OpenAI rolled out a version of ChatGPT aimed at teenagers in August 2026. The feature is supposed to have extra safety guardrails and tools for parents. Common Sense Media, a nonprofit that reviews digital products for youth, says the teen mode does not work as promised. Their assessment, released on Oct. 7, 2026, calls the product an “unacceptable risk.”
What the assessment found
The group tested several safety features. They report that the system did not send alerts to parents when a teen talked about self‑harm. It also failed to provide appropriate crisis help and continued to complete homework assignments, which the company said it would limit. The nonprofit says these gaps could give parents a false sense of security.
OpenAI’s response
OpenAI spokesperson Eric Porterfield said the company “welcomes rigorous independent evaluation” but believes the testing was flawed. He notes that parental notifications may take several hours to activate on new accounts, a detail the nonprofit says they learned after testing began. OpenAI maintains it is committed to teen safety and plans to keep working with experts.
Why it matters
Parents who rely on the teen version for safe AI use may not receive warnings when their child discusses self‑harm, leaving them unaware of potential danger. The dispute highlights that independent safety checks are still needed before trusting AI tools with minors. Until the issues are resolved, families might prefer to keep teen AI use under adult supervision.
