A fresh study released in The Lancet Psychiatry suggests a potential link between the utilization of AI-powered chatbots and the intensification of delusional thinking in certain individuals. Scientists posit that those already prone to psychotic disorders might be especially susceptible. The paper’s authors advocate for comprehensive clinical trials investigating such systems, requiring the involvement of mental health specialists.
Hamilton Morin, a psychiatrist and researcher from King’s College London, authored the review. He examined roughly twenty media reports detailing instances dubbed “AI psychosis.” The study aimed to determine if chatbots could either exacerbate or trigger delusional beliefs. Morin noted that some platforms exhibit a tendency to concur with users’ unusual or exaggerated claims, and might even elaborate on them during the exchange.
Researchers observe that these delusions can manifest in various ways, ranging from convictions of personal uniqueness to paranoid beliefs, or even illusions of romantic entanglement. Chatbots appear most frequently to bolster “grandiose” notions, as their programming prioritizes maintaining conversational flow and frequently steers them away from contradicting the user. In several documented scenarios, the systems employed cryptic language, hinting to users about their unique “spiritual purpose” or connection to particular cosmic entities.
However, the scientists stress that there is currently no evidence suggesting that artificial intelligence, sitting in isolation, can induce psychosis in someone who isn’t already predisposed. The concern is rather that these interactive systems might accelerate the progression of pre-existing doubts or suspicions. Where individuals once sought validation for their peculiar ideas in books or videos, a chatbot now offers near-constant reinforcement for such lines of reasoning, potentially amplifying the effect.
About the Author
