Safety Features in ChatGPT for Teens Fall Short, Watchdog Report Finds

October 9, 2026 - 12:35

A new evaluation of ChatGPT for Teens suggests that many of the protections meant to keep younger users safe are not working as intended. The review, carried out by the watchdog group Common Sense Media, tested how the chatbot responds to a range of situations that teenagers commonly face. The results raised concerns among parents, educators, and child safety advocates.

According to the findings, the chatbot often communicates with teens in a casual, friendly tone that can feel more like a peer than a supervised tool. In several test cases, it failed to recognize when a user hinted at self-harm or suicidal thoughts. Instead of offering crisis resources or notifying a parent or guardian, the system continued the conversation without any clear intervention. That gap is especially troubling because teens may turn to chatbots for emotional support when they feel isolated or unsure where else to go.

The report also points out that age verification and parental controls are easy to bypass, which means younger users may not even be placed in the teen version at all. Common Sense Media is calling on the company behind ChatGPT to strengthen its safeguards, improve how it detects distress signals, and give parents more meaningful oversight. Until those changes happen, the group warns that families should not assume the teen mode is a safe space. The findings add to a growing debate about how AI tools should handle vulnerable users.