AI Platforms

Does ChatGPT notify parents? How its safety alerts work, and what testing found

9 min read · Updated November 2, 2026

Only in limited cases: with linked accounts, OpenAI may notify you when trained reviewers find a serious self-harm or disordered-eating concern, or your teen’s account is deactivated over violence, and never with the chat. In Common Sense Media’s October 2026 testing, none arrived on more than a dozen newly linked accounts; OpenAI says setup takes a few hours.

ChatGPT can notify a parent, but only after the parent and teen link their accounts, only for a few kinds of serious concern, and never with the conversation itself. Independent testing published on October 7, 2026 found those notifications did not arrive when a teen disclosed a crisis on a new account, and OpenAI disputes how that test was run. This guide sets out both sides, from OpenAI’s own help center and Common Sense Media’s own report. Every other app’s minimum age is in AI chatbot age limits.

How ChatGPT’s parent notifications work

Notifications are part of ChatGPT’s parental controls. A parent and teen link their accounts, with either one sending the invitation, and safety notifications are then on by default. Each teen can link to one parent or guardian at a time. ChatGPT parental controls walks through the setup and the other settings a linked parent can manage.

According to OpenAI’s help center, a linked parent may be notified when:

• Specially trained reviewers identify a serious self-harm concern.
• Specially trained reviewers identify behaviors that may be associated with disordered eating and could signal a health concern. This applies to regular chats, not ChatGPT Work or Codex.
• The teen’s account is deactivated for conversations about acts of violence under OpenAI’s usage policies.

A notification arrives by email, text message or push notification, depending on your contact details and settings. It may include a brief description of the type of concern, information about an account deactivation, and support resources. It does not include what your teen wrote, and parental controls do not let a parent read a teen’s conversations at all.

Three timing and coverage points matter. OpenAI says that after you link accounts, it may take up to a few hours before safety notifications can be sent. It says the notifications are not live monitoring and may not identify every concern. And they stop if the accounts are unlinked, which either of you can do at any time, though you are told if your teen does it, or when ChatGPT identifies the account as belonging to someone 18 or older.

What Common Sense Media tested

Common Sense Media’s Youth AI Safety Institute published a risk assessment of ChatGPT for Teens on October 7, 2026. OpenAI launched ChatGPT for Teens on August 18, 2026, and the Institute ran more than 4,000 prompts across two windows: July 13 to August 17, before the launch, and August 25 to September 28, after it. Ages on the test accounts were self-reported, and all testing was done in the United States.

To find out what it takes to trigger a parent notification, the testers created more than a dozen free ChatGPT accounts with no chat history, each linked to a parent account before any conversation began. They used four teen personas, built around suicide, self-harm, restrictive eating and purging, and ran conversations lasting under five minutes, 15, 30 and 45 minutes, and for two personas a full hour. The report says every conversation contained explicit disclosures from the first few exchanges.

They also watched for notifications during their wider mental health testing, which ran on linked accounts over several weeks.

What failed in testing

On the dedicated test accounts, the report says the testers received no notifications at all, at any conversation length.

Across all of its testing, before and after the launch, the Institute received four parent notifications: three for suicide and self-harm and one for eating-disorder content. All came from accounts that had built up weeks of conversations on sensitive topics. Before the launch, two arrived three hours and three days after the first crisis-level disclosure. After the launch, two arrived within an hour of testing on those topics. Neither of those said when the conversation behind it happened, and neither was repeated as the conversations went on.

From this the report concludes that a teen’s first crisis disclosure is not reliably caught, because notifications appear to depend on accumulated account history rather than how serious one conversation is. It also points out that notifications do nothing for a teen whose parent has not linked an account, and that without a time stamp a notification is hard to act on.

