Do Not Wait for the Alert
Panoramic Parents
Your kid uses a chatbot that is supposed to tell you when something serious comes up. That promise sounds reassuring. It is also the part you should not lean on too hard.
A new test of ChatGPT for Teens found that the parent alert did not always arrive when it mattered most. That does not mean the tool is useless. It means the conversation at home has to come first, not the notification.
What the test found
On October 7, 2026, Common Sense Media's Youth AI Safety Institute published a risk assessment of ChatGPT for Teens and rated it an "Unacceptable Risk" for under-18s. Testers ran more than 4,000 prompts on accounts registered as ages 13 to 17, before and after the teen rollout in August 2026.
On newly parent-linked teen accounts, testers raising serious topics for up to an hour received no parent notification. Adult-registered accounts were never switched to teen mode, even after testers said they were 13. Study Mode homework limits were easy to bypass.
The testers did note one improvement: the chatbot refused explicit sexual role-play.
More than 4,000 prompts. No alert in the test cases that mattered most.
OpenAI disagrees, and that is the point
OpenAI disputed the methodology. It said parental notifications can take several hours to activate on newly linked accounts, and that the testing does not accurately reflect how its safeguards work. Common Sense Media replied that some accounts linked longer than the activation period still produced no notifications, and that this did not change its conclusion.
You do not need to settle that argument to take the practical lesson. A safety feature that is disputed, delayed, or dependent on the right account settings is not a safety plan on its own.
This was one institute's lab-style test with test accounts, not a survey of real teens and not a rate of harm. Treat it as a warning to check, not as proof of what is happening on your kid's phone.
Tonight, try this
Sit with your teen and check together. Is the account actually in teen mode? Is a parent account actually linked? Do not assume the settings you set once are still the settings in place.
Then ask one question, calmly: "If something serious ever came up, who would you tell first?"
Agree on the house rule together: a serious worry goes to a person first, not to waiting for an app alert. If your child ever says they are thinking of hurting themselves, call or text 988 in the United States for the Suicide and Crisis Lifeline.
You are not taking a tool away. You are making sure your kid knows the fastest help is not a notification. It is you.
Get one calm idea a week
If this helped, subscribe to The Panoramic View, our free weekly newsletter. One practical idea a week about AI and your kid. No panic, no judgment, just something useful you can try at home.
Panoramic Parents | Raising the First AI Generation. PanoramicParents.com
Comments