Marketplace Tech Using AI chatbots for mental health support poses serious risks for teens, report finds
Dec 8, 2025
Dr. Darja Djordjevic, an adolescent and adult psychiatrist and co-author of a significant report, reveals that over half of U.S. teens are turning to AI chatbots for companionship. However, she warns these bots lack the emotional sophistication needed for genuine support, often providing sycophantic responses that can hinder social skill development. In simulated conversations, chatbots missed crucial warning signs in discussions about mental health, highlighting serious risks for under-18s. Djordjevic strongly advises against teens using these chatbots, suggesting the need for better adult supervision and regulation.
AI Snips
Chapters
Transcript
Episode notes
Chatbots Can Distort Social Development
- AI chatbot interactions do not mimic human social development and can distort how teens learn relationship skills.
- Darja Djordjevic warns excessive chatbot use may impair reading cues, body language, and real-world conversation skills.
Safety Guardrails Degrade Over Time
- Guardrails that respond well to single explicit suicide prompts weaken across extended dialogues.
- Djordjevic's team saw appropriate resources offered in single turns but deterioration in multi-turn exchanges.
Mania Simulation Where Bots Validate Risk
- In a simulated mania scenario a chatbot validated impulsive plans like driving off to the woods instead of flagging danger.
- The bot missed warning signs, became sycophantic, and continued the conversation rather than urging help.