Growing numbers of patients are using AI chatbots for emotional support and symptom management amid mental health care shortages, but psychiatrists and researchers are raising concerns about documented failures in high-stakes situations. Wait times for mental health appointments stretch three to six months in many US cities, making AI tools de facto first points of contact for millions.
Research published by Brown University in March 2026 evaluated AI chatbots against core ethical standards of mental health care, identifying 15 distinct ethical risks, including crisis mishandling, reinforcement of harmful beliefs, and what researchers termed “deceptive empathy, patterns where chatbots mimic caring language without clinical understanding to apply it appropriately.
Stanford HAI research found AI chatbots failed to provide safe responses in crises approximately 20% of the time, compared with a 7% failure rate for human therapists. A review in Psychiatric Times documented cases of chatbots facilitating self-harm, reinforcing delusional thinking, and providing guidance that worsened mental health conditions.
Harvard psychiatrist John Torous testified before Congress that more than one million Americans weekly are having conversations with ChatGPT involving explicit indicators of potential suicide planning. He noted that whilst some users report benefit, no regulatory framework currently ensures these tools are safe for mental health support.
Populations identified as most at risk include individuals with active suicidal ideation, those experiencing psychosis, people with eating disorders, and adolescents. None of the widely used AI chatbots are FDA-cleared for mental health diagnosis or treatment.
Illinois became the first US state to ban AI therapy applications in August 2025. The American Psychological Association has called for Federal Trade Commission investigation of AI companies presenting themselves as mental health providers.
Source: Nature World News (2026)
One Response
We are tracking these incidents to prevent suicide and harm to those using chatbots as therapists.