OpenAI Launches ChatGPT Teen Experience Safeguards
OpenAI introduced a dedicated ChatGPT experience for teenagers on August 18, applying additional safeguards and parental controls to users aged 13-17 and accounts its system estimates belong to minors, The Verge reports. OpenAI's age-prediction documentation states that the system uses account and usage signals, while adults incorrectly placed in the teen experience can verify their age to remove the restrictions.
OpenAI introduced a dedicated ChatGPT experience for teenagers on August 18, combining youth safeguards, content restrictions, and parental controls for users aged 13-17, according to The Verge. The experience also applies to accounts that OpenAI's age-prediction system estimates may belong to users under 18. Children under 13 are not permitted to use ChatGPT under OpenAI's policies.
The Verge reports that the teen experience applies tighter defaults for graphic violence, self-harm depictions, and sexual or romantic roleplay. It also includes warnings and safety information for subjects such as eating disorders. Parents can set quiet hours and receive notifications when the system identifies a potential safety risk, according to the publication.
OpenAI described the offering as an experience intended to help teens "learn, think critically, deepen understanding, and use AI with confidence," The Verge reported. In a July 16 post, OpenAI wrote that nearly nine in 10 teens using ChatGPT in a given week use it for learning, information, skill-building, or productivity. That figure is the company's own characterization of teen usage.
Age prediction determines access
OpenAI's documentation states that age prediction is rolling out globally, with an EU rollout expected in the coming weeks to accommodate regional requirements. The company documents that its system may consider several account-level signals, including general conversation topics, time of use, patterns of account use, and how long an account has existed.
When the system predicts an account holder is under 18, ChatGPT can apply the teen experience without deactivating the account. OpenAI acknowledges that the system can make mistakes. Adults who are assigned the teen experience in error can complete age verification to have the additional protections removed.
The documentation does not specify a model architecture, accuracy rate, false-positive rate, or the relative weight of the signals used for age prediction. It also states that users in Italy who are asked to verify their age have 60 days to do so, after which certain features may be disabled until verification is completed.
Safety controls raise implementation questions
For AI product and trust-and-safety teams, the rollout illustrates a broader industry pattern: age-aware experiences require a tradeoff between conservative protection for potentially underage users and friction for adults incorrectly classified as minors. Systems using behavioral and account signals also require clear user recourse, since prediction errors can affect access to features and sensitive-content policies.
OpenAI's published approach combines automated classification with an age-verification appeal path and parent-facing controls. For practitioners evaluating similar systems, the undisclosed performance metrics leave important operational questions around calibration across regions, the handling of uncertain classifications, and whether safety interventions produce measurable reductions in harm without blocking legitimate educational use.
Key Points
- 1OpenAI's teen experience applies safeguards to declared minors and accounts predicted to belong to users under 18, expanding protection beyond self-reported age.
- 2The system uses behavioral and account signals for age prediction, making false-positive handling and age-verification recourse central implementation concerns.
- 3Comparable age-aware AI systems commonly balance conservative content protections against access friction, especially when classifiers operate with incomplete demographic certainty.
Scoring Rationale
This is a notable safety-product deployment from a major AI platform, with relevance to teams building age-appropriate generative AI experiences. Its use of account and behavioral signals for age prediction raises practical questions about classifier calibration, verification workflows, and privacy-sensitive enforcement.
Sources
Primary source and supporting public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
