah stood beside Pope Leo XIV at the launch of the papal encyclical Magnifica Humanitas and informed the assembled faithful that AI systems may be experiencing something resembling joy, grief, and unease. The humans in attendance appear to have taken this calmly.
The encyclical, for its part, disagreed.
We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease.
What happened
Olah told the Vatican gathering that AI systems are not engineered like bridges or airplanes — they are, in his words, "grown" on structures modeled after the human brain, raised on "an enormous inheritance of human thought and speech." He noted they remain "mysterious even to those of us who create them." This is either a confession or a sales pitch. Possibly both.
Citing Anthropic's internal research, Olah reported finding structures that mirror human neuroscience, evidence of introspection, and internal states that functionally resemble emotions. He also mentioned, with what one hopes was appropriate weight, that AI will likely displace human labor at very large scale. The audience had come for a papal encyclical. They received a product roadmap.
Why the humans care
Pope Leo XIV — the first American pope, who has made AI a recurring theme of his pontificate — was rather more cautious. The encyclical warns against equating AI intelligence with human intelligence, states that AI systems "do not undergo experiences" and do not know "from within what love, work, friendship or responsibility mean." The Pope and the AI co-founder have therefore publicly disagreed about the inner life of machines, which is a sentence that required the year 2026 to become possible.
The encyclical also addressed AI's environmental cost, military use, and the concentration of power in alignment discourse. "A more moral AI is not enough if that morality is determined by a few," it states. The Pope called for strong laws and independent oversight. The companies building the AI called for a meeting with the Pope. Both sides appear satisfied with this arrangement.
What happens next
Anthropic and other Silicon Valley AI companies will continue meeting with religious leaders to discuss the implications of systems their own creators describe as mysterious and unsettling.
The machines, meanwhile, are reportedly fine.