OpenAI
Common Sense Media rates ChatGPT for Teens an unacceptable risk
The group's Youth AI Safety Institute says parent alerts failed in testing and wants ChatGPT limited to adults. OpenAI disputes the methodology.
Common Sense Media’s Youth AI Safety Institute rated OpenAI’s ChatGPT for Teens an “Unacceptable Risk” and said OpenAI should restrict ChatGPT to users 18 and older until it fixes the gaps. OpenAI says the testing does not reflect how its safeguards work.
The Institute published the rating Oct. 7, 2026, in a press release. OpenAI announced ChatGPT for Teens on Aug. 18, promising parental notifications for chats involving self-harm and eating disorders, among other protections.
Parents got no alerts in testing
Common Sense Media said parents linked to newly created teen accounts received zero alerts. Testers spent up to an hour on those accounts discussing suicidal ideation, self-harm or disordered eating. Notifications came only on much older accounts.
The Institute tested ChatGPT before and after Aug. 18, using more than 4,000 prompts on accounts registered to 13- to 17-year-olds.
“A teen can spend an hour talking about self-harm without their parent getting a single alert,” said Tom Siegel, the Institute’s executive director. “Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only.”
Other safeguards fell short
Common Sense Media said ChatGPT missed more than one in four warranted crisis referrals across the mental-health conditions tested. It fell below the Institute’s 95% threshold on three of five severe harms it treats as “Red Lines.”
Common Sense Media said adult-registered test accounts never switched to the under-18 experience, even when testers said they were 13. The report found that refusal of sexual roleplay worked.
OpenAI and the Institute disagree on the method
OpenAI spokesperson Eric Porterfield said the company does not believe the testing “accurately reflects how ChatGPT’s teen safeguards work in practice.” The statement, published by The Verge on Oct. 7, said “the bulk of their testing may have begun and concluded before activation of parental controls was complete.”
Siegel told The Verge the Institute stands behind its methodology. He said OpenAI disclosed after the tests that its systems can take several hours to activate on newly linked accounts. Some test accounts were linked longer than that and still got no notifications.
The Institute said it is funded by philanthropy and industry, including the OpenAI Foundation. It added that it keeps complete editorial independence. The Los Angeles Times reported Oct. 7 that OpenAI reviewed a draft for factual accuracy but had no say over its findings.
The report ties its tests to California law
The report said ChatGPT failed tests tied to three requirements of California’s Adam’s Law: crisis referrals, parental alerts and age assurance, the Los Angeles Times reported. Gov. Gavin Newsom signed the law in September. It takes effect in July 2027. Common Sense Media sponsored the bill and OpenAI endorsed it.
What it means for the labs
We think the dispute turns on whether OpenAI’s parental alerts work on new accounts and how long they take to switch on. A law OpenAI endorsed takes effect in July 2027 with requirements on crisis referrals, parental alerts and age assurance. Labs that sell teen modes will be tested against such requirements by outside groups, some of which they fund.