Meta Platforms announced on Monday that it would expand its artificial intelligence-driven age-estimation technology globally, a move designed to place millions more teenagers on Instagram into its Teen Account protections even when those users have supplied birth dates indicating they are adults. The decision comes as the social media giant confronts a growing wave of litigation alleging that its platforms cause harm to young users, and it represents one of the company’s most concrete attempts to demonstrate that it is taking adolescent safety seriously.
The Teen Account framework, which Meta has made available across Instagram, Facebook and Messenger, incorporates a set of built-in safeguards that automatically restrict who can contact teenage users, limit the content they are exposed to and cap the amount of time they can spend on the applications. These protections are switched on by default for accounts identified as belonging to minors, and the company has positioned them as a meaningful barrier against some of the most commonly cited risks of adolescent social media use, including unwanted contact from strangers and compulsive scrolling. By extending AI-based age inference to users who have misrepresented their age, Meta is effectively acknowledging that self-reported birth dates have long been an unreliable gatekeeping mechanism, a point that critics and regulators have raised for years.
Under the expanded rollout, parents in the United States will begin receiving Instagram notifications this month explaining how their teenagers are being placed into age-appropriate protections. The notifications are intended to give families greater visibility into the classification process, though they also raise questions about how parents will respond if they believe their child has been incorrectly categorized. Meta has not indicated whether it will offer a straightforward appeals pathway for users who are placed into Teen Accounts despite being adults, a detail that could prove significant as the technology reaches a broader global audience.
The company is also broadening the inputs it uses to determine age. In Australia, Meta has begun incorporating Apple’s age-range signal as an additional method for estimating how old a user is, supplementing the AI analysis of profile activity and behavioral patterns. The use of platform-level signals from device manufacturers reflects a wider industry shift toward shared responsibility for age verification, though it also raises privacy considerations that are likely to attract scrutiny from regulators in Europe and elsewhere. Meta has said it is continuing to roll out the technology on Facebook and will begin launching similar capabilities globally on Threads later this year, suggesting that age estimation will become a standard feature across its entire family of applications rather than a limited pilot.
The timing of the announcement is difficult to separate from Meta’s legal exposure. The company faces multiple lawsuits alleging that its platforms harm teenagers, and in August it reached settlements worth up to eighteen billion dollars with nearly every United States state and territory to resolve claims brought by attorneys general over the platforms’ effects on young people. As part of those settlements, Meta agreed to improve how it identifies teenagers and children under the age of thirteen, commitments that the expanded age-estimation rollout appears designed to fulfill. Last month, the company also settled with a teenage plaintiff during a trial in Los Angeles that tested claims Instagram and Facebook were deliberately designed to be addictive and contributed to mental health problems and an eating disorder. That case, and the broader litigation landscape, has made adolescent safety one of the most consequential reputational and financial issues confronting the company.
From a strategic standpoint, the expansion of AI age estimation serves several purposes at once. It gives Meta a tangible response to critics who argue the company has been slow to act, it provides evidence of good-faith compliance with settlement obligations, and it creates a defensible narrative for investors who worry that regulatory pressure could constrain growth in the company’s most lucrative markets. The technology itself, however, is not a complete solution. AI age inference is probabilistic rather than definitive, and any system that relies on behavioral signals will produce false positives and false negatives. Adults may find themselves placed in restricted accounts, while some determined teenagers may find ways to evade detection. The effectiveness of the approach will ultimately depend on how accurately the models perform across different languages, cultures and usage patterns, and on whether Meta is transparent about error rates and correction mechanisms.
