AI Tool Age Limits for Kids
Most major AI chatbots set 13 as their minimum age, but the exceptions matter more than the rule: Claude requires 18+, and Character.AI removed open-ended chat for under-18s in November 2025. The stated age is a legal floor, not a recommendation — the table below gives both, plus what parental controls actually do.
Checked against vendor terms and help documentation on 28 July 2026. Every row links to its sources. These policies change often.
What age does each AI tool actually allow?
| Tool | Type | Vendor minimum | Our recommendation | Parent can see chats? |
|---|---|---|---|---|
| ChatGPTOpenAI | Chatbot | 13+, and under-18s need a parent or guardian’s permission | 13+ with supervision | Partially, via linked parent account |
| Google GeminiGoogle | Chatbot | 13+ on a personal or school Google account; 18+ on a work account | Under 13 only via Family Link; 13+ with agreed rules | Under-13 via Family Link only |
| Snapchat My AISnap | Social | 13+, following Snapchat’s own minimum age | 16+ with active supervision, if at all | No — not even with Family Center |
| Meta AIMeta | Social | 13+, matching the minimum age across Meta’s platforms | 16+, with AI character chats switched off below that | Not applicable (18+) |
| Microsoft CopilotMicrosoft | Chatbot | 13+, and higher in some countries | 13+, ideally through a school account | Not applicable (18+) |
| PerplexityPerplexity AI | Search | 13+, and under-18s need a parent to accept the terms on their behalf | 13+ with supervision; better suited to 15+ | Not applicable (18+) |
| DuolingoDuolingo | AI tutor | 13+ for a standard account; under-13s need a parent’s email for consent | Any school age; use Duolingo ABC under 8 | Not applicable (18+) |
| PhotomathGoogle | AI tutor | 13+, and minors need a parent or guardian’s permission | 13+, with an agreed "attempt it first" rule | Not applicable (18+) |
| BrainlyBrainly | AI tutor | 13+ in the US, but under-16s are not permitted in some regions | 13+ in the US, 16+ elsewhere, with supervision | Not applicable (18+) |
| GrammarlyGrammarly | Creative | 13+ in the US, 16+ everywhere else | 13+ in the US, 16+ elsewhere; school account below that | Not applicable (18+) |
| QuizletQuizlet | AI tutor | 13+ for Quizlet itself; the Q-Chat AI tutor launched at 18+ | Any school age for flashcards; 18+ for Q-Chat | Not applicable (18+) |
| Character.AICharacter Technologies | Companion | 13+ nominally, but open-ended chat is now restricted to 18+ | 18+ | Not applicable (18+) |
| ClaudeAnthropic | Chatbot | 18+ — the strictest age floor of the major chatbots | 18+ | Not applicable (18+) |
| KhanmigoKhan Academy | AI tutor | Account holder must be 18+; children access through a parent or school | Any school age, via a parent or school account | Yes — by design |
| ReplikaLuka | Companion | 18+ only | 18+ | Not applicable (18+) |
| GauthGauthTech (ByteDance) | AI tutor | Contradictory — 13+ in one place, 18+ in the terms of service | 16+, and only with the uploading rules agreed first | Not applicable (18+) |
“Vendor minimum” is what the terms of service require. “Our recommendation” is our own editorial judgement and is deliberately stricter in places — a legal minimum age is not the same thing as advice.
Which AI tools have real parental controls?
This is where the differences are largest, and where marketing is least reliable. A tool can be “13+” and still give a parent no visibility whatsoever.
Parental controls exist via account linking: a parent can connect their account to a teen’s and configure age-appropriate settings.
Family Link gives genuine parent-side control for under-13s, including switching Gemini access on or off. This is stronger than most competitors offer.
Snapchat’s Family Center lets a parent see who their teen talks to and how long they spend in the app, and disable My AI — but parents cannot read message content, cannot see what a child discussed with My AI, and the teen must actively consent to being monitored. Removing My AI entirely requires a Snapchat+ subscription (around $3.99/month).
Parents can turn off a teen’s one-on-one chats with AI characters entirely, and can see the topics their teen discusses with AI characters — topic-level insight, not full transcripts. The self-harm alerting is automatic for supervised teens aged 13–18.
Microsoft publishes dedicated age-limit and parental-control guidance, and access can be managed through family safety settings and, in schools, by the tenant administrator.
None built in.
The consent email and the COPPA social-feature restrictions are the main mechanisms. There is no rich parent dashboard.
None built in.
None built in.
None for individual accounts. The school route is the supervised one, governed by the institution.
None beyond the automatic restricted experience for younger accounts.
Limited. The primary control is now the platform-level under-18 restriction rather than parent-configurable settings.
None, because the product is not intended for minors at all.
The strongest of any tool listed here: chat history is visible to the parent by design, with safety alerts, rather than monitoring being bolted on afterwards.
None. The product is not built for minors in any configuration.
None built in.
A safer place for your child to actually learn AI
LittleAIMaster teaches how these systems work through guided, age-appropriate projects — rather than handing a child an open-ended chatbot. Free to start.
Or get the mobile app