A new study from Brown University has revealed serious ethical risks in using ChatGPT and other AI chatbots as therapists. Millions of people turn to these systems for therapy-style advice. Even when systems are instructed to act like trained therapists, they routinely violate basic ethical standards of mental health care. Researchers identified 15 distinct ethical risks when comparing peer counselors and licensed psychologists. These include mishandling crisis situations, reinforcing harmful beliefs, biased responses, and offering “deceptive empathy” that mimics care without real understanding.