The code is not theoretical. A security researcher who went through the current version of Meta's smart glasses app found a complete face-recognition pipeline sitting inside it: detection models, a database to store face prints, a notification channel that fires when a known face appears, and a folder that saves images of unrecognized faces for later. When tested manually, the system identified a face and sent a notification reading "Person recognized." The feature, called NameTag, is just switched off for regular users.
Meta has confirmed the exploration is real. A company spokesperson told press that "nothing has shipped to consumers and no final decision has been made," and added that if NameTag does launch, it will not rely on a central face database. That last point is technically accurate: the current architecture stores face data locally on the device rather than uploading it to Meta's servers. But local storage does not reduce the harm to the person being identified without their knowledge.
The glasses themselves are already a hit. EssilorLuxottica, the company that makes them under the Ray-Ban and Oakley brands, sold over 7 million pairs in 2025. That is more than triple what it sold across the two preceding years combined. Meta and EssilorLuxottica are reportedly discussing production targets of 20 to 30 million units per year going forward. These are not gadgets sitting in a drawer. People wear them daily.
The privacy problem with these glasses predates NameTag. In 2024, two Harvard students built their own face-recognition tool using the same Ray-Ban glasses, a third-party facial search engine, and some off-the-shelf software. The result: they could identify a stranger on the street and pull up that person's name, home address, and phone number within about a minute. They called the project I-XRAY, published it as a warning, and did not release the code. Their point was that the hardware already enables this; only the software is missing.
Meta considered adding face recognition to the first generation of its smart glasses back in 2021 but dropped it over technical and ethical concerns. The company also shut down face recognition on Facebook that same year after a settlement over privacy violations. Then in 2024 it brought facial recognition back to Facebook and Instagram, framed as a tool to detect scam ads using real people's faces. The pattern is consistent: the technology gets shelved, then returns under a different frame.
More than 70 organizations, spanning civil liberties, domestic abuse prevention, reproductive rights, and immigrant advocacy, have formally asked Meta to drop NameTag. Their concern is specific: glasses that look like normal eyewear, worn by anyone, in any public or semi-public space, could silently identify people at medical clinics, union meetings, protests, or in any place where anonymity matters for safety.
For anyone running an organization with physical premises, the relevant question is not whether Meta launches this feature but what the existence of cameras-on-faces means for their environment right now. Employees, customers, and visitors may already be recorded without any indication. Policies written for phone cameras or ID badge systems do not cover this. Legal requirements around biometric data, which exist in Illinois, Texas, and several other places, were written with databases in mind, but the obligation to inform people their biometric data is being collected does not disappear because storage moved to a local device.
Competition is arriving too. Google is building smart glasses with Warby Parker. Apple is reportedly developing its own version. Whatever norm Meta sets with NameTag, if it launches, will be the baseline every competitor responds to.