Meta Removed Hidden Facial Recognition Code From Its Smart Glasses App After WIRED Revealed It

Meta Removed Hidden Facial Recognition Code From Its Smart Glasses App After WIRED Revealed It

Just 24 hours after WIRED uncovered that Meta had secretly built an unannounced facial recognition system into a mobile app used by more than 50 million smartphone users, the company pulled all related code from the platform, according to a new WIRED analysis of the app’s latest update.

The most recent update to Meta AI — the companion app for Meta’s line of smart glasses — eliminates all unused software components tied to the internal project codenamed NameTag. The version of the app published the same day WIRED released its initial investigation included multiple code libraries explicitly labeled for facial recognition functionality; none of those assets remain in the release that rolled out Friday.

Andy Stone, Meta’s vice president of communications, told WIRED on Monday that the work on NameTag was purely exploratory, adding: “No final decision has been made on what path we will take here, if any at all.”

In our initial report published Thursday, WIRED detailed that Meta had quietly embedded core components of the full NameTag system directly into the Meta AI app. While the feature was never turned on for public use, its design would allow the app to turn faces captured by smart glasses into unique biometric identifiers, commonly called faceprints, and cross-reference those against a database of faceprints stored entirely on the user’s personal device. WIRED also confirmed that any faces the system failed to identify would be cropped, indexed, and saved locally on the device for future matching attempts.

NameTag first made headlines back in February, when The New York Times reported, citing internal Meta documents, that the company was building facial recognition for its smart glasses and was evaluating a launch as early as this year. One internal memo reportedly framed a potential launch during a “dynamic political environment,” when privacy and civil liberties watchdogs would be distracted. Last week, WIRED confirmed that core infrastructure for NameTag had already been included in the Meta AI app — downloaded by millions of users — as early as January, even as Meta publicly maintained it had not made any final decision about rolling out facial recognition.

Following WIRED’s initial publication, Stone pushed back against our findings, writing that the company could not answer questions about how the system would function because “the feature does not exist.” Andrew Bosworth, Meta’s chief technology officer, also criticized the reporting, calling it “incredibly misleading” and “absolutely dishonest.”

| Got a Tip? |

|---|

| Are you a current or former Meta employee with insight into the company’s technology development? We want to hear from you. Contact the reporter securely via Signal on a non-work phone or computer at the handles dmehro.89 or dell.3030. |

Meta declined to answer 10 separate questions WIRED submitted ahead of our Thursday publication. Those questions included whether the company had already built out the face profile database NameTag would rely on, how long the app would retain photos and biometric data of unrecognized faces stored on a user’s device, and whether any of that local data would ever be transmitted back to Meta’s corporate servers.

Additionally, Meta did not answer a question about whether NameTag was being developed specifically for blind and low-vision users. The company also declined to address criticism from privacy advocates, who have warned the tool could enable stalkers and abusers to identify random strangers in public spaces. It also did not respond when asked whether the company planned to give users the option to opt in to or opt out of the system entirely.

The new update to Meta AI removes nearly every trace of the feature Meta claimed did not yet exist. The core facial recognition software is gone, along with the code that powered NameTag’s matching process and the “Person recognized” in-app alert that would have popped up when a user’s glasses successfully identified someone. The update also deletes the dedicated folder where the app would have stored cropped images and biometric signatures of faces it captured but could not identify.

Meta has not responded to WIRED’s new questions about why the NameTag code was removed, or whether these changes were scheduled to roll out before our initial investigation was published.

A handful of tiny fragments of the NameTag system still remain in the latest version of Meta AI, including an internal debug menu label and a dormant link designed to open a recognized contact’s profile. The leftover code only points to parts of the system that have been fully removed.

Kade Crockford, director of the Technology for Liberty Program at the American Civil Liberties Union of Massachusetts, says removing the code does not reverse Meta’s original choice to ship the hidden software to millions of users. They pointed to the incident as clear evidence that consumer privacy needs far stronger legal protections than what Congress has so far been willing to enact. Crockford noted that the Massachusetts House of Representatives last week unanimously passed a comprehensive consumer privacy bill that would include strict enforcement measures if passed into law, and urged other U.S. states to follow suit — particularly by including a private right of action that allows harmed users to file suit against companies for privacy violations.

“Meta’s underhanded tactics in sneaking facial recognition code into its smart glasses app show exactly why data privacy legislation needs the teeth of strong enforcement,” Crockford said. “Companies like Meta prioritize their bottom line above all else, so lawmakers need to speak the only language their C-suite actually understands.”

Updated June 8, 2026 at 3:05 p.m. EDT: This story was updated to add additional comment from a Meta spokesperson and clarify context around the company's communication with WIRED.