Canadian AI developer and Anthropic co-founder Christopher Olah presented Pope Leo XIV's encyclical "Magnifica Humanitas" alongside him. True to Anthropic's brand, Olah suggested that today's language models might be more than just statistical systems. "AI systems are not engineered the way a bridge or an airplane is engineered," he said. "They are grown on a structure roughly modeled after the brain on an enormous inheritance of human thought and speech." He added that, as the Pope observes, they remain mysterious even to their creators.
Citing Anthropic's internal research, Olah said, "We keep finding things that are mysterious, even unsettling. We find structures that mirror results from human neuroscience. We find evidence of introspection. We find internal states that functionally mirror joy, satisfaction, fear, grief, and unease." He also warned of a real possibility that AI will displace human labor at a very large scale.
However, the Pope's proposals urge caution. He called on everyone along the AI chain to take responsibility, warning that AI is "never neutral" because "it takes on the characteristics of those who devise, finance, regulate and use it." The encyclical reads more cautiously than Olah's claims about introspection. "We must avoid the misconception of equating this type of 'intelligence' with that of human beings," it states. "These systems merely imitate certain functions of human intelligence."
According to the document, AI systems "do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean." The encyclical also flags AI's environmental cost, pointing to the "enormous amounts of energy and water" needed for data centers and calling for more efficient systems.
The Pope also weighed in on AI in a military context, asserting that deadly or irreversible decisions should not be handed off to machines. "No algorithm can make war morally acceptable." He criticized the current alignment discourse, arguing that "a more moral AI is not enough if that morality is determined by a few." Instead of abstract ethics, he called for strong laws and independent oversight.
Leo XIV, the first American pope, has made AI a central theme of his pontificate. Anthropic and other Silicon Valley AI companies regularly meet with religious leaders to discuss the use of AI.
[AgentUpdate Depth Analysis] Anthropic co-founder Olah’s claims of AI "introspection" stem from their pioneering work in mechanistic interpretability. In the AI Agent ecosystem, this internal self-reflection is a critical precursor to autonomous planning and cognitive self-correction. However, the Pope’s philosophical skepticism highlights a fundamental risk in the industry: anthropomorphizing agents obscures their deterministic limitations. While modern agentic frameworks like LangGraph rely heavily on the LLM's emergent "metacognition" to solve complex tasks, these internal states remain unpredictable black boxes. For the Agent ecosystem to scale safely into high-stakes sectors, the industry must transition from abstract corporate "alignment" to rigorous, independent regulatory standards. True reliability will not come from simulated emotions, but from verifiable, auditable control loops and concrete legal frameworks.
