OpenAI Announces Age Verification for ChatGPT Following Teen Suicide Case

OpenAI said in September 2025 that it would build a system to estimate whether ChatGPT users are under 18. In January 2026 it began rolling that system out. This page explains what OpenAI has said about how it works, what happens when it gets an age wrong, and the privacy trade-off.

The plan announced in September 2025

On September 16, 2025, OpenAI CEO Sam Altman wrote that the company would put teen safety ahead of privacy and freedom for minors. NBC News reported that the plan was to sort people into a teen experience (13 to 17) or an adult experience (18 and over), and to default to the under-18 experience when the company is unsure. It also said ID might be required in some cases or countries. OpenAI called this a privacy compromise for adults and a worthwhile one.

The January 2026 rollout

OpenAI announced on January 20, 2026 that it was rolling out age prediction on consumer plans. We could not open the page, so this summary relies on the help center and news coverage. According to OpenAI’s help center:

• The model looks at account and behavior signals, such as how long an account has existed, when it is active, how it is used over time, and the age the person gave.
• It still runs if you entered a birth date at sign-up. It works as an extra check.
• If it decides an account may belong to someone under 18, ChatGPT can apply the teen experience. The account keeps working. Extra limits apply to sensitive content, such as graphic violence and sexual or romantic roleplay.

If an adult is placed in the teen experience

Adults can verify their age through Persona, an outside company. Depending on the country, Persona may ask for a live selfie or a government ID. OpenAI says it does not see what you share with Persona, only the result and date of birth. Rules differ by region, so read the current help page.

Limits and open questions

OpenAI has said that even advanced systems sometimes struggle to predict age. That means two kinds of mistakes are possible: adults treated as teens, and teens treated as adults. We have not found independent testing of the system’s accuracy.

A guess based on how someone uses an app can also feel invasive. Whether the trade-off is worth it is a personal call. OpenAI has framed it as necessary for protecting minors.

Age checks help at the margins. They do not tell you whether a teen is leaning on a chatbot for comfort. For that, see our parent guide to OpenAI’s age rules.

Need to talk to someone today? The 988 Lifeline takes calls and texts in the US. Crisis Text Line answers when you send HOME to 741741. You can also reach SAMHSA on 1-800-662-4357. If anyone is in danger right now, phone 911.

Will ChatGPT ask me for ID?

Only if the system places you in the teen experience and you choose to verify as an adult. Then an outside provider may ask for a selfie or an ID, depending on your country.

Does age prediction read my conversations?

OpenAI says the model uses account and behavior signals such as account age and usage times. We have not seen a full technical description, so check OpenAI’s current help page.

This page is for education only. It is not legal advice.