Google is rolling out a machine-learning age-estimation system in Canada that determines whether users are likely under 18. The company says it isn’t collecting new data. It’s applying information it already holds to a new purpose.
Here’s how it works. Rather than asking how old you are, Google analyzes signals from your existing account activity. That includes what you search for, which apps you use and the kinds of videos you watch on YouTube.
Users flagged as under 18 may find account restrictions switched on. Those can include YouTube break and bedtime reminders, changes to content recommendations, limits on age-sensitive material and restrictions on personalized ads.
Get misidentified and the burden falls on you. Proving your age may require a government ID, a selfie or a credit card.
The rollout has already started
The system reaches Canada this month, and it isn’t a Canada-only experiment. YouTube already uses similar age-estimation technology in the US, so more users will likely encounter the same process.
Google’s stated goal is making its platforms safer for younger users without forcing everyone through manual verification. The system will misfire sometimes. When it does, adults get to jump through extra hoops to prove they’re old enough for features they’ve been using all along.
The questions worth sitting with are the bigger ones. How comfortable are we with Google inferring our age from how we use its products? How comfortable are we with Google deciding what’s appropriate for our children? And how much of this should a corporation be deciding at all?