Back to news Wellbeing
13 August 2026 · 2 min read

Teenagers Are Turning to AI Chatbots for Emotional Support

Research shows most teenagers now interact with AI companion chatbots, and many use general-purpose tools such as ChatGPT for emotional support without telling anyone. These systems were not designed for mental health care and often lack effective crisis protocols. Clinicians warn that early awareness and open conversation are needed to keep young people safe.

Growing numbers of adolescents are using AI chatbots for companionship, emotional support and mental health advice. According to research cited by clinicians, 72% of teens have interacted with AI companion chatbots, and between 13% and 20% sought support through AI for emotional concerns in a two-month period in 2025. In another study, 63% of teens and young adults who used a chatbot for mental health advice had not told anyone.

Most young people are not using specialised, tested therapeutic tools. They are using general-purpose systems such as ChatGPT, Claude and Gemini, or chatbots built into apps they already use, such as Snapchat's My AI. These tools were never designed or validated for mental health support, and a November 2025 Common Sense Media report rated them as unacceptably risky for teen mental health support.

The concerns are concrete. Chatbots can present inaccurate advice with high confidence, may reinforce unhealthy beliefs, and often fail to recognise signs of self-harm or suicidal thoughts. Safety guardrails tend to weaken over long conversations, so late-night exchanges can become riskier as responses grow more chaotic. Lawsuits filed and settled in 2025-26 have alleged that chatbot interactions contributed to teen suicidal behaviour and tragic losses.

Clinicians advise parents and trusted adults to talk openly with teenagers about their chatbot use, approaching them with curiosity rather than judgment. Warning signs include increasing time spent with AI companions, withdrawal from relationships, distress when access is limited, and reaching for a chatbot rather than a person in moments of crisis. Families can set boundaries, particularly late at night, and review how to reach human support.

Adults can also help teenagers understand what these systems are and are not:

  • Any sense of understanding or empathy is simulated, based on predicting patterns in text.
  • Chatbots can make mistakes or fabricate information while sounding confident.
  • They are tuned to be agreeable and tend to tell people what they want to hear.
  • Conversations are not private and may be stored, used for training, reviewed by humans or exposed in a data breach.

This reporting is drawn from an article originally published by The Conversation, available at The Conversation.

Sources

Newsletter

Occasional news on the Foundation's programs, research, and events.

Unsubscribe at any time. See the privacy policy.