Meta Rolls Out AI Age Estimation Worldwide to Move Suspected Teens Into Teen Accounts

In a blog post, the company said the expansion would put "millions more teens" into Teen Account protections. It gave no exact figures. The same technology is still being rolled out on Facebook, and Meta plans a similar global launch on Threads later this year. Before this week the system was running in more than 20 countries, Al Jazeera reported.

The expansion follows the settlement Meta reached in August with attorneys general across almost every US state and territory. Meta agreed to pay up to $18 billion to settle claims that its platforms harm teenagers, and committed to improving how it detects teens and children under 13. In its post, Meta tied the new investment directly to that agreement with 52 bipartisan attorneys general. It said it was building stronger tools to catch accounts that may belong to under-13s and to spot 13- to 17-year-olds who give false ages.

According to Meta, the system looks across a profile for contextual clues, such as birthday celebrations or mentions of school grades, in posts, comments, bios and captions. It also uses visual cues from photos and videos, including a person's height or bone structure, TechCrunch reported. Meta told Al Jazeera this does not extend to facial recognition.

Teen Accounts are private by default. They limit the content teenagers see, restrict direct messages to people they already follow or are connected to, and cap time spent in the apps. Parents can supervise them. In countries that ban social media below a certain age, a flagged account may be shut down entirely. Al Jazeera also reported that Meta will add US-specific limits for under-18s: a two-hour daily cap and a curfew from midnight to 6am. The company's blog post does not mention those limits.

In Australia, Meta now also uses an Apple feature that lets parents assign an age range to a child's device and passes that range to apps when they are downloaded. Meta repeated its call for laws that would make app stores verify ages and get parental consent. It argues that a single verified signal at the device level would be more effective and more private than each app collecting age data separately. Under its proposal, developers' own estimates would act as a backstop.

Outside researchers are not convinced. Sonia Livingstone, a professor at the London School of Economics who directs the Digital Futures for Children Centre, told Al Jazeera there was an "absence of clear evidence" that the system keeps young children off the platforms or puts teenagers into Teen Accounts. She said age assurance "poses a privacy risk to all users, not just children". She added that without independent audits, "it is hard to see why we should trust them with our personal data".

Australia's experience shows the limits of age checks. Six months after the country banned under-16s from social media in December, its eSafety regulator found that more than 80 percent of preteens and young teens were still using the platforms, Al Jazeera reported. Over the same period Meta closed 462,000 Instagram accounts to comply with the ban, but eSafety found that more than 69 percent of underage users still had their accounts.

US parents will start seeing an in-feed video on Instagram this month that explains how Meta assigns teens to age-appropriate settings. Threads is next in line for the age-estimation system.