Meta licensed face recognition from Rank One Computing, a vendor that earns most of its money from US police and the military, to prototype identifying people through its glasses. It shipped the code dormant to 50 million phones, then deleted it.
The Meta face recognition system for its smart glasses was built on software licensed from Rank One Computing, a Pentagon and police contractor, according to a WIRED investigation. Reporters Dell Cameron and Dhruv Mehrotra found a leaked, still-active licence tying Meta to a vendor that draws roughly 80 per cent of its revenue from government clients.
The Denver firm, founded in 2015 and newly listed on Nasdaq this February, supplies face recognition to US police and the military. Its technology has verified prisoners for the US Marshals Service since 2021; the Navy’s criminal investigators bought its video tool; and US Special Operations Command funded work that, the company says, can identify a face from up to a kilometre away.
Its board is stacked with former CIA, FBI and Pentagon officials, and its chief executive once ran the FBI division that keeps the bureau’s biometric databases.
The Meta face recognition code, already pulled
This lands on top of a story Meta hoped had ended.
The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!
WIRED reported on 5 June that Meta’s AI companion app, installed on more than 50 million phones, carried a dormant face-recognition pipeline, internally called NameTag, that could turn a glance through the glasses into a name. Meta deleted the code the next day.
What it did not say was where the technology came from. The Rank One licence, which WIRED says supports up to 10 million facial templates, is the answer: it sat dormant inside the app alongside Meta’s own face-recognition system, and was never switched on for users.
A defence-grade tool, and a public contradiction
The detail matters because of what Meta says in public. The company insists it will not add face recognition to its Ray-Ban and Oakley glasses without robust privacy safeguards, and it switched off Facebook’s photo-tagging in 2021.
Quietly licensing a military-grade engine to prototype exactly that feature is the gap between the message and the build. Pairing always-on camera glasses with this kind of recognition is the scenario privacy researchers have warned about for years: identifying a stranger on the street, in real time, without consent.
There are reasons to be wary of the technology itself, not just the optics. In NIST testing, a version of Rank One’s algorithm produced false matches at sharply different rates by sex and country of birth, a proxy the agency uses for race, with higher error rates for women. The US has almost no national rules on face recognition.
Meta declined to say why it licensed the software, when the relationship began, or whether it continues; Rank One declined to comment. The takeaway is uncomfortable but simple: the tech to put a name to every face you look at has been licensed and tested, and the only real question left is who ships it first, and on what legal footing.
With expertise in digital marketing, product management, and branding & identity, Ana Maria Constantin develops strategies that resonate (show all) With expertise in digital marketing, product management, and branding & identity, Ana Maria Constantin develops strategies that resonate with our target audience in the software/SaaS industry. Collaboration and teamwork are paramount to her, as she loves empowering her colleagues to achieve outstanding results and unlock their full potential.