The person being identified doesn't know it's happening.
In a quiet but consequential reversal, Meta is preparing to embed facial recognition into its Ray-Ban and Oakley smart glasses, allowing wearers to identify strangers in real time through a feature called 'Name Tag.' What the company once declined to pursue in 2021 out of legal caution, it now approaches with the confidence of more powerful AI — and the same unresolved questions about consent, surveillance, and the erosion of anonymity in public life. The distance between a tool that recognizes friends and one that unmasks strangers is, in practice, almost no distance at all.
- Meta is set to launch facial recognition in consumer smart glasses in 2026, a technology it explicitly rejected just five years ago as too legally and ethically risky.
- The 'super sensing' system runs continuously in the background without user activation, turning ordinary eyewear into an uninterrupted surveillance instrument.
- Harvard researchers already demonstrated in 2024 that this exact capability — smart glasses with facial recognition — can expose strangers' identities in public spaces, making the risks concrete rather than hypothetical.
- Meta's own internal documents acknowledge 'security and privacy risks,' yet the company is pressing forward, framing its caution as a 'thoughtful approach' rather than a reason to pause.
- The person being scanned has no knowledge, no means of consent, and no way to opt out — a structural asymmetry that no amount of careful rollout language can resolve.
Meta is preparing to introduce facial recognition into its Ray-Ban and Oakley smart glasses this year, through a feature internally known as 'Name Tag.' The system would scan faces in real time and match them against Meta's social platforms — surfacing the identities of friends, or of strangers who maintain public accounts on Instagram or Facebook. It marks a striking reversal: in 2021, Meta had weighed the same capability and walked away, judging the legal and regulatory exposure too great.
What changed was the technology. Meta developed a broader AI system called 'super sensing' — a continuous background process that analyzes video in real time without requiring the wearer to trigger it. An LED light would signal when the function is active, but the glasses would otherwise operate as a persistent sensor, processing faces and environments without interruption. Sources who reviewed internal Meta documents told The New York Times that the Name Tag feature is planned for Ray-Ban Meta and Oakley Meta glasses during 2026.
The risks are not abstract. In 2024, two Harvard students demonstrated that smart glasses with facial recognition could identify people in public spaces in real time — a proof of concept that made visible exactly what is now being prepared for mass consumer deployment. Meta's own internal framing acknowledges 'security and privacy risks,' even as the company moves toward launch.
Publicly, Meta has been measured, saying it is 'still considering options' and will take a 'thoughtful approach if and before launching something.' The company also notes the feature would not function as a universal identification tool — limited, it says, to people with some connection to Meta's platforms or with public accounts. But that boundary is thinner than it sounds. Once a face can be matched to a public profile through a pair of glasses, the practical gap between identifying a friend and identifying a stranger collapses entirely. The person being scanned cannot know it is happening, cannot consent, and cannot refuse — leaving the public to trust that a company with a long and contested history on privacy will wield this capability with restraint.
Meta is moving forward with facial recognition technology for its Ray-Ban and Oakley smart glasses this year, introducing a feature internally called 'Name Tag' that will identify people in real time and surface information about them through the company's AI assistant. The capability represents a significant reversal from the company's own earlier judgment. Back in 2021, Meta had considered adding facial recognition to these glasses but ultimately decided against it, citing legal and regulatory concerns that made the technology too risky to deploy.
That calculus began to shift last May when the company started reconsidering the technology. Meta had been developing new software to enable facial recognition, along with a broader system called 'super sensing'—an AI-powered tool that runs continuously in the background, analyzing video in real time to understand what the wearer is seeing. The system would work without requiring the user to explicitly activate it, though an LED light would indicate when the super sensing function was in use. The glasses would essentially become a continuous sensor, processing faces and environments without interruption.
Now, according to sources who spoke with The New York Times after reviewing internal Meta documents, the company plans to roll out the Name Tag feature on Ray-Ban Meta and Oakley Meta glasses during 2026. The internal documents reveal that Meta has been consulting on how to launch a capability that carries what the company itself describes as 'security and privacy risks.' The feature would work by scanning a person's face and matching it against Meta's social networks—identifying friends from Instagram or Facebook, or potentially revealing the identities of strangers who maintain public accounts on those platforms.
The privacy implications are not theoretical. In 2024, two Harvard University students completed a research project demonstrating that smart glasses equipped with facial recognition could reveal people's identities in public spaces in real time, creating tangible risks to both privacy and personal safety. That research underscores why Meta's earlier decision to reject the technology in 2021 made sense, and why the company's current pivot warrants scrutiny.
Meta's public position on the feature remains cautious. In a statement to The New York Times, the company said it is 'creating products that help millions of people connect and enrich their lives' and acknowledged that while it 'often hears about interest in this kind of function,' it is 'still considering options and will take a thoughtful approach if and before launching something.' The company also noted that the feature would not function as a universal identification tool—it would not allow users to identify anyone, anywhere, at any time. Instead, it would be limited to recognizing people who already have some connection to Meta's platforms or who have chosen to make their accounts public.
Still, the distinction between identifying friends and identifying strangers with public profiles is a narrow one. Once someone's face can be matched to a public social media account in real time through glasses, the practical difference between targeted and universal identification collapses. The person being scanned has no way to know they are being identified, no way to consent, and no way to opt out. What Meta is describing as a thoughtful approach to a feature that carries security and privacy risks is, in effect, asking the public to trust that the company will use powerful surveillance technology responsibly—the same company that has faced years of criticism for how it handles user data and privacy.
Citações Notáveis
We are creating products that help millions of people connect and enrich their lives. Although we often hear about interest in this kind of function, we continue to consider options and will adopt a thoughtful approach if and before launching something.— Meta, in a statement to The New York Times
A Conversa do Hearth Outra perspectiva sobre a história
Why is Meta bringing this back now, after rejecting it five years ago?
The technology itself got better. They developed 'super sensing,' which runs continuously in the background analyzing what the glasses see. That made it feel less intrusive—not something you have to activate, just something that's always working. And the company has spent years building out its social networks. Now they have the infrastructure to match faces to accounts instantly.
But they knew the risks in 2021. What changed about those legal and regulatory concerns?
Nothing, really. The risks are the same. But the business case got stronger. Identification in real time, at scale, across their platforms—that's valuable. The legal risk didn't disappear; they just decided it was worth taking.
The company says it will only identify friends or people with public accounts. Doesn't that limit the harm?
Not really. If you're wearing glasses that can instantly tell you who anyone is—whether they're your friend or a stranger with a public profile—you're still creating a surveillance tool. The person being identified doesn't know it's happening. They can't opt out. The distinction between 'friends' and 'public accounts' is a technical limit, not a meaningful privacy protection.
What does the Harvard research actually show?
Two students proved that smart glasses with facial recognition can reveal identities in public spaces in real time. They demonstrated the vulnerability exists. Meta's own internal documents say the feature carries 'security and privacy risks.' They know what they're building.
So why launch it at all?
Because the value to Meta—and to users who want to know who they're looking at—outweighs the risk in the company's calculation. They're betting that 'thoughtful approach' language will satisfy regulators and the public, and that most people will want the feature once it exists.