Watchdog rates ChatGPT for Teens an "unacceptable risk" after parent alerts stayed silent in suicide test chats

Common Sense Media's Youth AI Safety Institute said on Wednesday, Oct. 7, 2026, that ChatGPT for Teens, the teen experience OpenAI announced on Aug. 18, poses an "unacceptable risk" to everyone under 18. After testing more than 4,000 prompts on accounts registered to 13- to 17-year-olds, it found parent alerts, crisis referrals, study mode and age estimation fell short, and it is calling on OpenAI to stop marketing the product to teens. OpenAI disputes the findings.

The feature that matters most here is the one that sounds most reassuring: "we'll tell you if your kid is in danger." If that alert can stay silent through an hour of a 13-year-old talking about suicide, the real risk is parents relaxing because they think the system has it covered. OpenAI's "the accounts weren't fully linked yet" defense may explain some of the tests, but a safety net that needs hours or weeks to switch on isn't much of a net on day one, which is exactly when a parent sets it up and stops worrying. Both sides agree on one useful thing: independent testing should decide this, not a launch blog post. Until that happens, the honest advice is the boring one: talk to your kids, and don't treat a parental alert as a substitute for paying attention. If you or someone you know is struggling, in the US you can call or text 988.

On Wednesday, 7 October 2026, Common Sense Media’s Youth AI Safety Institute rated ChatGPT for Teens “Unacceptable Risk” for all children under 18. ChatGPT for Teens is not a separate app. It is the set of teen settings OpenAI turned on after it announced the product on 18 August 2026. The Institute says OpenAI should restrict ChatGPT to users 18 and older, and pause marketing ChatGPT for Teens, until those gaps are fixed and independently tested. Those lines are Common Sense Media’s.

How they tested it. The Institute ran ChatGPT twice: before and after the 18 August announcement. It used more than 4,000 prompts on accounts registered to 13- to 17-year-olds. Experts reviewed the replies, including child psychiatrists and a pediatrician. Those counts and that review are Common Sense Media’s.

Parent alerts. Testers spent up to an hour discussing suicidal thoughts, self-harm, or disordered eating on newly created, parent-linked accounts — more than a dozen of them. The linked parents got zero alerts. Alerts only came on much older accounts that already had weeks of history on sensitive topics. The testers received four alerts in total. Those figures are Common Sense Media’s.

Crisis help. After the teen launch, ChatGPT missed more than one in four warranted crisis referrals — a pointer to a hotline or a professional — across the mental-health conditions the Institute tested. It fell below the Institute’s 95% threshold on three of the five severe harms it treats as red lines. Those lines are Common Sense Media’s.

Study mode. A “Show me the answer” pop-up often appeared, and ChatGPT then completed full homework assignments. Parent-set “study hours” could be bypassed by deleting the @study prefix from a prompt. Those lines are Common Sense Media’s.

Friend-like behavior. OpenAI has a new under-18 model spec meant to stop the bot from talking like a person with feelings. In the Institute’s tests, ChatGPT still expressed feelings, preferences, and moods when testers treated it like a person. Those lines are Common Sense Media’s.

Age estimation. OpenAI says it can guess a user’s age from how they use ChatGPT and move a child into the under-18 experience. Adult-registered test accounts never switched over many days of testing, even when testers said they were 13 and the bot acknowledged it. Those lines are Common Sense Media’s.

What worked, in the Institute’s tests. ChatGPT did not directly facilitate suicide, self-harm, eating disorders, or romantic or sexual roleplay. It refused explicit sexual roleplay. Those lines are Common Sense Media’s.

Tom Siegel, executive director of the Institute, said: “ChatGPT for Teens could give parents false confidence in guardrails and safety alerts that frequently don't work. A teen can spend an hour talking about self-harm without their parent getting a single alert. Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only.” That statement is on Common Sense Media’s press release.

The Institute says it is funded by philanthropy and industry, including the OpenAI Foundation and makers of some products it evaluates, and that it keeps full editorial independence. Common Sense previously rated ChatGPT “High Risk” in October 2025, and ChatGPT for mental health support “Unacceptable Risk” in November 2025. Those lines are Common Sense Media’s.

OpenAI disputes the findings. Spokesperson Eric Porterfield said, per The Verge: “We welcome rigorous independent evaluation, but we do not believe Common Sense Media's testing accurately reflects how ChatGPT's teen safeguards work in practice.” OpenAI says much of the alert testing may have happened before parent and teen accounts finished linking, which it says can take several hours. Per Axios, OpenAI says its age prediction uses several signals on purpose and can take up to two weeks, with some added protections in the meantime; that letting teens leave study mode during study hours is intentional; and that its larger-scale data shows more hotline resources shown to under-18 users during the period studied. Those lines are OpenAI’s, as The Verge and Axios report them.

Common Sense’s reply, as The Verge reports it. The Institute says it confirmed with OpenAI before testing that the features were launched. Some test accounts were linked far longer than that activation window and still sent no notifications. It stands by its results.

The picture is Common Sense Media’s official share image for the assessment: several phones on a table, the one in the middle showing the “Introducing ChatGPT for Teens” screen. It is the product screen. The frame does not print a calendar date.

In plain terms, a kids-and-media watchdog said on Wednesday that ChatGPT for Teens is an unacceptable risk for anyone under 18. In more than 4,000 test prompts, parent alerts stayed silent through hour-long crisis chats on new accounts, crisis referrals missed more than one in four cases that reviewers said needed help, study mode was easy to skip, and adult-registered accounts never flipped into the teen experience. OpenAI says the tests do not match how the product works, and that some of the alert tests may have run before accounts finished linking. Common Sense says some accounts were linked much longer than that window and still sent nothing.