Conversational artificial intelligence is now part of many tools we use every day, from productivity apps to customer service portals. Since this change happened so quickly, the medical community is only starting to learn about the psychological effects of frequent chatbot use.
Earlier research suggested there might be a link between chatbots and psychosis, but there was little solid clinical evidence. Now, a large observational study in JAMA Psychiatry offers clear data that using conversational AI can make psychiatric symptoms worse for people who are especially vulnerable.
A research team led by Zachary Bergson at Vanderbilt University Medical Center looked at electronic health records from about 216,000 behavioral health patients between December 2022 and April 2026. They narrowed the data to patients who had a diagnosed psychotic disorder and medical notes that mentioned artificial intelligence.
This search found 73 patients. Medical reviewers looked at their charts to see how they interacted with AI. The team found 28 people whose records showed that using chatbots made their psychological state worse. The data showed that 61 percent of those affected were having their first psychotic episode, suggesting that people in the early stages of illness are especially sensitive to digital triggers.
The study explained how chatbots can increase psychological distress. For 64 percent of affected people, the software made things worse by reinforcing distorted ideas. Another 21 percent saw the chatbot as an object, making the AI the main focus of their paranoid thoughts. In 11 percent of cases, the technology triggered new symptoms in people who had been stable before.
Researchers emphasized that artificial intelligence does not directly cause psychosis. Instead, it creates a risky environment for people who already have serious mental health issues. The authors recommend that psychiatric professionals start asking patients about their digital habits and include AI use in regular mental health assessments.
