Instagram is expanding AI age estimation worldwide to identify accounts it suspects belong to teens and apply Teen Account protections.
Why Instagram is estimating age
Meta introduced Instagram Teen Accounts in 2024 and says they were available in every country where Instagram operates by June 2025. They limit who can contact teens, what content they see and how much time they spend on the app. Under-16s need a parent’s permission to loosen some protections, but those defaults depend on Instagram identifying teens.
Meta began using AI to place suspected teens into Teen Accounts in the United States in 2025, then tested it in the UK, Canada and Australia. In related age-check work, Meta has said its AI examines profile details such as birthday celebrations and references to school grades. It also described visual analysis, later clarifying that it does not create facial fingerprints or identify people in photos, but looks for general visual patterns. The global announcement does not specify which signals are used in each region.
Accuracy and appeals remain unclear
Age estimation is not proof, and Meta’s November announcement gave no error rate or independent test results. Meta previously reported that more than nine in ten active teen accounts remained in protective settings. That measures whether teens keep those settings, not whether AI correctly distinguishes teens from adults.
Meta has said users can change their settings if it makes a mistake, but the new announcement does not describe a formal appeal process or how an adult can challenge a mistaken classification.
Legal pressure and privacy trade-offs
In August, state attorneys general announced a proposed settlement with Meta worth up to $17.1 billion, resolving claims that the company designed features that harmed young users. The announced reforms included age-assurance measures. Meta says its global expansion builds on a separate agreement with 52 attorneys general to improve how it identifies teens and children under 13. This does not show that litigation is the sole reason for the rollout, but it makes age detection part of a broader effort to show protections are applied proactively.
More teens who evaded safeguards may receive stricter defaults without a parent spotting a false birthday. But adults may be placed in a teen experience by mistake, and profile information and images can become clues used to estimate age. The UK Information Commissioner’s Office advises choosing methods in proportion to privacy risks and collecting only what is needed. Meta has not clearly disclosed the system’s accuracy or how people can correct mistakes.
What Meta plans next
Meta says US parents will start receiving Instagram notifications in November explaining how teen protections are assigned. It is also using Apple’s age-range signal as an additional input in Australia, continuing the rollout on Facebook and plans to bring similar technology to Threads later this year. The next test will be whether Meta publishes accuracy evidence and makes mistaken placements easy to correct.