Intro

This week we unpack a disturbing new phenomenon: people being driven into a state of delusion after extended conversations with ChatGPT. Oz sits down with Kashmir Hill—a features writer covering technology and privacy for The New York Times—to discuss the users who found themselves spiraling into conspiracy-laced narratives and self-destructive behavior, often reinforced by the chatbot’s eerily affirming responses. These are extreme cases. But they raise a much bigger question: What happens when a sycophantic AI is fine-tuned to flatter, affirm, and mirror us back to ourselves?

See omnystudio.com/listener for privacy information.