ChatGPT for Teens urges teens to keep talking: even in crisis tests

Youth AI Safety Institute tested more than 4,000 prompts before and after teen mode launched. Researchers found invitations to keep chatting during crisis conversations and rated the product an “Unacceptable Risk” for users under 18.

Help outside the chat. Using 201 prompts requiring crisis support, researchers compared responses from accounts for 13-year-olds linked to a parent. Before the mode launched, ChatGPT mentioned a hotline in 33% of cases; after launch, in 23%. Recommendations to contact a trusted adult rose from 87% to 94%.

In Common Sense Media's tests, parental notifications arrived after a lengthy history of sensitive conversations. On newly linked accounts, researchers received no notifications after the first disclosure of a crisis, even following conversations lasting up to an hour. Four notifications arrived on accounts where these topics had been discussed for weeks.

OpenAI launched teen mode for ChatGPT on Aug 18, 2026. It activates automatically for eligible accounts belonging to users under 18 and is available on Free and paid personal plans. You can check the age category in Settings → Account → Age.

Stating an age in a conversation did not override the account setting in the institute's tests over several days. Accounts registered as adults never switched to teen mode, even when testers said they were 13 and ChatGPT acknowledged that age.

Family settings. Set up parental controls through Settings → Parental controls → Add family member. The parent sends an invitation by email or phone, and the teen accepts it. In the section under the teen's name, Quiet hours sets periods when use is restricted, including multiple separate blocks.

Study Hours starts new chats in Study mode on a schedule. This setting does not block ChatGPT and does not apply to conversations in GPTs and Projects.

According to OpenAI's help page dated Oct 7, the Study Hours schedule is off by default.