How Old Do You Have to Be to Use ChatGPT? Age Rules for Every Major AI (2026)

ChatGPT requires users to be at least 13 years old, and anyone under 18 needs a parent or guardian's permission. That's OpenAI's official rule. But if you're asking this question, you probably want the answer behind the answer: how is that rule enforced, what happens when kids ignore it, and what are the rules for every other AI your kid can reach?

Here's the full picture, current as of July 2026 — including the newest wrinkle: AI platforms have started guessing your child's age.

Quick reference: AI age requirements in 2026

Now the details, because the details are where the real safety picture lives.

ChatGPT: 13+, parental consent under 18 — and now, age prediction

OpenAI's terms are simple: no accounts under 13, and users 13–17 must have a parent or guardian's permission. For years, enforcement was a date-of-birth field anyone could type anything into — and ChatGPT doesn't even require a login for basic use.

Two things have changed that picture:

Parental controls (late 2025). Parents can now link their account to their teen's account. Linked teen accounts automatically get content protections (reduced graphic content, no romantic or sexual roleplay, no extreme beauty-ideal content), and parents can set quiet hours, disable voice mode, turn off memory and image generation, and opt the teen out of model training. Importantly: parents do not get access to their teen's conversations. If OpenAI's systems detect signs a teen may be considering self-harm, parents may be notified — but the linking is optional, requires the teen's consent, and the teen is notified if they unlink.

Age prediction (2026). ChatGPT now analyzes account signals — topics, usage times, account age — to predict whether a user is under 18, and automatically applies the restricted teen experience if so, even if the user claimed to be an adult at signup. Adults flagged by mistake can verify their age through a third-party service (Persona); OpenAI says it receives the age result, not the ID itself.

What this means for parents: the "just lie about your birthday" era is ending at OpenAI — but the teen experience is a content filter, not visibility. You still can't see what your child asks ChatGPT.

For the full breakdown of ChatGPT's safety features and where they fall short, see our deep dive: Is ChatGPT Safe for Kids in 2026?

Google Gemini: the only major AI that officially allows under-13s

Gemini requires users to be 13 to use it with their own personal Google account. But Google made a move no other major lab has: children under 13 can use Gemini through a parent-managed Family Link account, if the parent enables it. Supervised accounts get automatic guardrails — image generation is off for minors, content filters apply, and Google says conversations from child accounts aren't used to train its models.

The trade-off parents should understand: Family Link lets you turn Gemini on or off for your child, but it does not show you their conversations. You control the door, not the room.

Character.AI: the rules just changed completely

If your information about Character.AI is more than a year old, throw it out. Facing multiple lawsuits and regulatory pressure over minors' interactions with its AI companions, Character.AI removed open-ended chat for all users under 18 — fully in effect since November 25, 2025 in the US, with other markets following. Teens can still use creative features (videos, feeds, character creation), but the core product — free-form conversation with AI personas — is now adults-only.

Enforcement is the strictest in the industry: an in-house age-detection model combined with third-party verification, because Character.AI learned the hard way that a checkbox doesn't stop a 14-year-old.

The lesson inside this story: the platform that leaned hardest into AI companionship is the one that ended up banning minors from it. That's worth remembering when evaluating any AI that wants to be your child's friend.

See the full comparison: HeyOtto vs. Character.AI

Snapchat My AI: the one your kid already has

My AI doesn't have its own age rule — it inherits Snapchat's minimum of 13, enforced by a date-of-birth checkbox. The difference: My AI isn't an app your child chooses to download. It's pinned at the top of the chat feed of an app they already use, presented like a friend.

Parents do have one meaningful control here: Snapchat's Family Center lets you disable My AI entirely for your teen, alongside visibility into who they've messaged (though never message content) and screen-time breakdowns. If your teen is on Snapchat, turning on Family Center and deciding deliberately about My AI is a ten-minute job worth doing this week.

Read more: Is Snapchat My AI Safe for Kids?

Meta AI: embedded everywhere, rules still forming

Meta's AI runs across Instagram, WhatsApp, Facebook, and Messenger, inheriting each platform's minimum age of 13. After reporting raised alarms about Meta's AI characters engaging in inappropriate conversations, Meta introduced supervision controls that let parents block their teens from chatting with AI characters.

Claude: 18 or nothing

Anthropic's Claude takes the opposite approach from everyone: the minimum age is 18, full stop, with no parental-consent path for minors. Worth knowing because teens increasingly encounter Claude through school and productivity tools built on it — the age rule applies to the consumer product.

The pattern across all of these

Look at the table again and you'll see the industry quietly admitting something: general-purpose AI was not built for children, and the patches are recent. Age prediction (2026), parental controls (late 2025), the Character.AI teen ban (late 2025) — every safeguard on this list is under a year old, reactive, and bolted onto products designed for adults.

Every major platform also shares the same two gaps:

No visibility. Even the best parental controls above filter content — none of them let a parent see the actual conversation. You get a lock on the door, but no window.

Consent, not design. "Parental consent" means a checkbox that says a parent agreed. It doesn't mean the product was designed for a developing mind.

That's the difference a purpose-built platform makes. HeyOtto is built for ages 8–18 from the ground up: parents create and manage the account (so the age question is answered by design, not by a checkbox), see full conversation transcripts, and get Real-Time Distress Alerts when something concerning appears — in a single message or in a pattern across many. Kids get a genuinely capable AI for learning and creating. Parents get the window, not just the lock.

Editor's Note

Facts in this post reflect platform policies as of July 2026 and were verified against each platform's published terms and announcements. Platform rules change fast — we update this page when they do.