Meta Killed Facial Recognition on Facebook. Now It's Putting It on Your Face.

Internal documents show Meta is waiting for privacy groups to be overwhelmed before launching "Name Tag", a feature that identifies people through smart glasses. They're betting you're too distracted to notice.

State of Surveillance Editorial Team

TL;DR: Meta is building facial recognition into its Ray-Ban smart glasses through a feature internally called “Name Tag.” An internal document obtained by The New York Times reveals Meta plans to launch during political turmoil: specifically because privacy groups are “focused on other concerns.” The feature would let wearers identify people connected to Meta platforms on sight. Meta shut down facial recognition on Facebook in November 2021 and deleted over 1 billion face templates after paying $2 billion in biometric privacy settlements. Now the company wants to put the same technology on 7 million pairs of smart glasses walking around in public. The difference: this time, it scans strangers, not friends.

The Quiet Part, Out Loud

On February 13, 2026, The New York Times published internal Meta documents showing the company’s plan to add facial recognition to its smart glasses line: Ray-Ban Meta, Oakley Meta HTSN, and Meta Ray-Ban Display [1].

The most revealing detail wasn’t the technology. It was the timing strategy.

An internal document from May 2025 stated: “We will launch during a dynamic political environment where many civil society groups that we would expect to attack us would have their resources focused on other concerns” [1].

Read that again. Meta isn’t worried about whether facial recognition on glasses is ethical. They’re worried about getting caught. And they’re betting that the ACLU, EFF, and CDT are too busy fighting government surveillance to fight corporate surveillance at the same time.

Alexandra Reeve Givens, CEO of the Center for Democracy & Technology, called it a “privacy and surveillance nightmare” and added: “People should not have to duck and dodge their way through society to avoid being surveilled by facial recognition-enabled glasses” [2].

What “Name Tag” Actually Does

The feature is internally called “Name Tag.” Here’s what Meta’s documents describe [1][3]:

  • Wearers look at someone through their smart glasses
  • Meta’s AI assistant identifies the person by matching their face against Meta’s platform data
  • The system surfaces information about the person: their name, profile details, mutual connections
  • It works on people with Meta accounts, Facebook profiles, and public Instagram accounts

Meta says it wouldn’t function as a universal facial recognition tool, only identifying people connected to its platforms. But Meta has 3.3 billion monthly active users across Facebook and Instagram. That’s nearly half the planet’s population. “Limited” is doing a lot of work in that sentence.

Meta is also developing “super sensing” glasses that continuously run cameras and microphones to record your entire day, like an AI note-taker for real life. Facial recognition would be key to this system: the glasses could remind you of tasks when you see a specific colleague, for example [1]. Mark Zuckerberg has questioned whether the glasses should keep their LED indicator light on during “super sensing” or use a different signal [1]. The fact that this is even a question tells you where Meta’s priorities are.

Meta Already Tried This. It Cost Them $2 Billion.

Here’s the timeline Meta is hoping you’ve forgotten:

  • 2010: Facebook enables facial recognition for photo tagging, without user consent [4].
  • 2019: After years of backlash, Facebook makes the feature opt-in instead of default. The same year, Meta pays a $5 billion FTC settlement over privacy violations, including deceptive consent practices [5].
  • March 2021: Meta settles an Illinois biometric privacy class action for $650 million, one of the largest privacy settlements in history [5].
  • November 2021: Meta shuts down facial recognition on Facebook entirely. The company says it will delete more than 1 billion face templates [5].
  • July 2024: Meta pays $1.4 billion to settle a Texas biometric privacy lawsuit over the same defunct facial recognition system [5].

Total bill for the last time Meta scanned faces: roughly $7 billion in settlements and fines.

And now they want to do it again, except instead of scanning photos you upload, they want to scan everyone you walk past on the street.

The FTC Problem Meta Is Trying to Ignore

As part of its $5 billion FTC settlement in 2019, Meta agreed to review every new or modified product for potential privacy risks before launch [4].

In January 2025, Meta restructured. The company reduced its privacy team’s influence over product releases and imposed time limits on how long privacy risk reviews could take [1].

Andie Millan, a director of risk review in Meta’s Reality Labs division, told employees the changes would “push the bounds” of Meta’s FTC agreement [1].

That’s a Meta employee saying, on the record internally, that the company is testing how far it can bend a consent decree before it breaks. And the product they’re testing it with is facial recognition on glasses worn in public.

The EFF put it bluntly: “Meta’s conclusion that it can avoid scrutiny... is craven and morally bankrupt. It is also dead wrong” [5].

Two Harvard Students Already Proved What’s Coming

In October 2024, Harvard juniors AnhPhu Nguyen and Caine Ardayfio built a tool called I-XRAY that did what Meta is now planning to productize [6].

