Biometric Data Privacy: Walmart Keeps 512 Face Numbers

When you walk into certain Walmart stores, a camera can turn your face into a string of numbers, and then throw away the photo it started from. The picture disappears. The numbers can stay for years. And most of us have no idea the two things follow completely different rules.

If you've ever unlocked your phone with your face,

If you've ever unlocked your phone with your face, this already touches your life. Biometric collection isn't science fiction anymore, it's happening at the checkout line. And I get why that feels unsettling. When a store's privacy notice says it collects your "face geometry," most people either glaze over or quietly panic. But once you understand what's actually being kept, you stop feeling powerless. So let's answer one question, when a company deletes your photo, does that mean your face is really gone?

The short answer is no. And to see why, you have to understand that your face on camera creates two separate things. The first is the photo, the raw image. The second is something called a biometric template. A template is basically a mathematical map of your face, turned into numbers. A good system does something clever with those two objects. It looks at you in real time, confirms you're a living person, builds the number-map, and then deletes the original photo. The photo and the template don't have to live together. They can be deleted on totally different schedules.

Picture a cashier checking your I.D. to confirm you're old enough to buy something. They look, they verify, and they hand the card back. They don't photocopy it for the register. But a biometric system flips that. The measurement, the thing the cashier just did in their head, that's what gets stored forever. The card itself gets tossed. A store can delete the camera footage after thirty days and still hold onto the mathematical profile it built from your face.

What's actually inside that profile

So what's actually inside that profile? According to biometric researchers, face templates built with modern A.I. tend to boil your face down to around five hundred and twelve numerical values. Those numbers describe things like the distance between your eyes, the shape of your nose, the size and position of your ears. And this is the part people miss, when a company matches you later, it's not comparing two photos side by side. It's comparing those numbers.

The system measures how close your five hundred and twelve numbers are to a stored set, using a math distance calculation. Closer numbers mean a higher similarity score, which means a more confident match. For a security team, that's efficient. For you, it means the thing identifying you isn't your picture, it's a numerical fingerprint you never see.

Now, here's the myth I really want to correct. Most people assume that if a store says "we delete photos after thirty days," their biometric data is gone too. And that's a completely reasonable assumption, both things came from the same camera, so surely they die together, right? Companies also use soft language like "we don't keep your photos," which feels reassuring and quietly skips the template question entirely. But the photo and the template run on separate clocks. Walmart's own policy, for example, says it destroys biometric data within three years of your last interaction, a timeline that has nothing to do with when the photos get deleted.