Meta Deletes Face-Recognition System From Its Smart Glasses App After WIRED Report
Meta removed unreleased face-recognition code from its Meta AI glasses app a day after WIRED reported it had been embedded in software on more than 50 million phones.
Intelligence analysis by GPT-5.4 Mini
WIRED says Meta stripped face-recognition libraries, a NameTag process, and local storage code from the Meta AI app after reporting it had quietly bundled the system into software for its smart glasses. The episode raises fresh questions about how far the company had gone with biometric features before publicly launching them.
Meta had a secret face-matching tool inside its glasses app, like hiding a name-checking machine in a toy before showing it to anyone. After WIRED found it, Meta removed most of the machine, but the story raises worries about what else could be tucked away in phone apps.
Analysis
What changed
WIRED says the latest version of Meta AI removes the code it previously found tied to an unreleased face-recognition system Meta internally called NameTag. The updated app no longer includes the face-recognition libraries, the recognition workflow, the “Person recognized” alert, or the folder meant to store cropped images and biometric signatures for faces the app could not identify.
Why the earlier report mattered
Before the removal, WIRED reported that Meta had quietly embedded substantial pieces of NameTag into the companion app for its smart glasses, which is installed on more than 50 million phones. The feature was described as turning faces seen by the glasses into unique biometric signatures, then comparing those signatures against a database stored on the user’s device. WIRED also reported that faces the system could not identify were cropped, indexed, and stored locally for later processing.
Meta’s response and the remaining questions
Meta had previously said the feature did not exist. After WIRED’s report, company spokespeople again pushed back, with one calling the reporting misleading and another saying Meta had nothing new to add. WIRED says Meta did not answer questions about whether it had already built a face-profile database, how long the app would retain biometric data, whether anything would be sent to Meta’s servers, or whether users would be able to opt in or opt out.
What the code removal does and does not prove
The removal shows Meta backed out the code, but it does not explain why the code was shipped in the first place or whether the project is dead. A few remnants still remain, including an internal debug label and a dormant link for a recognized person’s profile, suggesting parts of the system were built and then partially removed.
Broader implications
Kade Crockford of the ACLU of Massachusetts argued that the episode shows why consumer privacy needs stronger legal protections and stronger enforcement. The story lands as a warning about biometric features being developed inside mainstream consumer apps before the public gets a clear say.
Key points
- WIRED says Meta removed face-recognition code from the Meta AI app after its report.
- The code was tied to an unreleased system Meta internally called NameTag.
- WIRED says the app had code for biometric matching and local storage of unrecognized faces.
- Meta previously denied the feature existed and did not explain why the code was removed.
- Privacy advocates argue the incident shows why stronger consumer privacy laws are needed.
If the removal is permanent, it could mean Meta is backing away from a risky biometric feature before it reaches users. The report may also push companies to be more transparent about face-recognition tools and privacy controls before shipping them.
The code removal does not answer whether Meta still intends to build NameTag or whether similar biometric systems are already under development elsewhere. The episode also suggests that sensitive face-processing code can be embedded quietly and only removed after outside scrutiny, which keeps privacy risks alive.
