AI Chatbot Parental Controls & Teen Safety Comparison

This is a rules-and-lookup tool, not a calculator. Each of the 9AI assistants documents its minimum age and, where they exist, a set of teen-safety controls. The tool stores those facts verbatim from the provider's own policy or help page — nothing is inferred — and turns the age and country you enter into a plain-English verdict per provider.

Every provider record holds an own-account minimum age, an optional guardian-managed path (a linked teen account, Family Link, or guardian consent), an optional EU/EEA override, and six documented safety dimensions: a dedicated teen experience, parental controls, content filtering for minors, memory or training turned off, usage time limits, and self-harm safeguards with parent notification. The classification runs in order:

1. Resolve the effective minimum age for the region. In the EU/EEA a stricter floor is used where a provider sets one (Character.AI is 16 there instead of 13).
2. Decide permission. The child is permitted if their age meets the own-account minimum, or if it clears the guardian-path floor. Otherwise the verdict is Not permitted at this age, with the years until the first path opens.
3. If a provider confines minors to a cut-down experience (Character.AI removed open-ended chat for under-18s in November 2025), the verdict is Limited teen-only experience.
4. Otherwise, score the three supervision dimensions — parental controls, content filtering, and self-harm safeguards — with a documented control counting 1 and a partial one 0.5. A score of 2.5 or more gives Allowed with parental controls (strong), 1 to 2 gives (basic), and 0 gives Allowed, no dedicated child safeguards.

Two providers set no explicit lower age for their guardian path — Google Gemini through a Family Link supervised account, and DeepSeek “with the consent of your legal guardian” — so both are encoded as an any age with a guardian path rather than a guessed number. Where a provider does not document a control, the cell reads Not documented; it is never filled in with a guessed “No”. To keep the results trustworthy, a cross-check recomputes each permitted-or-not verdict a second way — from the youngest usable age for the region — and confirms all 9 agree, the same idea as verifying a tax figure with a second formula.

Each control and age is copied from and linked to the provider's own live policy, help centre, or newsroom post, verified on 2026-07-18:

• ChatGPT (OpenAI) — OpenAI — Parental controls FAQ
• Claude.ai (Anthropic) — Anthropic — Usage Policy
• Gemini (Google) — Google — Gemini Apps & Family Link
• Copilot (Microsoft) — Microsoft — Copilot age limits & parental controls
• Meta AI — Meta — Teen Accounts & AI protections
• Character.AI — Character.AI — Safety Center & newsroom
• Grok (xAI) — xAI — Terms of Service
• Perplexity — Perplexity — Terms of Service
• DeepSeek — DeepSeek — Terms of Use

This page summarises public safety and terms pages for convenience and is not legal or safety advice. AI providers change their policies often; the facts are reviewed on the last-verified date above and whenever a provider announces a teen-safety change. For the minimum sign-up age on its own, see the AI Chatbot Age Requirement Checker, and for training and retention, the AI Data Privacy Comparison.