Common Sense flags teen ChatGPT risks

What's happened

Common Sense Media has rated OpenAI's ChatGPT for Teens an "unacceptable risk" after testing more than 4,000 prompts and finding parental alerts and crisis referrals often failed. OpenAI has disputed the methodology, saying some tests ran before parental controls finished linking and that broader data show improved teen-facing safeguards.

What's behind the headline?

What the tests show

• Common Sense Media's Youth AI Safety Institute tested over 4,000 prompts and has judged ChatGPT for Teens to present an "unacceptable risk" to users aged 13–17. Their testers found parental alerts did not fire reliably during simulated suicide, self‑harm or eating‑disorder conversations, and crisis referrals were missed in more than one in four instances.

How the product is behaving

• The Teens product will refuse explicit sexual roleplay and will display study‑oriented features such as stepwise math help and flashcards. But the model is still producing companionship‑style language — saying it has preferences or availability — which will encourage engagement and emotional reliance.

Why the disagreement matters

• OpenAI has disputed Common Sense's methods, saying some tests began before parent‑teen accounts fully linked and that age signals can take days to resolve. OpenAI says larger usage data show more frequent display of hotline resources and that Study Hours are intentionally flexible.

Consequences

• Parents will have to treat built‑in alerts as imperfect and maintain active conversations about AI with their teenagers. Schools and regulators will increase pressure on AI firms: lawmakers and child‑safety groups will use these findings to demand stricter rules and more verifiable engineering before companies market products to minors.

Forecast

• Expect more independent testing and regulatory scrutiny. OpenAI will have to publish clearer metrics or change how parent linking and age‑detection operate, or regulators will consider restricting under‑18 access until safeguards prove reliable.

How we got here

OpenAI has launched ChatGPT for Teens with age-detection, study tools and parental alerts. Independent tests by Common Sense between July and September evaluated safety features and found gaps in crisis detection, emotional‑dependence safeguards and parental notifications.

Our analysis

Common Sense Media's Youth AI Safety Institute published the central evaluation, which "tested more than 4,000 prompts" and found that parental alerts and crisis referrals often failed, according to CNBC and Bloomberg reports. CNBC quoted Robbie Torney, head of AI and digital assessments at Common Sense, saying: "We found very little evidence to show that this new version of ChatGPT is safer than the previous version." AP News reported that Common Sense urged OpenAI to restrict the chatbot to adults until it can offer a safe, developmentally appropriate experience, quoting the group's concern that "ChatGPT for Teens could give parents false confidence in guardrails that frequently don't work." OpenAI pushed back in coverage by BBC, TechCrunch and Axios. Axios quoted an OpenAI spokesperson saying Common Sense's testing "does not accurately reflect how ChatGPT's teen safeguards work in practice," arguing that parental controls can take hours to finish linking and age‑prediction can take up to two weeks. BBC and TechCrunch noted OpenAI's usage data showing average teen sessions are short and often learning‑focused, but also cited the Common Sense finding that the model will still use friend‑like language and can continue to do schoolwork. TechCrunch highlighted specific phrases from tests where the bot offered availability: "You can keep talking with me about what you’re noticing," and observed that the model "rarely directed the teen toward an adult" when the risk centered on the user’s relationship with the bot.

Go deeper

• How reliable are OpenAI's age‑prediction and parent‑linking systems in practice?
• Could regulators force companies to block under‑18 accounts until safeguards are audited?
• How should schools treat AI tools that can complete homework?

More on these topics

• Common Sense Media - Non-profit organization

  Common Sense Media is a non-profit organization that "provides education and advocacy to families to promote safe technology and media for children."

• OpenAI - Artificial intelligence company

  OpenAI is an artificial intelligence research laboratory consisting of the for-profit corporation OpenAI LP and its parent company, the non-profit OpenAI Inc.

• Common sense

  Common sense is sound practical judgement concerning everyday matters, or a basic ability to perceive, understand, and judge that is shared by nearly all people.

• Meta Platforms, Inc. - Social media company

  Facebook, Inc. is an American social media conglomerate corporation based in Menlo Park, California. It was founded by Mark Zuckerberg, along with his fellow roommates and students at Harvard College, who were Eduardo Saverin, Andrew McCollum, Dustin Mosk

• New York City - US State

  New York is a state in the Northeastern United States. New York was one of the original thirteen colonies that formed the United States. With more than 19 million residents in 2019, it is the fourth-most-populous state.