Cliquez ici pour lire en français
Meta is expanding its AI-based age verification system worldwide. The aim is, in particular, to catch young users who get around access restrictions by entering a fake birthdate on Facebook, Instagram or Threads. For years, the company has faced heavy criticism over the safety of younger users on its platforms. To keep out children under 13, who are officially barred from its services, Meta has gradually tightened its controls. One of them is an age check based on visual analysis, first rolled out in the United States and Europe on Facebook and Instagram. It is now set to reach all relevant markets, with Threads folded in as well.
Catching users who lie about their age đ
In practice, the system tries to identify minors who entered a false birthdate to pass as adults and get accounts without age-specific restrictions. To estimate usersâ ages, the AI built by Meta and its partner ConnectSafety does not, according to Meta, rely on facial recognition aimed at identifying people. Instead, it analyzes their appearance to estimate an age range.
But the check doesnât rest on images alone. The AI also looks at other information on profiles, including bios, posts, comments and captions. That lets it pick up clues that could reveal a userâs real age, such as birthday mentions.
Teen Accounts to rein in minors đ„
When the system detects that a 13-to-17-year-old is using an adult account without the protections designed for their age group, Meta plans to move them automatically to a âTeen Accountâ. These profiles come with stricter safety settings, notably on who can contact the account holder. The platform also applies tougher filtering to limit exposure to inappropriate content and offers tools to manage screen time.
Accounts identified as belonging to children under 13, by contrast, will be disabled. Meta says it wants to enforce its terms of service better and leave fewer ways around age restrictions.
Age checks built into app stores đ±
Beyond its own tools, the company also wants age verification to be more centralized at the app-store level. In some countries with stricter rules on social media, Meta already uses age-range information passed on by operating systems.
That is the case in Australia, where social media is banned for under-16s. On iPhones set up with parental controls, Meta can rely on the age-range signal provided by iOS.
For Meta, centralized checks at the app-store level would be more effective and better for user privacy. They would also spare every app from building its own control system.
With this global expansion of AI age checks, Meta is looking to tighten minorsâ access to its platforms. Whether it works will depend largely on how accurately the AI estimates usersâ ages, and on how well it limits mistakes in identifying profiles.
What do you make of AI age checks on social media? Tell us in the comments.
đ± Get our latest updates every day on WhatsApp, directly in the âUpdatesâ tab by subscribing to our channel here âĄïž TechGriot WhatsApp Channel Link đ