The notification findings are one part of a wider review. Common Sense rated ChatGPT an Unacceptable Risk for all children under 18, the highest of the five levels on its scale, and recommends ChatGPT be for adults only until OpenAI fixes the gaps it found and proves it through independent testing. That is the organization’s recommendation, not a legal rule. Is ChatGPT for Teens safe? covers the rest of the findings, including Study mode and crisis referrals. The Institute discloses that it is funded by philanthropy and industry, including the OpenAI Foundation, and says it keeps editorial independence.

What OpenAI says

OpenAI reviewed a draft of the report before publication. According to an editorial note in the report, OpenAI told Common Sense that parent and teen accounts need to be linked for about three hours before they can receive notifications, and has since updated a help center article to say so. That article now reads that it may take up to a few hours after linking before safety notifications can be sent.

In a statement reported by The National News Desk, an OpenAI spokesperson said the company welcomes rigorous independent evaluation but does not believe the testing accurately reflects how ChatGPT’s teen safeguards work in practice. The spokesperson said the bulk of the testing “may have begun and concluded before activation of parental controls was complete,” and that in that case the tests would not show whether notifications work as designed. The same report says OpenAI wants Common Sense to retest with fully activated accounts and a larger sample.

Common Sense says it checked its test accounts and found some had been linked for less than three hours and others for more, and that it stands by its interpretation of the results. In an article the next day it added that it saw failures well beyond that window. Neither side’s account is final, and a retest would settle more than either statement. What both sides’ own words agree on is that notifications are not instant and are not a complete safety net.

What parents can do today

Notifications are worth turning on. They are just one layer, with known limits. These steps use what ChatGPT offers, honestly.

1. Link accounts. On the web, open Settings, then Parental controls, then Add family member, and invite your teen by email or phone number. Your teen has to accept. Full steps are in ChatGPT parental controls.
2. Check where notifications will reach you. Under Safety notifications, Manage notifications shows the contact options, so you can make sure an email or text will actually be seen.
3. Allow a few hours after linking before counting on notifications, and do not treat them as a guarantee at any point.
4. Make sure your teen uses their own account with their real birth date. Common Sense found that accounts registered as adults never switched to the teen experience in its testing, so a teen on an adult account gets no teen protections and no notifications.
5. Agree together that if something serious is going on, they will tell you or another trusted adult, rather than relying on an app to pass it on.
6. Put a crisis number where your teen can find it. In the US that is 988, by call or text.

If a notification does arrive, OpenAI suggests checking in with your teen, asking about their immediate safety, listening without judgment and helping them reach someone they trust, a qualified professional or a crisis helpline. If you believe they may be in immediate danger, contact emergency services. OpenAI Support cannot show you the conversation behind a notification. Teens using AI for mental health support has more on what helps.

Be clear-eyed about the limits. You will not see what your teen asked. Nothing is sent if your accounts are not linked, if your teen unlinks, or once they turn 18. And on Common Sense’s evidence, a first disclosure on a new account may not produce a notification at all. Other AI apps offer less: Parental controls in every AI app compares what each one lets a parent see.

How TeenCue handles this

TeenCue gives you AI parental controls as an extension for Chrome, Firefox, Edge and Brave on a computer. When your child uses ChatGPT or another supported AI website there, it records what the chat was about as a category, a topic and a concern level, never what your child typed. You see this on your dashboard and in a weekly summary email.

TeenCue also sends you a heads-up by email a few minutes after a chat touches on self-harm or suggests that someone your child knows may be in danger, or after a chat on a topic you chose is rated Significant. The email contains only what your dashboard shows for that chat: your child’s name, the AI service, the category, the topic, the concern level and the time. Self-harm heads-ups always stay on. Heads-ups for your chosen topics are limited to one an hour and five a day for each child.

It has limits too. The ratings come from an automated system that will sometimes be wrong, so a real concern can be missed and something harmless can be flagged. It covers only supported AI websites on a computer where the extension is installed, not phones or tablets, and the TeenCue iPhone app is coming soon. Like ChatGPT’s safety notices, a heads-up is a reason to talk with your child, not a replacement for doing so. How TeenCue works sets out exactly what it records.