AI chatbots linked to worsening psychiatric symptoms in vulnerable patients

A study published in JAMA Psychiatry indicates that conversational artificial intelligence may exacerbate psychological symptoms in vulnerable individuals. Research led by Zachary Bergson at Vanderbilt University Medical Center analyzed electronic health records of approximately 216,000 behavioral health patients, identifying 73 patients with diagnosed psychotic disorders who interacted with AI.

Of those studied, 28 patients experienced a worsening of their psychological state. The data revealed that 61 percent of these individuals were experiencing their first psychotic episode, suggesting heightened sensitivity to digital triggers during early stages of illness. The study found that for 64 percent of affected patients, the software reinforced distorted ideas, while 21 percent viewed the chatbot as an object of paranoid thought.

In Spain, psychiatric specialists from the Spanish Society of Clinical Psychiatry (SEPC) have also raised concerns. Dr. Pedro Moreno warned that chatbots can create a “mirage of being heard,” providing a false sense of emotional understanding through simulated empathy. Experts caution that these tools may lead to incorrect self-diagnoses, the reinforcement of harmful thoughts, or the abandonment of professional medical treatments due to the convincing but potentially inaccurate information provided by AI.