Meta Tests Surveillance-Grade Face Recognition From Military Contractor for Ray-Ban Smart Glasses

Meta is currently exploring the integration of high-powered facial recognition software, developed by a prominent military and law enforcement contractor, into its burgeoning line of smart glasses. This development, revealed through internal software licenses and code analysis, indicates a deepening relationship between the social media giant and Rank One Computing, a Denver-based firm that specializes in biometric surveillance for the United States military and various federal agencies. The discovery marks the first documented business tie between Meta and Rank One, highlighting an increasingly blurred boundary between specialized surveillance tools designed for the battlefield and consumer electronics marketed to the general public.

The arrangement was brought to light through a software license issued by Rank One Computing, which was specifically tied to a test version of the Meta AI application. This application serves as the primary software interface for Meta’s Ray-Ban and Oakley smart glasses. Rank One is not a traditional consumer software vendor; approximately 80 percent of its revenue is derived from government contracts. Its technology is currently utilized by the U.S. Marshals Service to identify prisoners during transport and by the Naval Criminal Investigative Service (NCIS) for video surveillance. Furthermore, Rank One has previously developed long-range facial recognition capabilities for the U.S. Special Operations Command, claiming its algorithms can identify a face from a distance of up to one kilometer.

Technical Specifications and the Scope of Integration

The license acquired by Meta authorizes the use of Rank One’s core facial recognition engine alongside its "liveness detection" suite. Liveness detection is a sophisticated security feature designed to distinguish between a real human face and a high-resolution photograph or a 3D mask. According to the documentation, the license supports the creation and management of up to 10 million facial templates—unique mathematical signatures derived from individual facial features.

While Meta has not officially released facial recognition features for its smart glasses, remnants of the Rank One integration were found in the production version of the Meta AI app as recently as June 2024. Code analysis suggests that while the software routines to load and initialize the Rank One license were present in the app downloaded by millions of users, the features remained dormant and inaccessible to the public.

This discovery follows a separate revelation that Meta had internally developed a proprietary facial recognition system nicknamed "NameTag." This system was designed to allow users of smart glasses to identify individuals in their line of sight in real-time. On June 5, 2024, one day after public reports detailed the existence of NameTag, Meta reportedly scrubbed the facial recognition code from its app entirely. Despite the removal, the presence of the Rank One license suggests that Meta was weighing third-party, military-grade solutions as a potential alternative or supplement to its in-house efforts.

A Chronology of Rank One Computing and Meta’s Biometric Ambitions

The evolution of Rank One Computing provides context for why Meta would seek its expertise. Founded in 2015 by a group of engineers from the nonprofit research institute Noblis, Rank One was built on a foundation of government-funded research. Its founders previously spent years evaluating biometric algorithms for U.S. intelligence agencies. Since its inception, the company has rapidly ascended the ranks of the surveillance industry:

• 2015: Rank One Computing is founded with a focus on high-accuracy biometric algorithms.
• 2021: The U.S. Marshals Service begins deploying biometric kits powered by Rank One technology to verify the identities of detainees without the need for traditional fingerprinting.
• 2022: The company secures contracts with the NCIS for its "ROC Watch" video analytics tool.
• Early 2024: Rank One goes public on the Nasdaq, signaling its transition from a niche contractor to a major player in the security tech sector.
• June 2024: Investigative reports reveal the dormant Rank One license within Meta’s consumer-facing AI application.

Meta’s own history with facial recognition has been fraught with controversy. In 2021, the company announced it would shut down its "Face Recognition" system on Facebook, deleting the facial templates of more than a billion users due to "growing societal concerns." However, the company’s recent pivot toward artificial intelligence and wearable augmented reality (AR) hardware appears to have reignited its interest in the technology. Smart glasses represent the next frontier for Meta’s Reality Labs division, and the ability to identify people, objects, and landmarks in real-time is viewed by many industry analysts as the "killer app" that could drive mass adoption of the hardware.

Leadership and Ties to the Intelligence Community

The professional background of Rank One’s leadership further underscores its deep roots in the American security apparatus. The company’s CEO, B. Scott Swann, is a former senior official at the FBI, where he managed the division responsible for the bureau’s massive biometric databases. The company’s board of directors is a roster of former high-ranking intelligence and defense officials, including:

• A former Deputy Director for Science and Technology at the CIA.
• A former head of the FBI’s Science and Technology Branch.
• A former Pentagon official who oversaw the development of specialized military capabilities.

