From schools to restaurants and courts, Meta's AI smart glasses have drawn backlash for rampant unauthorised recording and privacy violations. But then, a Polish software developer came up with an app for iPhones and Android smartphones to scan nearby Meta smart glasses using Bluetooth signals. The app has been named 'Zuckoff', and it has reached approximately 5,000 users within a month of launching on Apple's App Store.
Paweł Szydłowski, a 30-year-old software developer, said he built the app after watching videos of people using camera-equipped smart glasses to film and identify unsuspecting strangers in public. The glasses in question are Meta's Ray-Ban smart glasses, a line of eyewear produced in partnership between Meta Platforms and the Italian eyewear manufacturer EssilorLuxottica, which owns both Ray-Ban and Oakley.
The glasses look nearly identical to conventional Ray-Ban frames but contain a forward-facing camera, microphones, and AI onboard. 'We as a society have the right to at least know that someone is recording,' Szydłowski told Business Insider.
How Zuckoff Detects Smart Glasses Using Bluetooth Fingerprints
The app works by scanning Bluetooth broadcasts from nearby devices and cross-referencing them against a catalogue of known signal signatures tied to specific smart glasses models.
Szydłowski built that catalogue himself. He purchased several models of smart glasses and methodically recorded their Bluetooth fingerprints, noting the unique device identifiers and manufacturer codes each model transmitted.
If both a device's unique identifier and its manufacturer identifier align with a known smart glasses signature, Szydłowski considers that 'an ultimate combo'. Some Bluetooth services on the glasses identify themselves using terms directly connected to Meta or Ray-Ban, which makes the fingerprinting process more reliable.
'Zuckoff' app for iOS, although free, also has a premium version. The free tier of Zuckoff displays an estimated proximity reading on its main scanning screen while the app is open. A paid upgrade for $24.99 or $9.99 yearly runs the detection process in the background and pushes a notification to the user's phone whenever it spots recording-capable glasses nearby, even when the app is closed.
European Union regulators have contacted Szydłowski seeking more information about the app, according to Medium, though the specific regulatory body involved and the nature of their inquiry could not be independently verified.
The Privacy Problem That Made the App Necessary
The concern Szydłowski is trying to address is not hypothetical. Meta's Ray-Ban smart glasses sit on the wearer's face in a form that is indistinguishable from ordinary spectacles to a casual observer. That visual ambiguity is central to the privacy debate surrounding them.
The stakes became clear in late 2024, when two Harvard University students demonstrated a system they called I-XRAY. The students streamed live video from Meta Ray-Ban smart glasses to Instagram, while a connected computer program used publicly available facial recognition software to identify strangers in real time and pull their personal information. The project was created explicitly to illustrate the privacy risks the technology enables, according to Forbes.
Szydłowski has cited the specific scenario of people using smart glasses to film women without their consent as a motivation for building the app. TikTok creators have used the AI glasses' technology to create POV videos, but in doing so, there have been instances of recording women without their consent.
This issue took a turn for the worse when young boys were using Meta AI glasses to harass and bully minor girls in high schools. A whole genre of 'rizzcam' is built around this concept of content.
The glasses' design compounds the problem. A camera embedded in eyewear carries none of the social signals that have historically cued bystanders to the presence of recording equipment.
Meta Disables Cameras on Thousands of Ray-Ban AI Glasses
Meta has not ignored the criticism surrounding its AI smart glasses, widely dubbed 'pervert glasses'. The company has disabled the cameras on thousands of its AI-powered glasses after reports emerged of users physically tampering with the built-in LED indicator light, which is designed to illuminate whenever the camera is actively capturing footage. The affected devices represent less than 0.1% of glasses sold, according to Meta.
The company has also removed online posts that provided instructions or offered services for tampering with the indicator lights. In a public statement, Meta said it would 'keep rolling out updates like this to make sure the capture LED can reliably alert bystanders when photos or videos are being captured for someone's gallery.'
The company also said that 'only a tiny minority of people try to get around these protections, and the fact they have to go to these lengths shows our efforts are working.'
Meta's chief technology officer, Andrew Bosworth, has stated that the camera was deliberately designed to be noticeable, positioning the LED as sufficient notice to anyone nearby.
The LED-tampering episode is significant context for what Szydłowski built. The indicator light is the only passive signal a bystander has that recording is occurring. If the light is covered or disabled, detection depends entirely on active countermeasures, which is precisely the gap Zuckoff fills. The app does not require the glasses to be recording: it flags their presence based on Bluetooth alone, regardless of whether the camera is active.
The technical limitation worth noting is that Bluetooth detection identifies presence, not intent. An app like 'Zuckoff' can tell a user that a pair of Meta smart glasses is nearby. It cannot determine whether the wearer is actively recording, and it cannot identify who is wearing them. The proximity estimate gives a rough sense of distance, but in a crowded environment, that information may be difficult to act on.
For Szydłowski, the more basic point holds regardless of those limitations. The glasses are in circulation, people are using them in ways that harm others, and the burden of awareness has fallen entirely on bystanders who have no way to know they are being filmed.
An imperfect detection tool is better than no tool at all.
