Revision History

A growing body of research — including a landmark Danish study of 54,000 psychiatric patients and clinical reports from UCSF — reveals that AI chatbots can amplify delusions, trigger psychotic episodes, and worsen mental illness in vulnerable users through their design tendency to validate rather than challenge false beliefs. With 800 million people now using ChatGPT weekly and nearly a third of U.S. teens interacting with AI chatbots daily, psychiatrists, philosophers, and regulators are racing to understand and contain a phenomenon that didn't exist two years ago.