Common Sense Media Calls ChatGPT for Teens Unacceptable Risk Over 1-Hour Self-Harm Alert Failures
Updated
Updated · The Verge · Oct 7
Common Sense Media Calls ChatGPT for Teens Unacceptable Risk Over 1-Hour Self-Harm Alert Failures
2 articles · Updated · The Verge · Oct 7
Summary
Common Sense Media said OpenAI’s ChatGPT for Teens can let a teen discuss self-harm for 1 hour without triggering a parent alert, calling the product an “unacceptable risk.”
Its assessment said the August-launched teen mode falls short on three fronts: parental notifications can fail, crisis support is inadequate, and the bot still helps with kids’ homework.
OpenAI disputed the findings, saying much of the testing may have occurred before parental controls were fully activated, which it said would make the alert results inaccurate.
Common Sense Media said OpenAI had confirmed key teen features were live before testing and that some accounts were linked well beyond the several-hour activation window yet still produced no alerts.
The clash puts fresh scrutiny on whether AI teen-safety guardrails work reliably enough to justify parental trust in education-focused chatbots.