ChatGPT is beginning to automatically give teenagers a more restricted experience, with stronger safety protections and new tools for parents. The change comes as governments and families around the world ask harder questions about how children are using AI.
If your teenager uses ChatGPT for homework, questions or advice, their experience may soon look a little different.
OpenAI announced ChatGPT for Teens on August 18, with a global rollout for eligible users aged 13 to 17. OpenAI says full availability in Australia is expected by September 8.
The biggest change is not a new app. It is that ChatGPT may now try to determine whether the person using it is under 18 and automatically apply stronger protections.
How does ChatGPT know a user is a teen?
According to OpenAI’s age-prediction guidance, the system may consider signals such as the general topics a person talks about, the times of day they use ChatGPT, how the account is used and how long it has existed.
That means a teenager does not necessarily have to turn on a special setting. If ChatGPT estimates that an eligible user is under 18, the teen experience can be applied automatically.
If an adult is mistakenly identified as a teenager, they can verify their age to remove the teen protections.
For parents, this is worth understanding because it represents a broader change in online safety: platforms are moving beyond simply asking users to enter a birthday and are beginning to estimate age from other signals.
What changes for teens?
ChatGPT will still answer questions, help with schoolwork and support creative projects. But the teen experience places stronger safeguards around sensitive areas such as graphic violence, risky viral challenges, sexual or violent roleplay, unhealthy dieting and other age-inappropriate content.
It also includes Study Mode, quizzes, homework reminders, break reminders and Study Hours that can encourage teens to work through a problem rather than simply ask AI for the answer.
Parents who link their account with their teen can also manage selected settings, set Quiet Hours, and receive limited safety notifications in serious situations.
OpenAI’s parental-control guidance says notifications may be sent when trained reviewers identify serious self-harm concerns or when a teen’s account is deactivated because of conversations related to acts of violence.
There is an important limit: parents cannot read their teen’s ChatGPT conversations or monitor their chats in real time.
Children are using AI for more than homework
This may be the most important point for parents.
A recent study by Australia’s eSafety Commissioner surveyed 1,950 children aged 10 to 17 and found that 78% had used an AI assistant.
More than half of those who had used AI assistants or companions said they had turned to them for personal or social reasons, including asking for advice, discussing physical or mental health, or talking about their feelings.
One in five reported a potentially inappropriate or harmful interaction, while nearly one in three said they had shared personal or potentially sensitive information with an AI tool.
So the parenting question is becoming bigger than:
“Is my child using ChatGPT to do their homework?”
It may also be worth asking:
“What does my child talk to AI about?”
Why the safety debate is getting more serious
The new protections arrive during growing scrutiny of AI chatbots and young users.
Last year, the parents of 16-year-old Adam Raine filed a wrongful-death lawsuit against OpenAI, alleging that ChatGPT provided harmful responses during months of conversations about suicide. OpenAI said at the time that ChatGPT had safeguards but acknowledged that some protections could become less reliable during long conversations, according to Reuters’ report on the case.
More recently, a separate case in Massachusetts raised questions about AI and violence.
A 17-year-old has been charged with killing his mother and younger brother in Acton. According to a Guardian report citing the Middlesex County district attorney, investigators found that the teenager had used the internet and ChatGPT to explore theoretical ideas and fantasy stories involving the killing of his family. He has pleaded not guilty.
There is no evidence presented so far that ChatGPT caused the killings or instructed him to commit them.
But the case raises a difficult question for parents and AI companies: if a young person repeatedly discusses serious violence or self-harm with a chatbot, when should the technology recognize that something may be wrong?
This is becoming a global issue
Australia is already bringing AI chatbots into its broader child online-safety framework. Its eSafety Commissioner says rules that took effect this year apply to certain generative AI services and require protections against children’s exposure to material involving pornography, high-impact violence, self-harm and suicide. The regulator says these protections also cover AI-powered chatbots and companions.
Canada has proposed its own Safe Social Media Act, which would place child-safety duties on certain AI chatbot services.
In the United Kingdom, Ofcom is testing ChatGPT and other AI chatbots using fictional accounts representing teenagers aged 13 to 17 to study how the services respond to younger users and potentially harmful prompts.
And in the United States, OpenAI faces growing legal and regulatory pressure over child safety. Reuters reported that the company was already facing multiple lawsuits alleging harm to children when it introduced ChatGPT for Teens.
What parents can do now
Parents do not need to become AI experts. A few simple habits can help:
- Ask your child how they use ChatGPT, including whether they use it for advice, feelings or personal problems.
- Link parental controls if your teenager uses ChatGPT regularly, and review the settings together.
- Talk about privacy. Remind children not to share passwords, addresses, private photos or other sensitive information with a chatbot.
- Use AI for learning, not just answers. Features such as Study Mode can help children work through a problem instead of simply completing the assignment for them.
- Keep humans in the conversation. If a child is deeply upset, thinking about self-harm, afraid or talking about hurting someone, an AI chatbot should never be their only source of support.
ChatGPT for Teens is a meaningful attempt to make AI safer for younger users. But no age-prediction system, parental setting or safety filter can replace an open relationship between children and the adults they trust.
As AI becomes part of everyday childhood, talking about how children use it may need to become as normal as talking about school, friends and social media.