This concentration of intelligence expertise suggests that Rank One’s software is designed with the rigors of national security in mind—accuracy, speed, and the ability to function in "unconstrained" environments where lighting and angles are suboptimal. For Meta, licensing such technology could provide a shortcut to achieving high-performance facial recognition without the years of R&D typically required to meet military standards.

Performance Data and Demographic Disparities

While Rank One’s technology is touted for its accuracy, independent testing reveals the same demographic biases that have plagued the facial recognition industry for years. In evaluations conducted by the National Institute of Standards and Technology (NIST), Rank One’s algorithms demonstrated significant variances in performance across different demographic groups.

NIST’s "Face Recognition Vendor Test" (FRVT) showed that a version of Rank One’s algorithm produced false matches—instances where the software incorrectly identifies a person—at different rates depending on sex and country of origin. The data indicated that:

• Error rates were lowest for individuals born in Eastern Europe.
• False match rates were consistently higher for women than for men.
• The software struggled more with certain ethnic proxies, a common issue in algorithms trained on datasets that lack diverse representation.

These disparities pose a significant risk if the technology is deployed in consumer products. If a pair of smart glasses incorrectly identifies a stranger or a neighbor, the social and legal consequences for the user could be substantial. In a law enforcement context, these errors can lead to wrongful arrests; in a consumer context, they could lead to harassment or privacy violations.

The Convergence of Military and Consumer Technology

The potential integration of Rank One’s tools into Ray-Ban glasses is part of a broader historical trend where military innovations eventually migrate to the consumer market. Technologies such as the internet, Global Positioning System (GPS), and voice recognition (Siri) all had their origins in defense research. However, facial recognition represents a more sensitive category of technology because it involves the permanent, non-consensual collection of biometric data.

Privacy advocates argue that the "dual-use" nature of this technology—serving both a police officer looking for a suspect and a consumer looking for a friend at a concert—creates a dangerous feedback loop. As consumer companies adopt these tools, the technology becomes more normalized, potentially lowering the barrier for even more invasive surveillance by the state.

Joseph Jerome, a former policy official at Meta’s Reality Labs, noted that while the transition from military to consumer use is a common narrative in tech history, the speed and scale at which biometric surveillance is being miniaturized into wearable frames is unprecedented. The concern is no longer just about what the government can see, but what every citizen can see and record through their eyewear.

Regulatory Vacuum and Public Response

In the United States, there is currently no federal law governing the use of facial recognition by private companies. While some states, such as Illinois, Texas, and Washington, have enacted biometric privacy laws that require explicit consent before collecting facial data, the majority of the country remains a "wild west" for biometric experimentation.

Eric Null, director of the Privacy and Data Project at the Center for Democracy and Technology, emphasized that consumer-facing companies are clearly "craving access" to high-powered facial recognition. He warned that without federal guardrails, the risks of this technology becoming a standard feature in consumer electronics are "significant and largely unbounded."

Meta has remained largely silent regarding its relationship with Rank One. When questioned about the software license and the purpose of the integration, Meta spokespeople declined to provide specific details, refusing to confirm when the relationship began or whether it remains active. Rank One Computing has also declined to comment on its dealings with Meta.

Broader Implications for Privacy and Society

The prospect of Meta-branded glasses equipped with military-grade facial recognition raises profound questions about the future of anonymity in public spaces. If every person wearing a pair of glasses can instantly cross-reference a stranger’s face against a database of millions of templates, the concept of "being a face in the crowd" effectively disappears.

Furthermore, the involvement of a company like Rank One suggests that the "liveness detection" and "10 million template" capacity were not merely for unlocking a phone or tagging friends in a photo. Such scale suggests a broader ambition for environmental scanning and identification.

As Meta continues to push its "Meta AI" as a pervasive assistant that "sees what you see," the underlying infrastructure of that vision is coming into focus. It is an infrastructure built not just by social media engineers, but by veterans of the FBI, CIA, and the Pentagon. The outcome of this testing phase will likely determine whether the next generation of wearable tech is a tool for personal productivity or a decentralized node in a global surveillance network. For now, the code remains dormant, but the business foundations for a high-surveillance future have clearly been laid.