Meta’s new age verification tool is on the lookout for chiselled jaws.
What happened: Meta has begun using AI in select countries to read faces and determine if a Facebook or Instagram user is under 13, and thus should be removed. The tool will scan photos and videos, analyzing visual clues like height and bone structure to make a decision.
Meta has stressed that this is not facial recognition as it only looks at general clues to determine an approximate age, not someone’s identity. It will bolster existing AI tools that determine user ages by looking for clues in bios, comments, and more.
Why it matters: Facial analysis has been one of the main ways of clocking minors as more governments pursue youth social media bans or age verification for porn and gaming platforms. With Canada potentially next in line, we could soon see more age-check tools like Meta’s.
That said, the ‘Big Tech cataloguing children’s bone structure’ of it all does leave us feeling icky. And on top of that, many of these systems can be pretty easily duped.
A recent report about the U.K.’s Online Safety Act found that kids were bypassing age verifications by submitting other people's faces, using images of video game characters and, in one instance, drawing a fake moustache with eyebrow pencil.
Are there any other options? One alternative is bringing age verification to the operating system level, where underage users are locked out by their device instead of an app or site. While this avoids some of the moral and practical pitfalls of facial scans, there are concerns about concentrating even more power with device makers like Apple and Google.—QH
