Meta is arguing over the verb “exist” while a face recognition system for its smart glasses sits at the center of a privacy fight that users did not ask to beta test.
WIRED reported that Meta placed inactive code for a system called NameTag inside Meta AI, the companion app for Meta Ray-Ban glasses. The app has been downloaded tens of millions of times, according to WIRED. The code was not available as a consumer feature, and ordinary users could not turn it on. Meta removed the NameTag code from the app the day after WIRED’s June 4 report, WIRED said.
Andy Stone, Meta’s vice president of communications, rejected WIRED’s framing on X, writing that the feature “doesn’t exist.” Meta CTO Andrew Bosworth later called the reporting “incredibly misleading” and “absolutely dishonest” in a reply to Stone. WIRED said Meta did not answer its questions about what, specifically, was misleading or dishonest.
WIRED’s technical account is more concrete than Meta’s public semantics. The outlet said NameTag code appeared in the Meta AI app as early as January, and that the main pieces were present by May. The New York Times reported in February that Meta had been working on NameTag face recognition. A researcher using the name Buchodi reviewed the code at WIRED’s request and used the system to identify a photograph of Michel Foucault, according to WIRED.
How Meta described NameTag
Bosworth described the idea during the July 8 episode of The Most Interesting Thing in AI, hosted by Nicholas Thompson, CEO of The Atlantic and a former WIRED editor in chief. Bosworth said NameTag would identify someone the glasses wearer had met before, such as a person who introduced themselves while the user was wearing glasses or someone the user asked the system to remember.
Meta spokesperson Ryan Daniels told WIRED there was no contradiction between Stone saying the feature does not exist and Bosworth describing it. Daniels emphasized that Bosworth said it “would” be a good feature. Daniels said Meta is exploring the idea, that it is not available to consumers, and that the company is not building a capability that connects glasses to a central database of people.
That last point has become Meta’s favored fence line. WIRED said its app analysis did not show a central lookup service. Instead, it found a design in which faces captured by the glasses could be converted into numerical faceprints and compared with a database stored on the user’s device, with that database populated by Meta’s servers.
The legal fight is about control
The location of the face data matters because state biometric privacy laws can require explicit consent before companies capture or use face recognition data. Bosworth mentioned Illinois’ Biometric Information Privacy Act and Texas’ Capture or Use of Biometric Identifier Act during the podcast, according to WIRED.
Meta has history here. Facebook ended its Tag Suggestions face recognition feature in 2019 after a $5 billion Federal Trade Commission privacy settlement, which included face recognition issues, and before a $650 million Illinois settlement over Tag Suggestions, according to WIRED.
Courts have split on whether on-device face data keeps companies clear of biometric privacy claims. WIRED cited a 2021 federal ruling allowing a BIPA case over Apple Photos to proceed, with the judge finding that a company could plausibly possess faceprints stored on users’ own devices. WIRED also cited rulings favoring Apple and Samsung where courts found the companies did not possess or access locally stored face data.
WIRED said Meta declined to answer questions in June about whether NameTag would be opt-in, how long it would retain faceprints and cropped images, whether data could leave the device, and why it licensed third-party face recognition software. Those are the boring questions. They are also the ones that decide whether smart glasses become a memory aid, a consent problem, or both.
This story draws on original reporting from WIRED.