Their system was simple:

  1. Ray-Ban Meta glasses stream video via Instagram Live to a phone
  2. An algorithm detects faces in the video feed
  3. The system runs detected faces through PimEyes, a commercial face search engine
  4. When a match is found, it scrapes public databases for the person’s name, address, and phone number

The whole thing took four days to code. “The meat is in the software,” Nguyen told The Harvard Crimson [6].

In a viral demonstration video that got over 20 million views, Ardayfio walked up to a stranger at a Boston T station, greeted her by name, and referenced her volunteer work, all information pulled from the glasses in real time. The New York Post called it “every stalker’s dream” [6].

Nguyen said they built it to raise awareness: “We think it’s best if people are straight-up aware that these things do exist” [6].

Meta saw the same demonstration. Their response wasn’t to prevent it. It was to build it themselves.

The Data Pipeline Nobody Talks About

Meta’s privacy policy for smart glasses already allows the company to collect photos and conversation transcripts from glasses users for advertising and AI training [4].

According to Proton’s analysis, Meta provided user account information to the federal government at a rate that increased 675% between 2014 and 2024 [4].

Now add facial recognition. Every person a glasses wearer looks at gets their face processed by Meta’s AI. Even if Meta says it won’t store the data permanently, the infrastructure exists. The pipeline is built. And Meta has a documented history of expanding data collection after promising limitations.

The EFF noted: “Meta cannot possibly obtain consent from everyone, especially bystanders who are not Meta users” [5]. You don’t get to opt out of being scanned by someone else’s glasses.

EssilorLuxottica, Meta’s manufacturing partner, has sold over 7 million smart glasses units in the past year [3]. That’s 7 million potential walking facial recognition cameras. And that number is growing fast as Meta competes with OpenAI’s wearable devices for market share.

The Bigger Picture

Meta’s Name Tag announcement lands in the same week that:

Government agencies are building facial recognition databases. Private companies are building the cameras. And Meta is trying to put one on every face in America.

The ACLU’s Nathan Freed Wessler warned: “Face recognition technology on the streets of America poses a uniquely dire threat to the practical anonymity we all rely on. This technology is ripe for abuse” [3].

Dozens of states now classify biometric data as requiring affirmative consent under privacy law. Meta has already paid $2 billion for violating those laws. And their internal memo shows they plan to launch anyway, just when nobody’s looking.

What You Can Do

Lock Down Your Instagram and Facebook

Name Tag reportedly works by matching faces to Meta platform profiles. Set your Instagram to private. Limit what’s visible on your Facebook profile to non-friends. Remove clear face photos from public posts. This won’t stop all matching, but it reduces your attack surface.

Opt Out of Meta’s Face Recognition

When Name Tag launches, watch for a facial recognition opt-out setting in your Meta account privacy controls. Meta may bury it. They have a history of hiding privacy settings deep in menus. Check under Settings > Privacy > Face Recognition or a similar path.

Remove Your Face from PimEyes

PimEyes is the face search engine the Harvard students used. You can request removal at pimeyes.com/en/opt-out-request. This doesn’t affect Meta’s system, but it closes one of the most accessible public face-matching tools.

Support Biometric Privacy Laws

Illinois’s BIPA is the reason Meta paid $650 million. Texas’s biometric law cost them $1.4 billion. Push for similar laws in your state. The EFF, ACLU, and CDT are tracking and supporting biometric privacy legislation nationwide.

What Happens Next

Meta says it’s “still thinking through options” and will “take a thoughtful approach if and before we roll anything out” [4]. That’s the same company whose internal memo says they’re timing the launch for when critics are overwhelmed.

The EFF has already signaled it will fight: “EFF, other civil liberties groups, and plaintiffs’ attorneys will be here to help” [5].

Meta is betting on two things: that you’re too tired of surveillance news to care, and that there are too many fires for privacy groups to fight them all. The internal memo says it plainly. They know this is wrong. They’re doing it anyway. They just want to pick the right moment.

Don’t let them.

References

  1. TechCrunch: Meta Plans to Add Facial Recognition to Its Smart Glasses, Report Claims (February 13, 2026)
  2. Gizmodo: The World Is on Fire, and Meta Sees an Opportunity to Add Facial Recognition to Smart Glasses (February 13, 2026)
  3. Engadget: Meta Is Reportedly Working to Bring Facial Recognition to Its Smart Glasses (February 13, 2026)
  4. Futurism: Meta Adding Facial Recognition to Its Smart Glasses That Identifies People in Real Time (February 14, 2026)
  5. EFF: Seven Billion Reasons Facebook Should Abandon Its Face Recognition Plans (February 13, 2026)
  6. The Harvard Crimson: Harvard Students Create Tool to Identify Strangers Using Ray-Ban Meta Glasses (October 19, 2024)