Meta expands AI age detection globally to strengthen teen protections

Meta is expanding its AI-powered age-estimation technology globally to identify teenage Instagram users who may have registered with an adult birth date, as part of efforts to strengthen age-appropriate protections across its platforms.

Unveiled on 9 November, the rollout will see millions more teenagers worldwide automatically placed into Instagram's Teen Accounts, which restrict who can contact them, the content they can access and the time they spend on the platform.

The technology is designed to identify users suspected of being teenagers, even when their registered birth dates suggest otherwise. According to Meta, the global expansion follows encouraging results from its earlier deployment of the technology.

Don't miss: Meta brings personal AI agent Muse to Singapore

The company is also continuing to roll out the technology on Facebook, with plans to introduce similar capabilities globally on Threads later this year. As part of the initiative, Meta will begin showing parents in the United States an in-feed video on Instagram this month, explaining how the platform determines users' ages and how parents can help ensure their teenagers' account information is accurate.

The video will direct parents to an article by Larry Magid, internet safety advocate and president and CEO of ConnectSafely, outlining Meta's approach to age assurance, including efforts to prevent children under 13 from accessing its platforms.

Meanwhile, in Australia, Meta has begun using Apple's age-range signal as an additional method of determining users' ages. The feature allows parents setting up a smartphone for their teenager to share an age range through the app store, which can then be passed to developers such as Meta when an app is downloaded.

However, Meta said the voluntary nature of the system could result in inconsistent protections across platforms, reiterating its support for legislation requiring app stores to verify users' ages and obtain parental consent.

The company argued that a single age-verification signal shared across apps would be more effective and privacy-preserving than requiring individual platforms to conduct their own checks. The expansion also follows Meta's recent agreement with 52 bipartisan attorneys general across US states, territories and the District of Columbia to strengthen how it identifies underage users.

Under the agreement, Meta is investing in technology to detect accounts potentially belonging to children under 13, while improving its ability to identify teenagers aged 13 to 17 who may have provided an inaccurate birth date.

The company also called on industry peers TikTok and YouTube to adopt similar age-assurance practices, arguing that inconsistent approaches across platforms could undermine efforts to protect younger users.

The move also builds on Meta's broader efforts to strengthen safety across its platforms, including its recent introduction of AI-powered tools to detect ads potentially linked to child sexual exploitation material. The company recently introduced additional AI tools to its ad review process to identify advertising that may direct users towards child sexual exploitation material hosted outside its platforms. The measures follow its discovery of tactics involving seemingly harmless ads containing hidden signals that direct users towards illegal content elsewhere.

As part of the measures, Meta has introduced large language model (LLM) detection to identify what it calls "signposting", where seemingly benign advertising content is used to direct users towards harmful material. The company has also expanded its investigations into related activity, disabling accounts found to have violated its policies and blocking links associated with the activity.

Related articles: 
India orders Meta to take down child sexual abuse ads   
Meta gives parents visibility into teens’ AI use as scrutiny grows  
X, TikTok under watch as SG joins global push to protect young social media users