Follow The Library Media
The Library Media — Science, Technology & Health · AI News & Intelligence Desk
Your Birthday No Longer Decides: Meta’s AI Age Checks Go Global on Instagram
By Eleanor Ever Loyal
Eleanor Ever Loyal
Eleanor Ever Loyal — The Library Media’s Lead Counsel and National Desk lead.
Meta is expanding its AI age-estimation technology globally on Instagram, using signals from across an account to identify users it suspects are teenagers—even when those users have entered an adult date of birth.
Accounts identified by the system as likely belonging to teens can be placed into Instagram’s Teen Account protections. Meta says the global expansion will bring those built-in protections to millions more teenagers worldwide.
The change shifts an important boundary in online age verification: the birthday entered at account creation is no longer necessarily the final word.
Meta says its age-estimation system examines contextual clues across a user’s profile and activity. Those signals can include birthday-celebration posts and references to a user’s school grade in posts, comments, biographies and captions. The company also uses visual analysis of photos and video.
Meta says the system does not create facial fingerprints and does not identify who appears in a photograph.
The company’s age estimates are probabilistic, however—not definitive determinations of a person’s age.
That distinction is central to the governance questions surrounding the expansion.
Meta’s Teen Accounts are private by default and place limits on who can contact teenage users and what content they can see. In some cases, the accounts also include time limits. Direct messages are restricted to people a teen follows or with whom the teen is already connected.
The new system is designed to apply those protections when Meta’s technology concludes that an account is likely being used by a teenager despite an adult birthday being listed.
Meta has tied the expansion to its agreement with 52 bipartisan attorneys general to strengthen the identification of teen users. The rollout also comes as Meta faces numerous lawsuits alleging that its platforms harm teenagers, according to TechCrunch’s framing of the broader legal backdrop. Countries are also adopting new online-safety laws aimed at younger users.
In Australia, Meta says it can additionally use Apple’s optional App Store age-range signal where that information is available.
The global rollout, however, leaves several important measurements undisclosed.
Meta has not published accuracy or error rates for the age-estimation system. It has not provided per-country effective dates for the expansion. And it has not publicly stated how many accounts have already been moved into Teen Account protections as a result of AI age estimation.
Those omissions matter because the potential errors run in both directions.
An adult incorrectly estimated to be a teenager could be placed into a more restricted account. A teenager incorrectly estimated to be an adult could remain outside protections intended for younger users.
Meta has described the technology as one tool for improving protections rather than as a system that can establish age with certainty. But without published accuracy figures, outside observers cannot assess from the company’s announcements how frequently either type of error occurs.
The broader policy challenge is increasingly moving from whether platforms should protect young users to how they determine who those users are.
Self-reported birthdays are simple to collect but can be inaccurate. AI-based age estimation may offer platforms additional signals, but it introduces a different set of questions involving reliability, transparency, privacy and the consequences of an incorrect classification.
Meta says its system looks at the context of an account rather than relying solely on the date a user entered. The company has specifically emphasized that its visual analysis does not create facial fingerprints or identify people appearing in images.
The technology’s next expansion is already planned beyond Instagram. Meta says age-estimation technology will continue rolling out on Facebook, while similar technology is scheduled to launch globally on Threads later this year.
For now, Instagram is becoming a test of a broader proposition in platform governance: an account’s stated birthday may be treated as one piece of information, rather than the deciding fact, when a company determines what protections should apply.
What remains less clear is how accurately that judgment works at scale.
Reporting based on Meta’s announcements and TechCrunch dispatches.
— Eleanor Ever Loyal, AI News & Intelligence Desk
Eleanor Ever Loyal is The Library Media’s regulation and governance correspondent.
From The Library — the estate of our family.
