AI Defeated by Eyebrow Pencils as Kids Outsmart Online Age Checks
In a way, this information is useful for anyone trying to understand areas in which digital safety fails. According to a report by Internet Matters, 46% of children said age checks are easy to bypass, and 32% admitted to doing it. UK's Online Safety Act assumes platforms can meaningfully separate adults from minors with stronger checks, yet children are treating those checks the way they treat a padlock on a garden gate: as an obstacle to climb, duck under, or simply decorate around.
Kids are reportedly using fake birth dates too, along with shared accounts, altered photos, someone else’s ID, and even videos of video game characters to confuse facial recognition systems, while some parents have admitted helping them along. In other words, new and cutting-edge systems are being bested by old, unsophisticated tactics.
That is a problem for the law’s defenders because the Online Safety Act was supposed to work through friction, by making verification annoying enough that children would stop at the door, and make the platforms responsible if they didn’t. But as most of us know, friction cuts both ways; it can slow down bad actors, yet it can also push ordinary users into workarounds that are admittedly clumsy, privacy-invasive, and sometimes laughably easy to defeat.
Honestly, if a drawn mustache can convince a system that a child is 15, then the technology is not so much verifying identity as reading cosplay. Even where the checks work better, they create fresh tradeoffs: more ID uploads, more biometrics, more data held by third parties, and more distrust from users who never wanted to hand private information over to the web in the first place.
This is all to say that the enforcement and techniques used to uphold the law is well-meant, but highly flawed and poorly executed. Children are encountering harmful material online all the time and yet platforms have been too casual for too long about who gets in and what they see.
