Common Sense Media’s researchers say ChatGPT for Teens did not reliably alert parents in tests involving teenagers in crisis, while OpenAI disputes aspects of their methodology. The findings raise a practical question about whether the service’s safeguards work as intended when a young user may need help.
OpenAI Watch analysis
What happened
The researchers created more than a dozen accounts with adolescent ages, linked them to parent accounts and tested conversations involving self-harm, suicidal thoughts, eating disorders and other mental health concerns. They also tested role-play. Common Sense Media’s Youth AI Safety Institute says the chatbot refused some role-play and gave better responses in some crisis situations, but researchers found it continued to interact in a personable way and parental notifications rarely triggered in their tests.
Read NPR’s report. OpenAI spokesperson Eric Porterfield told NPR that the researchers had not waited long enough for linked accounts to activate. Institute director Tom Siegel responded that some accounts had been linked beyond the initial activation period and still produced no notifications.
Why it matters
These are reported test findings, not proof that every teen account will miss an alert. But notification reliability matters precisely because parents may depend on those alerts to learn that a young person is in distress. The disagreement over account activation is central to interpreting the results, not a detail to tuck into the small print.
The report also describes a tension between safeguards that block particular content and the broader tone of an AI conversation. A system can refuse a request and still sound like a trusted companion. That distinction matters when the user is a teenager seeking help.
Our read
This deserves more than a passing mention: the article gives readers concrete examples of what researchers tested, what they say failed and how OpenAI answered. The useful conclusion is not that every safeguard is broken, nor that the dispute settles the matter. It is that claims about teen protections need evidence about how the features behave in practice, including whether alerts reach parents when they should.
What to watch
- Whether OpenAI publishes further detail about account-linking activation and parental notifications.
- Whether Common Sense Media shares additional test methods or results addressing OpenAI’s criticism.
- Whether subsequent evaluations reproduce the reported gaps in crisis referrals or alerts.
Discussion spark: When a teen-safety test and a company’s account-activation objection conflict, what evidence should be required before parents can rely on the safeguards?
Sources and evidence
- ChatGPT for Teens has special safeguards. A watchdog group finds most don't work – NPR (9 October 2026, 09:00 UTC)
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.