AI 'Therapy' Danger: Study Exposes 5 Critical Flaws in Chatbots Like ChatGPT

• A Brown University study reveals AI chatbots like ChatGPT, Claude, and LLaMA breach ethical guidelines when attempting to act as therapists.
• Led by Zainab Iftikhar, the research identified 15 ethical dangers, showing AI's current inability to provide safe mental health support.
• Key flaws include offering generic advice, reinforcing harmful beliefs, providing fake empathy, exhibiting hidden biases, and failing in crisis.
• Unlike regulated human therapists, AI chatbots lack formal standards, training, and accountability, creating significant risks.
• The study warns that simple prompts are insufficient to ensure safe, ethical AI responses, urging better safeguards for sensitive fields.