ChatGPT for Teens: Age Prediction and Study Mode

Who gets ChatGPT for Teens by age prediction or verification

If the account is eligible and its age information points to someone under 18, ChatGPT turns on the teen experience without anyone asking for it. The rollout began August 18 on Free and paid personal plans. For a parent, a teen or an adult who saw the label, the first question is which signal applies to the account.

Under 18 by stated age or OpenAI age prediction

Age prediction is OpenAI's estimate of whether someone is under 18. The Help Center names three signals for an eligible account:

• The age given on the account
• A verified age
• OpenAI age prediction

Going by the Help Center's wording, any one of the three pointing to someone under 18 is enough. The blog gives a shorter version: if OpenAI's system estimates someone is under 18, or they state an age between 13 and 17, they are placed into ChatGPT for Teens automatically. The launch builds on earlier work: parental controls, the updated under-18 Model Spec, the Teen Safety Blueprint and the rollout of age prediction.

An adult placed here by mistake has one documented way out. They can verify their age to ask ChatGPT to remove the teen protections.

Rollout dates, plans, and what happens at 18

The Help Center said the rollout to eligible teen accounts on Free and paid personal plans began August 18. It also said full availability in Australia was expected September 8.

A note at the top of the Help Center page is for anyone who does not see the teen features yet. It said the features may not be available for that account, and that availability would keep expanding over the coming weeks. Features may also differ by region.

An account does not stay in the teen experience forever. When ChatGPT identifies the account holder as 18 or older, the account may move out of ChatGPT for Teens, which can remove the teen protections and end a linked parental-controls connection. Certain protections, such as parental controls or reduced sensitive content, can be reinstated at any time.

How Study Mode and homework reminders work for schoolwork

Once an account is covered, the learning side comes in four parts: Study Mode, new responsible homework reminders, Quizzes and Learning Visualizations, and new Study Hours. The Help Center adds that ChatGPT for Teens is not limited to schoolwork, and teens can also use it to build, create and ask everyday questions. The blog says ChatGPT for Teachers, a separate product, continues to serve educators and school districts.

Study Mode guides with questions instead of answers

Picture a teen studying for a test. According to the blog, the teen can ask ChatGPT to explain a concept that is new to them, work through a practice problem and check their understanding. If the teen asks for a quick answer, ChatGPT for Teens can guide them toward the steps and questions that let them work it out. The wording is "can guide", so the blog does not promise that every quick-answer request is redirected.

OpenAI says Study Mode was built with teachers, scientists and pedagogy experts. The blog adds that the design is consistent with research showing that active retrieval and self-explanation can support stronger comprehension and longer-term retention than passive review alone. OpenAI also says its own early evaluation showed promising gains in student performance, but the blog gives no figure.

Responsible homework reminders and Study Hours

The responsible homework reminders can recognize when a teen appears to be trying to shortcut an assignment and redirect them toward Study Mode. TechCrunch says the product promises to address what it calls the educational crisis of ChatGPT-assisted cheating in schools.

Study Hours let the teen or a linked parent or guardian choose when eligible new chats start in study mode, and the Help Center says this does not block access to ChatGPT. Quiet Hours are a different setting: they can limit access during scheduled times. Quizzes and Learning Visualizations round out the set, with more ways to practice and test what the teen knows.

What the safety protections and parental controls cover

The safety protections are on by default. For users under 18, ChatGPT for Teens applies age-appropriate safeguards based on the Under-18 Principles in OpenAI's Model Spec, the rules the company publishes for how its models should behave.

Age-appropriate limits on self-harm and explicit content

The blog says ChatGPT for Teens applies model protections and product interventions in higher-risk areas such as self-harm, violence, eating disorders, dangerous activities, and explicit sexual or graphic content. The stated aim is to reduce exposure to content that may be harmful or developmentally inappropriate while keeping the ability to learn, create and explore.

OpenAI also says it has added under-18 evaluations to its system cards, which show how its models perform against teen-specific standards in sensitive-content cases. The areas it names include self-harm, eating disorders, violence, sexual content, and age-restricted goods and services. The blog gives no scores.

The under-18 Model Spec on emotional dependence

The updated under-18 Model Spec goes past blocking romantic or sexualized roleplay. It rules out romantic language, encouraging emotional dependence, and implying that ChatGPT has feelings or consciousness.

The blog lists product features as well. Break reminders encourage teens to step away, and product cues consistently identify ChatGPT as AI. Sensitive-image reminders warn teens not to upload private or sensitive pictures.

Parental controls: Quiet Hours, no access to chats

Parents and guardians can optionally link accounts through parental controls. With a linked teen account, a parent can do three things:

• Set Quiet Hours, which limit access to ChatGPT during scheduled times
• Manage selected settings
• Receive safety notifications in limited high-risk situations

The blog said OpenAI was adding notifications related to eating disorders, while limiting what is shared. Parental controls give a parent or guardian no way to read or monitor the conversations. A safety notification shares only the information needed to support the teen's safety.

How well does age prediction work? What is not published

Neither OpenAI's August 18 blog post nor its Help Center article on ChatGPT for Teens gives an accuracy figure for age prediction. TechCrunch says it is unclear how hard the new systems will be for teens to get around. The Help Center's stated route for an adult placed in the teen experience by mistake is to verify their age.

Lawsuits behind the launch and the question of timing

TechCrunch reports that the launch follows numerous lawsuits over AI chatbots' lack of safety measures, which it says led to teen suicides and other mental health concerns. It also asks why the protections were not part of ChatGPT from the beginning. In its telling, ChatGPT arrived in late 2022 and reached 900 million weekly users before meaningful safeguards designed specifically for teenage users were added.

TechCrunch also points out that teens are very good at working around parental controls. It adds that the teen mode has not yet been put to more strenuous tests.

ChatGPT for Teens is the default for an eligible account that a stated age, a verified age or age prediction places under 18, and it pairs Study Mode and homework reminders with age-appropriate protections. Linked parents can set Quiet Hours and receive limited safety notifications, but they cannot read the conversations. OpenAI's blog post and Help Center article give no accuracy figure for age prediction.