Meta Denies NameTag Face Recognition Exists Despite Code and CTO’s Detailed Description
Meta executives claim the NameTag face recognition feature does not exist, even though WIRED found its code in the Meta AI app and the company’s CTO described its functionality at length in a podcast.
Meta has spent weeks arguing that its in-development NameTag face recognition system for smart glasses does not exist, even after WIRED reported in June that the app Meta AI—downloaded tens of millions of times—contained robust but inactive code for the feature. Meta removed the code the next day.
Andy Stone, Meta’s vice president of communications, called WIRED’s report “intellectually dishonest” on X, while CTO Andrew Bosworth later described NameTag in detail on a podcast, saying it would identify people the user had already met while wearing the glasses. “That’s what we call a NameTags feature,” he said, adding it “would be a great feature.”
Meta stresses that Bosworth spoke conditionally—the feature does not yet exist. However, WIRED’s analysis, confirmed by two independent experts, found a technically functional face recognition system within the app. It converted faces into unique numerical faceprints and could match them against a database stored on the user’s device, populated by Meta’s servers. This local approach is designed to comply with state biometric laws like Illinois’ BIPA and Texas’ CUBI.
Legal precedents vary: some courts have ruled that on-device data is still possessed by the company; others have found it is not, especially if the feature is optional and data never leaves the device. Meta declined to say whether NameTag would be opt-in, whether faceprints leave the device, or why it licensed third-party face recognition software.
Strip away semantics: Meta built a face recognition system, distributed it to millions of phones, and a top executive praised it publicly. Whether NameTag “exists” remains a matter of definition.
