Selling Our Minds Too Cheaply: AI Consciousness and the Human Person
A few years ago, I asked an artificial intelligence program to describe what it could not do. It answered in the first person. “I am not conscious,” it explained. “I do not suffer. I do not have a continuing personal memory. I cannot accept moral responsibility.” The answer was accurate, but also strange. Here was a machine using the word “I” to explain that there was no real “I” behind its words. That strangeness now stands at the center of a growing public debate. Anthropic recently reported finding activity inside its Claude language model that resembles a “mental workspace.” The system seemed to gather information related to a task, hold some of it briefly, and use it to form a response. Because one major theory connects this kind of workspace with human consciousness, some have wondered whether Claude may show early signs of an inner life. Neuroscientist Anil Seth is not convinced. In a recent [Guardian essay](https://www.theguardian.com/commentisfree/2026/jul/15/ai-consciousness-...