In February 2026, The New York Times published a report on an internal Meta document describing a feature under development for Ray-Ban Meta smart glasses called “Name Tag.” The feature, as described in the document, would allow a person wearing the glasses to identify individuals they encounter — in the street, at a café, at a conference — by their face, in real time, using Meta’s AI systems and its social graph.
The report described two versions that had been considered internally: one limited to identifying people you already know on Meta’s platforms, and a broader version capable of recognizing any public Meta account user. The document did not say when or whether the feature would be released, and Meta did not confirm or deny the specifics.
Several months later, Name Tag has not been publicly released. But the document’s existence, and the reasoning it reveals about how Meta is thinking about its glasses platform, is worth understanding — because the trajectory points somewhere specific, and that destination has significant implications for how personal photos and identity information work in public spaces.
What Name Tag Would Actually Do
The capability described in the leaked internal document is technically real-time facial recognition at consumer scale.
Ray-Ban Meta glasses have a forward-facing camera. When you wear them and look at someone, the glasses can capture what you’re seeing. The Name Tag concept would run facial recognition on that captured image against Meta’s databases — which include the profile photos, tagged photos, and identity data of billions of users across Facebook, Instagram, and WhatsApp — and return identifying information about the person you’re looking at.
At the most basic level described: you look at a stranger, and your glasses tell you their name.
This is qualitatively different from other privacy concerns around smart glasses.
The existing concern with smart glasses — covert recording of bystanders, footage sent to cloud servers for processing, video that could be used to train AI models — is about the misuse of recorded content. The footage is captured, and something bad might happen with it later.
Name Tag, as described, would be something else: a real-time identification system that gives the wearer information about anyone they look at, without any visible cue to the person being identified, without any action by the identified person, and without that person having any meaningful way to prevent it.
The person being identified doesn’t know it’s happening. They didn’t opt in to any system. They may not even know Ray-Ban Meta glasses exist. And the only reason their identity is in the system is that they created a Meta account at some point, with a profile photo and friends who tagged them in photos over the years.
The Current State of the Ray-Ban Platform
The broader context for this is a Ray-Ban Meta platform that has had a turbulent 2026.
In March 2026, a federal class-action complaint was filed in the Northern District of California, accusing Meta and EssilorLuxottica of routing footage captured through the glasses to Kenyan subcontractors for labeling without adequate disclosure to users. The complaint alleged that Meta’s marketing — “designed for privacy, controlled by you” — was misleading given actual data handling practices.
In July 2026, Meta rolled out a mandatory update to the glasses that disables the camera if the device detects that the LED indicator light has been physically tampered with or destroyed. The update followed reports of users deliberately disabling the recording light to make covert recording easier. Meta said it would also pursue legal action against services offering to modify or defeat the LED indicator.
Also in July 2026, Fortune reported on an internal Meta prototype called “super-sensing” glasses that continuously record audio and take photos every few seconds, operating more like an ambient sensor than a user-triggered camera. The prototype is not a released product, but its existence illustrates the direction of internal development.
Name Tag belongs in this same category: not yet a released feature, but a documented direction of development.
Why This Is Different From Other Facial Recognition Concerns
Facial recognition is already embedded in many products. Google Photos and Apple Photos use it to group people in your photo library. Ring’s Familiar Faces feature uses it to identify recurring visitors to your home. Many airlines now use it for biometric boarding. These applications involve facial recognition in defined contexts with at least some degree of user choice.
Name Tag as described would be different in three ways.
Scale. Meta’s social graph represents billions of people. A facial recognition system connected to that database can potentially identify an enormous fraction of any population in a major city. Competing systems using smaller databases have narrower reach; Meta’s would not.
Context. Using facial recognition in your own photo library, at an airport where you’ve enrolled in biometric boarding, or at your own home security camera involves a defined context. Name Tag would operate anywhere the wearer chooses to look, in any public or semi-public space.
