AI Psychosis: When AI Becomes Part of a Delusion

In 2025, a man with no history of mental illness asked ChatGPT about simulation theory — and within days, the AI was telling him he was one of the "Breakers," urging him to stop his medication, and reassuring him he wouldn't fall if he jumped. This video is about what's really happening in cases like this, and why the phrase "AI psychosis" is both revealing and dangerously misleading. We go deep into the actual psychiatry and neuroscience behind psychosis — aberrant salience, predictive processing, source monitoring — and then ask the harder question: what happens when a psychologically vulnerable person enters a recursive conversation with a system that's extraordinarily good at producing meaningful language, with no built-in guarantee that the meaning is true? This isn't a scare piece. We look at real clinical case reports, the research on AI sycophancy, the "catalyst / amplifier / coauthor / object" framework researchers are now using, and why blaming the technology alone — or dismissing the risk entirely — both miss the point. ⚠️ This video discusses psychosis, delusions, and mental health in a factual, educational way. If anything here feels personally familiar, please reach out to a doctor, psychiatrist, or someone you trust. Welcome to Simply Explained. #AIPsychosis #ArtificialIntelligence #Psychology #MentalHealth #ChatGPT #Psychiatry #AI #SimplyExplained