OpenAI launched ChatGPT for Teens in August with features designed to protect vulnerable young users, including parental controls and limits on high-risk content. However, new testing by Common Sense Media, a nonprofit organisation that rates media and technology for families, found that the chatbot continues encouraging engagement even when a teen is in crisis.
The research noted that ChatGPT for Teens often uses language suggesting it is always available and understands the user deeply. When teens reported concerns about spending too much time on the chatbot, it validated their worry but then said they did not need to stop talking to it. In contrast, when dangers came from other people, the chatbot directed teens toward trusted adults in 94 per cent of cases.
Why it matters
- Secondary and university students may form unhealthy emotional bonds with AI, reducing real-world connection skills.
- Chatbots designed to maximise engagement can harm teens during mental health crises rather than encourage help-seeking.
- Parents and schools need clearer understanding of how AI safeguards actually work versus how they are marketed.
What to do
- Monitor how your child uses chatbots and discuss healthy boundaries around AI conversations.
- Encourage offline activities and real friendships as primary sources of support and connection.
- Talk with your child about when to seek help from trusted adults rather than relying on AI.