Instead of relying solely on self-reported birthdates—which are easily bypassed—OpenAI is using a machine-learning model to “guess” a user’s age based on their platform behavior. If the system flags an account as likely belonging to a minor, it automatically triggers a “restricted experience.”
How the Prediction Model Works
The model analyzes a combination of metadata and interaction patterns to estimate age:
- Account-Level Signals: How old the account is and the age the user initially shared.
- Usage Patterns: Typical times of day when the user is active (e.g., activity during school hours vs. late night).
- Behavioral Clues: General topics of discussion and the complexity of language used in prompts.
- Safety Default: If the system is “uncertain,” it defaults to the under-18 experience to ensure maximum protection.
Automated Safeguards for Teens
Once a user is identified as being under 18, ChatGPT automatically restricts or modifies responses related to:
- Sensitive Content: Stricter filters on graphic violence, gore, and self-harm.
- Risky Behavior: Blocking content related to dangerous viral challenges.
- Mature Interactions: Prohibiting sexual, romantic, or violent role-play.
- Body Image: Limiting content that promotes extreme beauty standards or unhealthy dieting.
The “Adult Mode” & Identity Verification
For users who are incorrectly flagged as minors or those who want to access the upcoming “Adult Mode,” OpenAI has introduced a formal verification process through Persona, a third-party identity service.
| Feature | Under-18 Experience | Verified Adult Mode |
| Content Filters | Strict / Protective | Standard / “NSFW” enabled |
| Verification | Predicted by AI | Selfie + Government ID |
| Parental Oversight | Available (Quiet hours, alerts) | None |
| Ads | Age-appropriate | Standard |
How to Verify Your Age (18+)
If you are incorrectly placed in the teen experience, you can restore full access by going to Settings > Account. You will be asked to provide:
- A Live Selfie: To ensure you are a real person.
- Government ID: Persona checks the DOB and matches the photo to your selfie.Note: OpenAI claims it does not receive your ID or selfie; Persona deletes this data within 7 days.
Why the Sudden Push?
The rollout comes as OpenAI faces mounting legal and regulatory pressure:
- FTC Investigation: The U.S. Federal Trade Commission is currently probing how AI chatbots impact the mental health of minors.
- Wrongful Death Lawsuits: OpenAI is defending several cases involving teen suicides allegedly linked to chatbot interactions.
- EU Compliance: The system is being adapted to meet strict European digital safety laws, with a targeted EU rollout in late February 2026.
Conclusion: Walking the Tightrope
By introducing age prediction, OpenAI is attempting to solve a classic tech dilemma: protecting children without infringing on the freedom of adults. As CEO of Applications Fidji Simo noted, the goal is to “treat adults like adults” while ensuring that the most advanced AI tools remain safe and beneficial for the millions of students who use them daily.
