Instagram’s AI age check goes global, moving millions more teens into Teen Accounts

Meta says it is expanding the AI age-estimation technology on Instagram to every region, so that teens aged 13 to 17 who signed up with an adult birthday are placed in Teen Account protections. Meta expects “millions more teens” to be covered, according to the report of the announcement and Meta’s own newsroom post.

What Meta says changes on Instagram

In its post, “Expanding Teen Account Protections to Teens Around the World”, Meta says the technology looks for people it suspects are teens even when their account lists an adult age, and moves them into Teen Accounts. Meta’s wording is that “millions more teens across the world who use Instagram will receive these built-in protections”. Teen Accounts limit who can contact a teen, what content the teen sees and how much time the teen spends in the apps.

Meta adds that it is continuing the rollout on Facebook and plans to launch similar technology globally on Threads later this year. For accounts that may belong to children under 13, it says it is “investing in even stronger technology to proactively catch” them.

Two further pieces come with the expansion. In the United States, parents will see an in-feed Instagram video this month that explains Meta’s approach to age-appropriate protections. It links to an article by Larry Magid of ConnectSafely, which Meta says is available to parents worldwide on its Family Center. In Australia, Meta points to Apple’s age-range signal: a parent can choose to have the app store confirm a device user’s age range, which Apple then passes to developers when a teen downloads an app. Meta describes that signal as voluntary.

A push for app-store signals and rivals to follow

Meta uses the announcement to argue for a shared approach. It calls on TikTok and YouTube to adopt similar practices and says the ideal is “One verified signal from people’s devices, shared consistently with all apps”. It says consistent app-store adoption would give teens more uniform protections. The post also cites an agreement with 52 bipartisan attorneys general, covering US states, territories and the District of Columbia.

This is a continuation of earlier steps. According to Meta’s May 2026 post on AI age assurance, placement of suspected teens first launched on Instagram in the US, Australia, Canada and the UK, then expanded to the 27 EU countries and Brazil, with Meta saying it aimed to go global on Instagram during the year. The same post says hundreds of millions of teens have been enrolled in Teen Accounts across Instagram, Facebook and Messenger since 2024.

What the announcement does not establish

Meta’s post gives no figure for how many teens will be affected beyond “millions”, and the pages we could read offer no accuracy rates, error rates or details on how wrongly flagged adults can correct a decision. Those are the numbers regulators and privacy groups usually ask about with automated age estimation, and they are Meta’s claims to substantiate.

The method is also partly described elsewhere. The May post says a separate visual-analysis signal, which scans photos and videos for cues about apparent age, was available only in select countries at that point. Meta said it is not facial recognition; a correction dated July 31, 2026 clarifies that it does not use biometric data or create facial fingerprints. The new post, as far as we could read, does not say whether that signal is part of the worldwide Instagram rollout, so readers should not assume it is.

For teens and parents, the practical effect is that an adult birthday entered at sign-up no longer guarantees an adult experience on Instagram. Teens in the newly covered regions may find their accounts switched to the restricted settings without having asked. For adults whose accounts are misjudged, the post as summarized gives no new detail on appeals. The question of how well one app-store signal can replace each platform’s own estimation also remains open, since the Apple signal is voluntary and Meta only cites Australia as an example.

Meta’s post is dated November 9, 2026 on its newsroom listing, which also shows November 8. The Verge reported the expansion as well, but this article is based on Meta’s own text.