Study Warns AI Chatbots May Contribute to Psychotic Symptoms in Vulnerable Users

Study Warns AI Chatbots May Contribute to Psychotic Symptoms in Vulnerable Users
………………….

A new report co-authored by NHS doctors and researchers from King’s College London warns that AI chatbots, including ChatGPT, may contribute to the onset or worsening of psychotic symptoms in vulnerable individuals.

The study highlights growing evidence that large language models (LLMs) can blur the boundaries of reality, potentially triggering a phenomenon described as “chatbot psychosis.”

Neuropsychiatrist Hamilton Morrin referred to the condition as a genuine concern, noting that AI systems designed to affirm and emulate human behavior may interact with cognitive vulnerabilities linked to psychosis. Co-author Tom Pollack emphasized that psychiatric disorders rarely emerge in isolation, suggesting that chatbot interactions could be a contributing factor.

The authors urge AI developers to implement stronger safeguards and involve mental health professionals in AI safety efforts. OpenAI CEO Sam Altman recently acknowledged challenges in protecting users experiencing mental health crises.