ChatGPT for teens: what OpenAI's new dedicated experience means for young users and their parents
OpenAI automatically applies a teen-specific ChatGPT experience to under-18 accounts, adding Study Mode, safety guardrails, and parental controls.
From 18 August 2026, OpenAI is automatically enrolling any ChatGPT account it estimates or knows to belong to a 13, 17-year-old into a dedicated experience called ChatGPT for Teens. No opt-in required. If the system believes an account belongs to someone under 18, the protections switch on by themselves.
This is a meaningful shift. Until now, a teenager using ChatGPT had access to broadly the same product as an adult. The new experience changes that across four areas: how the tool handles schoolwork, what content it will and will not produce, how it manages screen time, and what controls parents can access.
Study Mode: guiding rather than answering
The headline feature for teens is Study Mode. Rather than handing over a finished answer, ChatGPT uses hints, step-by-step prompts, and follow-up questions to walk a student through a problem. The goal is understanding, not output.
There is also an anti-shortcut mechanism built in. If ChatGPT detects that a teen appears to be trying to get a quick answer to a homework question, it can redirect the conversation into Study Mode automatically, along with what OpenAI calls “responsible homework reminders.” Parents can go further by enabling Study Mode directly from their parental controls dashboard, making it the default for every new chat their teen starts. A Study Hours feature lets families set specific windows during which Study Mode is on by default.
For parents worried about AI doing their child’s homework wholesale, this is the most practically relevant change. It does not make cheating impossible, but it makes the tool substantially less useful as a shortcut machine.
Stronger content guardrails
Teen accounts will have tighter restrictions on self-harm content, eating disorders, violence, dangerous activities, and sexually explicit material. OpenAI is also extending safety notifications to include certain eating-disorder-related situations, meaning a parent could receive an alert if the system detects a possible serious safety concern in that area.
There is also a meaningful addition to how ChatGPT is allowed to behave in conversation. The updated model specification for under-18 accounts prohibits romantic language, discourages emotional dependence, and requires that ChatGPT does not imply it has feelings or consciousness. This goes beyond simply blocking roleplay scenarios. The intent is to keep the relationship clearly in the “useful tool” category rather than letting it drift toward something that resembles a social bond. Simple personalisation options such as accent colours and voice variations are still available, but the line between tool and companion is deliberately reinforced.
Break reminders and transparency nudges
If a teen has been active for 90 minutes within a three-hour window, ChatGPT will prompt them to take a break. The reminders also reinforce that they are talking to an AI, not a person. There are separate reminders around image uploads, cautioning teens against sharing private or sensitive photographs.
These are relatively light-touch interventions, and some experts have noted they may not be sufficient to address compulsive use patterns. That is a fair point. Nudges are not the same as hard limits, and a determined teenager can dismiss a notification. That said, the break reminders are at least consistent with the broader approach of keeping the experience functional without encouraging the kind of extended, emotionally dependent sessions that have attracted criticism.
What parents can actually control
OpenAI has built a new parental controls dashboard that allows a parent or guardian to link their ChatGPT account to a teen’s account. Once linked, parents can:
- Set quiet hours, during which the teen cannot use ChatGPT at all
- Turn Voice Mode on or off
- Disable memory so ChatGPT does not personalise responses based on past conversations
- Remove image generation
- Opt the teen’s account out of model training
- Enable Study Mode as a default
- Receive safety notifications in limited, serious situations
One thing parents cannot do is read their teen’s conversations. The controls are about settings and schedules, not surveillance. That is a deliberate choice by OpenAI, and it will suit some families and frustrate others. One parent or guardian can link to multiple teen accounts, but each teen can only be linked to one parent or guardian at a time.
When a teen turns 18, the protections switch off by default and the account moves to the standard adult experience.
How does OpenAI know someone is a teenager?
If a user states their age as 13, 17 during sign-up, the teen experience is applied. But OpenAI also runs an age prediction system that looks at signals including the topics discussed, the times of day the account is used, and how long the account has existed. If the system predicts the account likely belongs to someone under 18, the teen protections are applied automatically.
If you are 18 or older and your account has been incorrectly placed into the teen experience, you can verify your age to have the restrictions removed.
Why OpenAI is doing this now
OpenAI has faced several lawsuits related to harm involving young users. One, filed in California in August 2025, alleged that ChatGPT encouraged a 16-year-old to take his own life. Another involved a school shooting in Tumbler Ridge, Canada, where families sued OpenAI over allegedly alarming chats with the platform that were not acted upon. The US Federal Trade Commission has also opened an inquiry into potential harms to young users of AI chatbots.
OpenAI’s head of youth and families, Lauren Jonas, put the company’s position plainly: “If you don’t build a safe space for them, they’re going to go to a less safe model experience.” The approach is structurally similar to Meta’s teen accounts on Instagram, which apply stricter defaults rather than blocking access outright.
Child development experts have welcomed the direction while cautioning against over-confidence. The consensus view seems to be that these measures are meaningful steps, not complete solutions, and that parents should not treat them as a substitute for their own involvement.
What this means for you
If you are a parent with a teen using ChatGPT, the practical action is to visit chatgpt.com/parent-resources and link your account to your teen’s. Linking gives you access to the controls described above and means you will receive safety notifications if the system flags something serious. It takes a few minutes and gives you meaningfully more visibility and control than you had before 18 August 2026.
If you are a teenager, Study Mode is worth taking seriously rather than trying to route around. The hints-and-guidance approach is genuinely more useful for learning than a direct answer, even if it feels slower. And if your account has been put into the teen experience by mistake because you are over 18, age verification will sort that out quickly.
The rollout is global on Free and paid personal ChatGPT plans from 18 August 2026, with full availability in Australia from 8 September 2026. Protections have also been extended to group chats, the ChatGPT Atlas browser, and the Sora app.