One of the biggest fantasies -- and fears -- in the science fiction-like realm of smart glasses is the idea of looking across a room, seeing someone, and instantly being able to recognize who they are. And maybe when the last time you met them was. And who knows what else.
The line between fantasy and reality, and fantasy and nightmare, often gets pushed and pulled to the limit in new tech. Meta, the biggest maker of smart glasses right now, is apparently already well underway with plans to introduce facial recognition into its glasses, according to a new report from the New York Times that mirrors reports that The Information wrote last year.
While Meta has goals for facial recognition tech to be used for assistive purposes, an internal company memo from 2025 cited in the Times story sees our current "dynamic political environment" as being a good landscape to launch a controversial feature like facial recognition, claiming that "many civil society groups that we would expect to attack us would have their resources focused on other concerns."
That sentiment alone is deeply worrying, and not surprising for a company like Meta that has been at the center of privacy scandals more than once. If facial recognition technology like this does come to smart glasses -- and I expect it will -- it'll need to be handled with extreme measures of control and responsibility. Sliding its debut into a chaotic political landscape in hopes it'll go unnoticed -- or unregulated -- is the worst type of outcome.
Meta didn't immediately respond to a request for comment.
Is facial recognition on glasses inevitable?
Nothing in tech is "inevitable," but even so, I don't see a way that facial recognition on glasses won't happen to some degree sooner or later. While no smart glasses have facial recognition capabilities now, it's totally possible to do this. AI can already recognize faces in photos and some of our phone apps have been using the technology to sort our photo libraries for years. The federal Immigration and Customs Enforcement agency, or ICE, which has been widely criticized for its heavy-handed tactics, already uses it in software via Clearview AI and Mobile Fortify without the public's consent. Two students hacked a way to make Meta glasses do it in 2024.
Facial recognition has been possible for so long, CNET wrote a feature package about it in 2019. It's the consent and privacy parts that I'm thinking about. Our brains already have facial recognition capabilities, but we don't share that data out of our brains to others. And, if a tech company had the capability to recognize and label faces, would that function be limited to personal and private access, or shared out with government agencies or within that company itself?
It's an idea that's already embedded deep in the very metaphor of Facebook, a company named for the little student handbooks at colleges that would list who's who in a given class, and Mark Zuckerberg's Facebook app is essentially a portable digital recognition tool. A pair of glasses that could do the same thing and accelerate recognition and connection seems like a natural bridge.
The second-gen (or third-gen if you count Stories) Meta Ray-Ban glasses get a big battery boost. Always-on AI modes could be improving every year.
Scott Stein/CNETLimits are needed now
I'd be lying if I said I didn't dream about having something like that in my own life, to recognize people at parties and conferences, to assist my aging brain in helping place how I knew someone. But that type of power needs limits. According to the Times report, Meta may limit the recognition of faces to people you're already connected to on Facebook. Maybe it would also let you know if someone else was also on Facebook, or was someone you didn't know, but it wouldn't tell you who they were. Maybe.
I think about connecting protocols like AirDrop, that have limits on sharing. You could leave AirDrop open and public, or limit to personal contacts, or set a 10-minute window. A future of facial recognition wearables would need to determine consent for whether someone wanted to be recognized, and also allow for bubbles around personal networks, locations, or time periods. Maybe it works for a conference networking dinner, and then deactivates later when I'm on my own at the hotel.
Facial recognition tools could be a huge help for vision-impaired communities, something the Times report also acknowledges. In fact, Meta was first planning to debut these facial-recognition modes, possibly called Name Tag, at a conference for the blind.
The father of one of my friends uses Meta's glasses for vision assistance on a daily basis, and these glasses already present huge benefits for assistive needs. Where will the line be drawn between assistance and surveillance, though, and how will the privacy be managed?
A white LED lights up when the glasses are recording. There's also a shutter snap sound. Both are subtle, yet still easy to miss.
Scott Stein/CNETWill we even know it's happening?
The reports from The Times and The Information discuss Meta's plans for an always-on type of AI "super sensing" awareness using cameras, which is already possible to a limited degree now. Live AI modes can be triggered on Meta smart glasses and on Google's upcoming glasses, but the limit is battery life. Right now, these modes can only work for an hour at a charge. But those limits likely will be extended as battery life and software improve.
And will you know the glasses are even on and scanning? It's already hard to tell when the not-so-obvious indicator lights are on and functioning, and glasses aren't widespread enough, or consistently designed, to build general awareness to look for indicator lights. In my own experience, most people who see my smart glasses nervously and jokingly ask if I'm recording. They don't really know.
These limits and protocols need to be solved, because the time could be coming -- very soon, in fact -- when glasses will be able to do this, and a whole lot more. One way to go would be to try to legislate limits or bans on how much these glasses could do. Another would be having tech companies figure out these responsibilities now. Meta may be pushing the door down soon enough.
With Google next on deck with glasses, and AI becoming more capable every month, widespread facial recognition is less an if than a when. The race will be on for smart glasses makers to distinguish themselves with new features, and with Meta leaning on improving smart glasses sales as its hot hand, facial recognition is likely one of the next magic tricks ready to go in its pocket.
English (US) ·