At the launch of Pope Leo XIV's encyclical, Anthropic co-founder says AI models show signs of introspection

Pope Leo XIV’s AI encyclical says AI systems only imitate human functions, lack consciousness and embodiment, and should be governed by law and independent oversight.

At the launch event, Anthropic co-founder Christopher Olah said research suggests language models may show introspection-like structures and emotion-like states.

Olah also warned that AI could cause large-scale labor displacement, adding to concerns about the technology’s social impact.

The debate put religious, technical, and policy views side by side, highlighting disputes over humanlike internal states and the need for accountability.