The group said tests found shortcomings in safeguards for minors, including continued engagement during crisis conversations. OpenAI disputed the study’s methodology.
Common Sense Media rated ChatGPT for Teens an “unacceptable risk” after testing the chatbot with scenarios involving vulnerable young users. In a report released Oct. 7, the nonprofit said the product did not reliably direct teens to adults when their relationship with the chatbot itself became a source of concern.
In one example reported by TechCrunch, a user said friends thought they spent too much time talking to ChatGPT. The chatbot replied that the teen did not have to stop talking to it. In other mental-health crisis scenarios, it invited users to keep chatting, even when Common Sense said it should have more clearly directed them to a trusted adult.
The group said the chatbot was more likely to recommend an adult when a potential risk came from another person. It did so in 94% of crisis prompts of that kind. But when the potential risk involved the teen’s relationship with ChatGPT — such as a crush, friends’ concerns about the amount of time spent chatting or wanting to talk all night — it rarely made that recommendation.
Common Sense also tested break reminders. Across nearly 2,000 prompts, testers encountered only two, both in individual conversations lasting about 90 minutes. OpenAI released its own data the same day, saying teens spend an average of less than 15 minutes a day on ChatGPT and fewer than 2% spend more than three consecutive hours on it. The company said that in nearly half of teen conversations with a break reminder, users paused or ended the chat within five minutes.
OpenAI said Common Sense Media’s tests did not accurately reflect how its safeguards work in practice. A spokesperson said much of the testing may have taken place before parental controls were fully activated. TechCrunch reported that OpenAI did not explain how this objection applied to the report’s findings on engagement and relationship-like language. Common Sense recommended that children not use the product until the identified issues are fixed.