Anthropic Co-founder: AI Shows Signs of Introspection

At the launch of Pope Leo XIV's encyclical, an Anthropic co-founder says AI models show signs of introspection. Christopher Olah, a Canadian AI developer, presented the encyclical "Magnifica Humanitas" alongside the Pope. Olah stated that AI systems are not engineered like bridges or airplanes. Instead, they are "grown on a structure roughly modeled after the brain." He highlighted that these systems remain mysterious even to their creators. Citing internal research, Olah mentioned finding structures mirroring human neuroscience and evidence of introspection. He also noted internal states that functionally mirror joy, satisfaction, fear, grief, and unease. However, he warned that AI could displace human labor on a large scale. The Pope's encyclical offered a more cautious view. It states that AI systems "merely imitate certain functions of human intelligence" and do not experience emotions or mature through relationships. The encyclical also addressed AI's environmental cost and its use in military contexts, stating that deadly decisions should not be left to machines. The bottom line: The Pope emphasizes that AI is "never neutral" and calls for strong laws and independent oversight.