Can You Tell If Smart Glasses Are Recording You? Every Method Has a Footnote

Five small amber cards in a row on a dark teal field, holding a dot of light, a lens, a hand, a sound wave and a telephone handset; a large asterisk hangs under each one, and a block of blank grey ruled lines below them fills more than half the frame
AI-generated editorial illustration. Not a documentary image.

Each One Comes With a Footnote

Type the question into a search box and the answers arrive with numbers in the headline. Five ways to tell. How to know if you're being filmed. The genre is well established and the advice is broadly the same everywhere.

Read one of them slowly, though, and something else shows up. Every method comes with a sentence explaining when it fails, and those sentences are doing more work than the headline.

The five methods, with their own small print

BGR's version is representative and, to its credit, honest about itself. Each of its five methods carries a caveat in the same paragraph:

Watch for the recording light — but it is, in the article's words, "distressingly simple to cover it up or otherwise disable the feature." Look for the lens in the corner of the frame or at the bridge — but lenses "can be hard to make out from a distance." Watch for someone tapping the side of the frames or moving their whole head to keep you in shot — but "a practiced hand may be able to keep these motions looking innocuous." Listen for the shutter click — but "the sounds produced by a pair of smart glasses are subtle if you're not the one wearing them." Use a detection app — but its creator "notes that it can produce false positives, and attempting to harass anyone who you think is wearing smart glasses could get you in trouble."

That is not a dishonest article. It is a careful one whose headline promises a skill and whose body concedes, five times, that the skill is conditional. Other outlets and syndicated versions run the same shape, as does a guide from a smart-glasses maker — which has an obvious interest in the impression that identification is manageable.

The caveats are all true. What nobody assembles is what they add up to.

Notice what each one is really saying. The light caveat is about a device that has been modified. The lens caveat is about distance. The behaviour caveat is about somebody who has practised. The audio caveat is about noise. Those are four different failure conditions, and a street, a train carriage or a bar supplies most of them at once — a stranger several metres away, in ambient noise, standing normally. The conditions under which the advice works are roughly: someone close, in a quiet room, being obvious about it. That is the one situation in which you did not need the advice.

What the app actually tells you

The fifth method is the one with real engineering behind it, and it deserves a fair description because it is the closest thing to an answer.

Nearby Glasses, an Android app by Yves Jeanrenaud with its source published openly, does not look at the glasses at all. It listens for radio. Bluetooth Low Energy advertising packets carry a manufacturer field — a company ID issued by the Bluetooth Special Interest Group — and the app watches for four of them: Meta, Meta Technologies, Luxottica and Snapchat. Cross the default signal threshold, roughly ten to fifteen metres outdoors, and your phone tells you something is there. TechCrunch and Malwarebytes carried the same account, and an iPhone equivalent has since gone around scanning for a similar set of devices.

This works in a way eyeballing a frame does not. Radio does not care about lighting, distance within range, or how casually somebody is standing. It is also, unlike every other method on the list, not a judgement — the phone either sees the identifier or it does not, and it is not asking you to decide whether a stranger's head moved oddly.

But read what the developer says it does, because it is not what the headlines say. The company ID covers each manufacturer's whole product line, so a VR headset in the next room triggers it. Bluetooth scanning on Android is unpredictable — bodies, walls and the angle a device is held at all move the reading. And the limit that matters most: it detects presence, not recording. The app can tell you a pair of smart glasses is probably nearby. It cannot tell you whether the camera is on.

Which means the one method with a radio behind it answers a different question from the one the reader asked. Even a mainstream headline saying the app "warns if you're being recorded" is describing something the app does not do.

That distinction matters more than it sounds. A presence alert and a recording alert would prompt completely different behaviour from a reader, and only one of them is what the phone is actually reporting. Someone acting on the stronger reading is acting on a claim no tool on the market makes.

What the people with a budget did instead

The most useful evidence about whether you can spot these glasses is not in any guide. It is in what happens when an organisation has to solve the problem for real.

DEF CON is a security conference. Its attendees are, professionally, people who notice things. It did not publish a spotting guide. It banned the device, and its stated reason was that there is no consistent way to tell whether a pair is recording. The ban makes no exception even for prescription lenses.

New York's court system reached the same place from a different direction, barring them from more than 1,240 courthouses and taking them at the door rather than asking anyone to judge from across a room. Four institutions have now done a version of this inside seven months.

And where a rule exists without a search at the entrance, an employment-law practice says plainly what it rests on: the honour system. The device cannot be remotely disabled. What it is running is invisible to everyone but the wearer. Unlike a phone held up in plain sight, it is far less conspicuous.

None of those parties concluded that observation works. Every one of them replaced observation with a rule, a search, or trust.

So what should a person actually do

The honest answer is unsatisfying, and worth having anyway.

You cannot reliably tell, in the moment, whether a stranger's glasses are recording you. Not from the light, which can be covered and is small even when it is not. Not from the lens, at conversational distance. Not from behaviour, which is easy to make look ordinary. Not from an app, which reports presence rather than recording and confuses a headset for a pair of glasses.

What you can know is when it is possible — which is roughly whenever anyone near you is wearing glasses, because the current generation shares a frame with ordinary eyewear rather than announcing itself the way the first attempt did. That is a worse answer than a five-item list, and it is the one supported by what the institutions taking the question seriously actually did.

There is a version of this that is more useful than a checklist, and it is closer to how people already handle being in public. You do not audit every phone in a train carriage for a running camera; you assume a phone can film and behave accordingly. The glasses have now joined that category, and the only thing that has changed is that the assumption no longer has a visible trigger. A detector app is worth having for the same reason a smoke alarm is worth having — it tells you something is present that you would otherwise have missed — as long as nobody mistakes it for a verdict about what that something is doing.

It is also worth taking the last caveat seriously. Acting on a suspicion — confronting somebody, filming them back, making an accusation in public — carries its own risk, and BGR's warning about that is the most practical line in any of these guides. A false positive on a phone is a notification. A false accusation is a different thing entirely.

The question people are asking is reasonable and the technology does not currently support an answer to it. That gap is the whole story, and it is the reason the rules keep arriving in a shape that looks blunt: a blunt instrument is what is left when the precise ones do not reach.

Related coverage

Is It Legal to Record Strangers With Smart Glasses? The Law Asks What They Expected — the legal tests behind the question, and why each of them needs the reader to know something the device does not show

What Does the Meta Glasses Recording Light Actually Cover? Read Meta's Own Wording — before asking whether the light can be seen, what the manufacturer says the light is reporting

Sources