Meta Shipped Face Recognition Code to Millions, Deleted It in a Day

Our Take

Credit to Verified Privacy VPN for digging into the code itself instead of taking Meta's denials at face value. The story here isn't that Meta got caught — it's how close NameTag was to shipping. This wasn't some speculative research sketch sitting in a lab. Researchers found a system built to detect a stranger's face through smart glasses, crop it, generate a biometric signature, and quietly store it on the wearer's phone for future recognition. That's a pocket-sized identification machine, distributed to tens of millions of devices before anyone outside the company knew it existed.

The 24-hour scrub-and-deny routine should alarm you more than reassure you. Deleting code after a reporter asks questions doesn't undo the fact that the architecture was built, tested, and ready to flip on. It tells you the appetite for real-time face-matching is there, and that the only thing standing between "in development" and "active on your face" is whether anyone's watching. Corporate promises not to ship a feature are not a policy. They're a mood, subject to change the moment press attention fades.

This is the same pattern we track with license plate readers: quiet deployment, minimal disclosure, and public pushback treated as a PR problem to manage rather than a civil liberties issue to answer. Face recognition and ALPR networks are two halves of the same surveillance infrastructure — one identifies your face, the other identifies your car, and together they can map your entire life without a warrant. Check our map to see what's already watching your streets, and visit take action if you want to push back before the next NameTag ships quietly to your phone.