Meta abandons facial recognition for smart glasses after Wired investigation

Meta has abruptly scrubbed hidden code from its smart glasses following a media investigation which revealed it planned to add facial recognition to its wearable devices. 

An investigation published by Wired disclosed that Meta had quietly embedded an unreleased, dormant facial recognition system internally code-named “NameTag” into its Meta AI companion app.

The software, which supports the company’s popular Ray-Ban and Oakley smart glasses, has already been downloaded onto more than 50 million smartphones worldwide.

According to code audits, the NameTag infrastructure was remarkably robust despite being inactive. The software contained three live AI models designed to detect passersby, crop their faces, and convert them into unique biometric signatures known as “faceprints.”

These faceprints would then be cross-referenced against a localised database to instantly push a “Person recognized” notification directly to the wearer’s lenses. Within 24 hours of the report’s publication, Meta quietly pushed an emergency software update that deleted virtually all code libraries, pending storage folders, and interface links tied to NameTag.

While the swift deletion suggests Meta recognised the extreme regulatory and public relations tightrope it was walking, the exposure triggered an aggressive, highly public backlash from the company’s upper management.

Meta leadership was deeply frustrated by the reporting, primarily because the facial recognition feature was entirely dormant and had never actually gone live for a single consumer.

Taking to social media, Meta’s Vice President of Communications, Andy Stone, launched a furious defence, blasting the investigation as “intellectually dishonest” and “pure advocacy-driven click bait.”

Stone criticised the publication for waiting until the fourth paragraph to clarify that the feature was entirely disabled, and taking until the 16th paragraph to note that the code was merely exploratory. He reiterated that Meta has no active plans to launch the scanner and is absolutely “not building a central face database.”

Meta’s Chief Technology Officer, Andrew “Boz” Bosworth, also lashed out publicly, calling the reporting “incredibly misleading” and “absolutely dishonest.” However, he stopped short of disputing any of the specific technical code findings.

The corporate anger underscores the immense sensitivity surrounding biometric data at Meta. The company previously built one of the world’s largest facial recognition systems on Facebook before being forced to dismantle it in 2021 due to severe privacy scrutiny.

Tech industry analysts noted that while the feature was inactive, discovering operational face-scanning machinery pre-loaded onto millions of consumer phones validates long-held anxieties among digital rights groups regarding the covert surveillance potential of everyday wearable technology.