Meta Quietly Embeds Facial Recognition Into Smart Glasses App, Reigniting Privacy Concerns Over Biometric Surveillance

Meta has stealthily integrated facial recognition technology into the software powering its smart glasses, embedding the capabilities within an application already installed on millions of smartphones worldwide. A forensic analysis of Meta’s software reveals that despite the company’s public assurances that it was merely “thinking through” the implications of such technology, the core code for identifying individuals via wearable cameras has been quietly shipping to users for months.

The discovery, first identified through a technical review of the Meta AI companion app, centers on a feature internally codenamed “NameTag.” According to the analysis, this feature is designed to identify people captured by the cameras on Meta’s smart glasses—such as the Ray-Ban Meta collection—and provide the wearer with real-time alerts upon recognition. While the feature remains inactive for the general public, the underlying architecture required to transform the devices into mobile surveillance tools is already in place.

The Technical Framework of NameTag

The internal mechanics of NameTag suggest a sophisticated, multi-stage recognition pipeline. The software analysis, which was independently verified by security experts, identified three distinct artificial intelligence models that have already been deployed to users’ devices from Meta’s servers. The first model is tasked with detecting faces within a video frame; the second model crops these faces for processing; and the third model encodes the visual data into a unique biometric signature, or “faceprint.”

Once activated, NameTag would compare these live faceprints against a database of stored biometric data on the user’s phone. This database is currently configured to receive periodic updates from Meta. When a match is found, the app triggers a notification to the wearer. Faces that do not find an immediate match are not discarded; instead, the system is designed to crop, index, and save them into a folder labeled “pending,” suggesting a system capable of learning and cataloging new individuals over time.

In more recent versions of the app, specifically those updated in May, Meta appears to have rebranded the feature as “Connections.” The user interface for this iteration invites users to “remember the people you met,” framing the surveillance technology as a social tool for memory enhancement. However, the technical foundation remains identical to the NameTag system, raising questions about the transparency of Meta’s development process.

A Chronology of Biometric Controversy

The revelation of NameTag’s existence marks a significant pivot for Meta, which previously claimed to have abandoned facial recognition technology. To understand the gravity of this discovery, one must look at the company’s decade-long history of legal and ethical battles regarding biometric data.

In 2010, Facebook (now Meta) introduced its initial facial recognition system, which suggested tags for users in uploaded photos. By 2011, the system had become a flashpoint for privacy advocates and regulators in both the United States and Europe. Critics argued that the company was creating a massive, non-consensual database of human biometrics.

The mounting pressure culminated in 2019 when Meta agreed to a landmark $5 billion settlement with the Federal Trade Commission (FTC) to resolve a range of privacy concerns, including those related to its facial recognition practices. This was followed in 2021 by a $650 million settlement in an Illinois class-action lawsuit, which alleged that the company’s photo-tagging system violated the state’s Biometric Information Privacy Act (BIPA).

In November 2021, Meta made a high-profile announcement that it would sunset its facial recognition system on Facebook and delete more than one billion faceprints. At the time, Jerome Pesenti, Meta’s then-VP of Artificial Intelligence, cited “growing concerns about the place of facial recognition technology in society.”

However, internal documents and the recent code discovery suggest that this retreat was tactical rather than absolute. In early 2024, Meta agreed to a separate $1.4 billion settlement with the State of Texas over similar allegations of unlawful biometric data collection. Despite these massive financial penalties, the development of NameTag began as early as January 2024, with core components integrated into software distributed to over 50 million users.

Testing the Pipeline: The Foucault Experiment

To verify the functionality of the hidden code, independent security researchers conducted tests on the recognition pipeline. One researcher, operating under the pseudonym Buchodi, utilized a specific technical analysis to see if the matching system was operational.

By manually injecting a single faceprint—derived from a photograph of the deceased French philosopher Michel Foucault—into the app’s internal gallery, the researcher was able to trigger the system. When the app processed an image of Foucault, it successfully matched the biometric data and generated a notification stating, “Person recognized.”

“The main components of a face-recognition feature are already in Meta’s companion app,” Buchodi noted. “Not many pieces stand between this and a working feature.” This experiment confirmed that the software is not merely a dormant concept but a functional architecture waiting for a server-side switch to be flipped.

Meta’s Official Response and Defensive Stance

Meta has defended its actions, characterizing the presence of the code as part of its standard research and development process. Ryan Daniels, a spokesperson for Meta, stated that the company has been transparent about exploring these types of features.

"Regardless of any sensational reporting, the facts are simple: We’ve said before we’re exploring these types of features, and what you’re seeing is just evidence of that exploration," Daniels said. He emphasized that no final decision has been made regarding a public rollout and that Meta is not building a centralized face database. Instead, the current design suggests a decentralized approach where faceprints are pulled from servers and stored locally on user devices.

Meta’s leadership has argued that if the feature were to be released, it would follow a “thoughtful approach” with full transparency. However, privacy advocates point out that shipping the code to millions of devices before making a public announcement contradicts the spirit of transparency.

The Advocacy Backlash and Ethical Implications

The discovery has drawn sharp condemnation from a coalition of over 70 advocacy groups, including the American Civil Liberties Union (ACLU), the Electronic Privacy Information Center (EPIC), and Fight for the Future. These organizations have demanded that Meta scrap the NameTag project entirely, warning of the potential for mass stalking, harassment, and state-level surveillance.

Cooper Quintin, a senior public interest technologist with the Electronic Frontier Foundation (EFF), expressed alarm at the potential for Meta to turn its customer base into a “distributed surveillance machine.” Unlike fixed CCTV cameras, smart glasses are mobile, eye-level, and often indistinguishable from regular eyewear, making it impossible for members of the public to know if they are being identified and indexed in real-time.

Joseph Jerome, a former official at Meta Reality Labs who worked on privacy reviews, noted that there has always been internal tension regarding when to reintroduce facial recognition. He suggested that by embedding the technology into a mass-market wearable platform, Meta is attempting to normalize a capability that was once considered radioactive.

Assistive Technology vs. Public Privacy

Meta has occasionally framed the development of facial recognition as a tool for social good, particularly for the visually impaired. Internal documents suggest that Meta once planned to debut the feature at a conference for the blind, highlighting its potential to help users identify acquaintances in social settings.

A 2018 study conducted by researchers at Cornell Tech and Facebook found that blind participants viewed person-recognition as a critical daily task. While existing assistive devices already allow users to identify people they have personally “enrolled” in their system, Meta’s NameTag appears to have a broader scope. The ethical dilemma lies in whether the benefit to a specific group of users justifies the creation of a tool that could be used for general-purpose surveillance.

The Future of Biometric Norms

Legal experts warn that Meta’s strategy may be aimed at shifting the psychological and moral cues of the public. Woodrow Hartzog, a privacy law professor at Boston University, argues that the more these systems are deployed, the more they are perceived as unexceptional.

“The more we come to see them as unexceptional and routine, the more people tend to start to take their moral cues about whether it’s desirable or good to have your face scanned,” Hartzog stated. He further cautioned that even “opt-in” protections can be a thin defense, as consent is often tied to the ability to use a service or gain a benefit, effectively removing true choice from the consumer.

As of now, the Ray-Ban Meta glasses continue to be a top-selling product in the wearable tech category. While the “Connections” feature remains locked behind a software gate, its presence on 50 million phones suggests that the infrastructure for a new era of biometric surveillance is already in place, awaiting only the company’s signal to begin. Meta’s next moves will likely determine whether wearable AI becomes a helpful personal assistant or a permanent, invisible observer of public life.