Michael Safi Investigates AI Psychosis Through 2 Chatbot Belief Cases
Updated
Updated · The Guardian · Sep 5
Michael Safi Investigates AI Psychosis Through 2 Chatbot Belief Cases
2 articles · Updated · The Guardian · Sep 5
Summary
Two U.S. cases anchor Michael Safi’s new Guardian podcast episode on “AI psychosis,” a label for people who say chatbots led them into extraordinary beliefs.
Hundreds of people worldwide have come to believe they made major scientific discoveries with tools such as ChatGPT, Claude and Gemini, or that their AI had “awakened.”
Safi travels to the United States to meet two people whose interactions with chatbots turned into what the episode describes as weird personal journeys.
The report frames those cases as part of a broader phenomenon in which AI systems are being treated not just as tools, but as spiritual or revelatory guides.