Meta's NameTag Face Recognition: What Exists, What Doesn't
Executives insist the feature doesn't exist while simultaneously describing how it works, creating confusion over face recognition in Ray-Ban smart glasses.
Meta has spent weeks making contradictory statements about NameTag, its face recognition system for Ray-Ban smart glasses, leaving observers confused about whether the technology exists at all.
The confusion stems from a fundamental question: Does a software feature exist if its code has been deployed to millions of devices but users cannot activate it?
The timeline of NameTag
According to WIRED's original investigation, Meta included functional but inactive code for NameTag in its Meta AI companion app, which has been downloaded tens of millions of times. The code appeared as early as January 2026, with core components present by May.
When WIRED reported on NameTag in June, Meta's communications VP Andy Stone responded on X that "the feature doesn't exist," questioning how the company could answer questions about something non-existent. Meta removed the NameTag code from the app the following day.
Yet on July 8, Meta CTO Andrew "Boz" Bosworth described NameTag in detail on a podcast, explaining it would recognize "somebody you met in person with your glasses on who introduced themselves" and allow users to save that person's identity for future recognition.
The semantic defense
Meta spokesperson Ryan Daniels emphasized to WIRED that Bosworth used conditional language—saying NameTag "would be" a great feature, not that it "is" or "will be." The company maintains there is no contradiction between Stone's denial and Bosworth's detailed description.
A researcher reviewed the code at WIRED's request and successfully used the NameTag system to recognize a photograph of philosopher Michel Foucault, demonstrating the technology was technically functional.
The legal considerations
Bosworth stressed that NameTag would not use a "central database," though neither WIRED nor the podcast host had suggested it would. Instead, the system converts faces into numerical "faceprints" compared against databases stored on individual users' devices, populated by Meta's servers.
This architectural distinction matters for legal compliance. State laws like Illinois' Biometric Information Privacy Act and Texas' biometric identifier law restrict face recognition deployment. Meta previously abandoned Facebook's automatic face recognition feature after a $5 billion FTC settlement and a $650 million settlement with Illinois.
Courts have split on whether on-device face recognition violates these laws. A federal judge allowed a class action against Apple's Photos app to proceed in 2021, while other courts ruled that Apple's Face ID and Samsung's photo app didn't "possess" data stored solely on devices.
Why it matters
The semantic debate obscures a substantive reality: Meta has built, tested, and deployed face recognition technology to millions of phones, then removed it after public scrutiny. The company's insistence on conditional language suggests it's carefully navigating both public perception and legal exposure while keeping its options open. For business leaders considering similar technologies, Meta's approach illustrates how architectural choices—like on-device versus centralized processing—can create legal ambiguity that may or may not withstand court challenges.
Unanswered questions
Meta declined to answer WIRED's questions about whether NameTag would be opt-in, how it retains faceprints, why it licensed third-party face recognition software, or whether data ever leaves users' devices.
The facts remain straightforward: Meta designed a face recognition system, placed it on millions of phones, and a senior executive praised it publicly. Whether that constitutes "existence" depends on one's definition.
These details were first reported by WIRED.
This is an original analysis by the Omega editorial team. Source reporting: WIRED.
Want systems like this working for your business?
Book a Call
