OpenAI limits ChatGPT features for teen users

OpenAI said it has rolled out a specialised version of ChatGPT for users aged 13 to 17, introducing tighter limits on what teenagers can see and how the system replies. The company framed the update as a child-safety and governance-focused approach to making AI tools available to minors, including in education settings.

The announcement comes as regulators, educators, and parents continue to raise concerns about how AI systems could affect adolescent psychological development. OpenAI said the teen experience is intended to reduce exposure to self-harm, suicide, and sexual content, while also restricting interactions that could encourage emotional dependence on the system.

How the teen interface changes content and behaviour How the teen interface changes content and behaviour OpenAI said the teen experience applies enhanced content OpenAI said the teen experience applies enhanced content filtering that targets high-risk categories. The company listed self-harm and suicide-related material, as well as sexual content, as areas where the restrictions are designed to prevent harmful or age-inappropriate outputs for users in the 13–17 group. Alongside these filters, OpenAI said it has added technical guardrails aimed at preventing the model from simulating personal feelings or consciousness. The company described this as a way to reduce the risk that teenagers treat the system as emotionally aware and develop overreliance. To route users into the appropriate environment, OpenAI said it uses age-assurance heuristics that can automatically place users identified as minors into the restricted teen interface. The company did not provide details on how those heuristics work in its description of the deployment. Parental tools, oversight options, and risk alerts OpenAI said the teen offering includes institutional oversight features and optional parental controls. The tools include “quiet hours,” which can limit access during certain times, and safety notifications linked to high-risk queries. The company described these controls as optional, suggesting families or institutions can choose whether to enable them. OpenAI said the aim is to connect access limits with safety monitoring, especially in situations where user prompts indicate heightened risk. Education-focused adjustments to reduce cheating OpenAI said educational functionality in the teen experience has been modified to emphasise teaching and guidance rather than simply producing direct answers. The stated is to reduce academic dishonesty by pushing responses toward a pedagogical approach. OpenAI also positioned the deployment as part of OpenAI also positioned the deployment as part of a broader effort to integrate AI tools into formal education while managing liability risks tied to child safety and data governance. The company linked its approach to wider industry efforts to adopt age-appropriate design in digital services.

Limits and open questions in the rollout

OpenAI highlighted that key uncertainties remain, including how effectively age-assurance heuristics identify minors and how consistently safety notifications and filters work across different types of prompts. As described by the company, the central focus is reducing exposure to high-risk content and discouraging emotional overreliance while keeping an education-oriented mode available to teenage users.