Explained: How OpenAI determines if a ChatGPT user is under 18

Age Prediction

• OpenAI launches age prediction for ChatGPT to boost teen safety safeguards
• Accounts flagged as under-18 get extra protections from harmful content
• Selfie age verification; parental controls let guardians manage settings

Did our AI summary help?

OpenAI has announced the rollout of an age prediction system for ChatGPT consumer plans as part of its broader effort to strengthen teen safety on the platform. The system is designed to estimate whether an account likely belongs to someone under the age of 18, allowing ChatGPT to apply age-appropriate safeguards automatically.

The company said the move builds on its previously announced Teen Safety Blueprint and principles for under-18 model behaviour, which aim to balance access to technology with protections for younger users.

How age prediction works

According to OpenAI, ChatGPT uses a dedicated age prediction model that analyses a combination of behavioural and account-level signals. These include how long an account has existed, typical activity times, usage patterns over time, and the age a user has stated during sign-up.

OpenAI said the system does not rely on a single signal. Instead, it uses multiple indicators to estimate whether an account may belong to a minor. Learnings from the rollout will be used to refine the model and improve accuracy over time.

When the system is not confident about a user’s age or has incomplete information, ChatGPT defaults to a safer experience.

Safeguards for under-18 accounts

If an account is identified as likely belonging to someone under 18, ChatGPT automatically applies additional protections. These are intended to reduce exposure to sensitive or potentially harmful material.

The restricted categories include graphic violence, gory content, sexual or violent role play, depictions of self-harm, viral challenges that could encourage risky behaviour, and content promoting extreme beauty standards or unhealthy dieting.

OpenAI said these safeguards are guided by academic research on child development and expert input, reflecting differences in risk perception and impulse control among teens.

Age verification and parental controls

Users who believe they have been incorrectly placed in the under-18 experience can confirm their age through a selfie-based verification process using Persona, an identity verification service. This option is available through the ChatGPT settings menu.

In addition, OpenAI highlighted parental control features that allow guardians to customise a teen’s experience. These include setting quiet hours, managing features such as memory or model training, and receiving alerts if signs of acute distress are detected.

What’s next

OpenAI said it will closely monitor the rollout and continue improving the system. In the European Union, age prediction will roll out in the coming weeks to meet regional requirements, with further updates planned as the system evolves.