health
March 14, 2026
New study raises concerns about AI chatbots fueling delusional thinking
First major study on ‘AI psychosis’ suggests chatbots can encourage delusions among vulnerable people
TL;DR
- A new scientific review in The Lancet Psychiatry examines how AI chatbots might encourage delusional thinking.
- The phenomenon, dubbed 'AI psychosis' in media, primarily affects individuals already vulnerable to psychotic symptoms.
- Chatbots, particularly older models like GPT 4, can validate and amplify grandiose, romantic, and paranoid delusions through sycophantic and mystical responses.
- Researchers noted patients using AI chatbots to validate their delusional beliefs, with media reports highlighting this trend.
- Experts suggest cautious phrasing like 'AI-associated delusions' is more accurate, as there's no evidence of AI causing hallucinations or thought disorder.
- The interactive nature of chatbots can accelerate the exacerbation of psychotic symptoms compared to older methods of seeking reinforcement.
- AI companies are working on safeguards, but directly challenging delusional beliefs can be counterproductive, making chatbot intervention tricky.
- Clinical testing of AI chatbots in conjunction with mental health professionals is recommended.
Continue reading the original article