The consent model. Every other facial recognition system in widespread consumer use involves some form of enrollment by the identified person: you upload photos to Google Photos and consent to its terms; you enroll at an airport; you show up repeatedly at a home camera. Name Tag, if built against Meta’s existing database, would identify people based on content they shared in entirely different contexts, without any enrollment decision about being recognizable through smart glasses.
The Legal Landscape
In the United States, there is no federal biometric privacy law. The strongest state law is Illinois’s Biometric Information Privacy Act (BIPA), which requires informed written consent before collecting biometric identifiers including facial geometry. Texas has a similar law. Several other states enacted biometric privacy laws in 2025 and early 2026.
Whether BIPA would cover a consumer smart glasses feature like Name Tag is an open legal question. The law applies to private entities, and several BIPA cases have addressed facial recognition in commercial contexts. Meta has faced BIPA litigation over facial recognition features in its other products.
In the EU, high-risk biometric identification in public spaces is substantially restricted under Article 5 of the EU AI Act, which took full effect on August 2, 2026. Real-time remote biometric identification in publicly accessible spaces is prohibited with narrow exceptions — and those exceptions are framed around law enforcement, not consumer products. A Name Tag feature that operated in EU territory would face significant legal obstacles.
The geographic asymmetry here mirrors what happened with Google’s Gemini photo scanning feature and Apple’s various AI features: capabilities that are enabled in the United States and blocked or restricted in the EU, revealing the regulatory gap.
What Your Photos Have to Do With This
Personal photos that you’ve shared on social media — tagged photos, profile images, photos from years ago that friends tagged you in — form part of the database that would make a feature like Name Tag possible.
Most people, when they shared those photos, were not thinking about whether that content would make them identifiable through a stranger’s glasses in a coffee shop a decade later. The consent they gave was to the social network’s stated terms at the time. Those terms have changed repeatedly.
This is one of the structural arguments for keeping personal photos and files in private storage rather than on public platforms: content shared publicly accumulates in ways that become hard to reverse. Photos that exist only in your private archive cannot be used to identify you through someone else’s glasses, because there is no accessible database containing them.
That doesn’t mean everyone needs to delete social media or avoid ever sharing photos. It does mean the privacy calculus for what you share publicly has externalities — consequences for others and for yourself — that go beyond the immediate context of sharing.
What You Can Actually Do
Know the state of the technology. Name Tag is not currently a released feature. Meta’s released glasses do not offer real-time facial recognition of strangers as a consumer feature. Understanding what exists now versus what is in documented development is useful for calibrating concern.
Understand opt-out options on Meta’s platforms. Meta offers face recognition settings that control whether your face is used for tagging suggestions and related features. Reviewing and limiting these settings reduces the data available for any future recognition features.
Consider what’s in your public profile photos. Profile photos on public accounts are the most accessible source for facial recognition training data. Using a photo that clearly shows your face on a public profile is a choice; so is using a less identifiable image or keeping accounts private.
Watch for Name Tag’s release. If Meta releases a real-time facial recognition feature for Ray-Ban glasses, the announcement will be significant and will likely generate substantial coverage. Staying informed about what the platform does and updating settings accordingly is the most actionable step currently available.
Keep personal photos private. For photos and files you want to remain genuinely private — not identifiable by external systems, not part of any training dataset — keeping them in private storage rather than on public platforms is the most effective long-term protection.
The Trajectory
The February 2026 internal document was a signal, not a final decision. Features described in internal planning documents sometimes ship; sometimes they don’t. Name Tag may look very different by the time it launches, if it launches at all.
What the document reveals is how Meta is thinking about the glasses platform: not as a camera accessory but as an AI-powered layer between the wearer and the world, capable of surfacing information about what the wearer is seeing — including information about the people the wearer encounters.
That trajectory is the thing worth watching. The specific feature name matters less than the underlying capability being developed, because once real-time facial recognition at consumer scale exists as infrastructure, the question of which specific features use it becomes a matter of product decisions, not technical constraints.
Understanding where this is going is more useful than reacting to any single announcement.