Is Meta AI Safe for Kids? WhatsApp, Instagram and the New Muse Agent (2026)

Short answer: Meta AI is built for 13 and up, it lives inside apps your child may already use (WhatsApp, Instagram, Messenger, Facebook), and it can't be fully switched off — so "is it safe" is really "what does my child's account actually allow, and what can I change". Muse, the personal AI agent Meta launched on 8 September 2026, is a different product: it's 18+, it does things on your behalf, and it isn't intended for teens at all.

Here's the whole picture as of September 2026, by product and by age, followed by the settings that exist today.

Last verified: September 2026.

The Three Things Called "Meta AI"

The one that usually prompts the question is the first: the assistant that appeared in WhatsApp and Instagram without anyone installing it.

Meta AI in WhatsApp and Instagram: What a Child Can Do

On a 13-to-17-year-old's account, Meta AI works like any chatbot: it answers questions, writes text, and (on adult accounts at least) generates images through the Muse Image model added in July 2026. In WhatsApp it can be messaged directly or brought into a group chat with @Meta AI. In Instagram it sits in the search bar and in DMs.

What that means in practice for a teen:

• It's always there. There's no install step and no separate app to spot on the home screen. A teen who has WhatsApp has Meta AI.
• It's a chatbot, with a chatbot's flaws. It can be confidently wrong, and it's designed to be pleasant to talk to. The paused AI characters were the bigger worry on that front; the assistant itself is closer to a search box that talks back.
• Group chats. If a friend @-mentions Meta AI in a group your child is in, the messages that friend shares with it are sent to Meta AI. WhatsApp's Advanced Chat Privacy setting, when turned on for a chat, blocks the use of that chat's messages with AI features.

Meta AI for Under-13s: There Shouldn't Be an Account

Meta's apps require users to be 13, so a child under 13 encountering Meta AI means they have an account they shouldn't. That's about to be enforced harder: under the settlement approved on 26 August 2026 with 29 US state attorneys general — about $18 billion — Meta must build an age-assurance system within a year to identify under-13 users and remove their accounts, and report annually on how it's going.

If your under-13 has a WhatsApp or Instagram account today, the honest advice is that the account, not the AI, is the issue.

What the August 2026 Settlement Changes for Teens

For 13-to-17-year-old accounts on Facebook and Instagram, the settlement commits Meta to:

• Overnight lockout: no access between midnight and 6 a.m.
• Notification curfew: push notifications off from 10 p.m. to 7 a.m.
• A daily time limit: a two-hour default, with parents in control of changes.
• Age assurance for under-13s, as above.

Separately, a court order in New Mexico requires Meta to prevent children there from romantic or sexualised interactions with its AI chatbots — the clearest sign yet that AI companions are now inside the scope of child-safety enforcement, not just feeds and notifications.

Two limits are worth knowing. Messaging — WhatsApp and Messenger — is excluded from the night mode, the time limit and the school mode. And the timelines run into 2027, so a teen's account today may not yet show all of this.

Muse: Why the Answer for Kids Is Simply No

Muse is not a chatbot with better manners. It runs on its own secure cloud computer with a web browser, can send emails, fill in forms, book appointments and make purchases, and keeps working after you close the app. Meta says a separate "Sentinel" system checks its actions before anything reaches the internet, that it has no visibility into your passwords or payment methods, and that conversations aren't shared with Meta's ad systems.

None of that is designed with a 14-year-old in mind, and Meta hasn't pretended otherwise: Muse is 18+, and its terms say you can't use it if you're under 18. The practical question for parents is enforcement. Muse is available inside WhatsApp, which teens use in huge numbers, and WhatsApp sits outside those limits. Age gating there is the same self-declared age as the rest of the app.

So the rule is simple: Muse is an adult's tool. If your teen mentions it, the useful conversation isn't "is it safe" but "what would you want an agent to do for you, and what could go wrong if it did the wrong thing?" — which is a genuinely good conversation to have at 15.

The Verdict by Age

What Parents Can Set Today

Nothing here removes Meta AI. These are the controls that actually exist as of September 2026:

1. Supervision on Teen Accounts. Meta's announced supervision tools for Teen Accounts let a supervising parent see the topics a teen discusses with Meta AI, but not the conversations themselves. Meta said these would arrive on Instagram first; check what your teen's account actually shows in Family Center.
2. Advanced Chat Privacy in WhatsApp. Turn it on in the chats that matter (family groups, your child's chats with you). It stops messages being exported or auto-downloaded and blocks the use of those messages with AI features.
3. Delete the Meta AI chat. In WhatsApp and Instagram the assistant's chat can be deleted like any conversation. It comes back if it's opened again, so treat this as a signal to your child, not a lock.
4. Device-level limits. Screen Time (iPhone) and Family Link (Android) can limit the apps themselves overnight — which, until the settlement's lockout arrives on every account, is the only version of it you control.
5. Keep AI characters off when they return. Meta has said the new version will let parents turn off chats with AI characters entirely and block specific ones. Turn them off first, then decide.

The Two-Line Rule

For a teen with Meta AI in their pocket, two sentences cover most of it, said out loud and not as a lecture:

"It's a computer program owned by an advertising company. Anything you tell it, assume you've told Meta."

Then the same one that works for every chatbot: "It sounds sure even when it's wrong. Check anything that matters."

FAQ

What is the minimum age for Meta AI?

13, because it inherits the minimum age of each app it lives in (WhatsApp, Instagram, Messenger, Facebook). Muse, Meta's personal AI agent, is separate and 18+.

Can I turn off Meta AI in WhatsApp?

Not fully. You can delete the Meta AI chat, and you can turn on Advanced Chat Privacy for individual chats to block their messages from being used with AI features, but the assistant itself can't be removed from the app.

Is Meta Muse safe for teenagers?

Muse's terms say you can't use it if you're under 18. It's an agent that takes actions — browsing, booking, buying — rather than a chatbot, and it's currently US-only. The realistic risk isn't Muse being marketed to teens; it's a teen with an adult-declared WhatsApp age finding it there.

Did the 2026 settlement change anything about AI?

Mostly it changed feeds, notifications and time limits for 13-to-17s, and forced age checks for under-13s. On AI specifically, a New Mexico court order requires Meta to prevent romantic or sexualised chatbot interactions with children, and messaging apps (WhatsApp, Messenger) were left out of its time limits and night mode.

Sources: Meta Newsroom — Introducing Muse (8 September 2026), Meta Newsroom — Introducing Muse Image (July 2026), CNN — Meta settles state child harm claims (26 August 2026), TechCrunch — Meta pauses teen access to AI characters (January 2026), Meta Newsroom — Teen AI safety approach (October 2025), WhatsApp Help Center — About Meta AI, Muse Terms of Service, Hunton — Court approves Meta settlement with 29 states, CNN — New Mexico judge orders Meta to pay $567 million (6 August 2026), EFF — What WhatsApp's Advanced Chat Privacy really does.

Further reading: Is Google Gemini safe for kids?, Is Claude AI safe for kids?, AI tools for teens: balancing independence and safety.