OpenAI Launches New Age Prediction System to Improve Teen Safety on ChatGPT

OpenAI has officially begun rolling out its long-announced age prediction and verification system, aimed at strengthening safety protections for teenagers and minors on ChatGPT. First revealed in September 2025, the AI-powered system works in the background to estimate a user’s age based on activity signals, helping prevent underage users from accessing potentially harmful interactions.

OpenAI Introduces AI-Based Age Verification System

OpenAI has finally launched its age verification system, an artificial intelligence model that operates behind the scenes in ChatGPT. While users are asked to enter their date of birth when creating an OpenAI account, this information is not verified. The company says the new system is designed to ensure additional protections for younger users, especially those who may have entered an incorrect birth date.

Global Rollout Excluding the European Union

According to OpenAI, the age verification system is being rolled out globally, except in the European Union, where it will launch in the coming weeks due to regional regulatory requirements.

The system primarily targets teenagers and minors who bypass system-level safety settings by entering a false date of birth. OpenAI’s teen safety settings already restrict access to sensitive content and certain types of interactions, and the new system strengthens these protections.

How OpenAI’s Age Prediction System Works

OpenAI explained that the system analyzes user activity across each account, monitoring signals such as:

• Topics discussed

• Interaction patterns

• Typical times of activity

Based on these signals, the AI estimates a user’s real age. The company admits the system is not perfect and may occasionally misclassify adult users as being under 18.

Whenever an account is classified as underage, non-removable safety settings are automatically applied by default.

What Teen Safety Mode Restricts on ChatGPT

Even with teen safety settings enabled, users can still use ChatGPT to ask questions, create content, and learn. However, the AI will not provide responses related to:

• Graphic violence or gore

• Dangerous or harmful viral challenges

• Sexual, romantic, or violent role-play

• Content promoting extreme beauty standards

• Unhealthy dieting practices

• Body shaming

These restrictions are designed to reduce exposure to potentially harmful material.

Age Verification for Adults Misclassified as Minors

For adult users who are mistakenly identified as minors, OpenAI is offering a way to verify their age. The company has partnered with Persona, a third-party online identity verification platform.

To complete verification, users must:

• Upload a live selfie

• Upload a government-issued ID

Persona checks the user’s date of birth and verifies whether the selfie matches the ID photo. Once Persona confirms the user is an adult, the restrictions are lifted. OpenAI notes that it may take some time for the changes to fully apply.

Privacy and Data Protection Measures

OpenAI stated that only the photo, name, and date of birth on the ID are verified. Users can hide other personal details when uploading documents.

Since the verification process is handled entirely by Persona:

• OpenAI has no access to the uploaded ID or selfie

• All verification data is deleted within seven days

OpenAI does retain access to the signals and analysis generated by the age prediction system but claims this data is securely stored and used strictly according to its privacy policy.

Option to Bypass Age Prediction Screening

OpenAI is also giving users the option to opt out of age prediction screening by voluntarily completing age verification themselves. Interested users can initiate the process directly from their account settings.