Women in Brussels filmed by smart glasses expose Europe’s wearable privacy gap

Women in Brussels say men wearing Ray-Ban Meta smart glasses filmed them during unsolicited street interactions, with footage allegedly used for social media content linked to dating advice businesses. The report matters because the device at the center of the complaint is not a specialist spy tool. It is a mass-market pair of fashionable glasses, sold through the language of convenience, creativity and artificial intelligence. The Brussels cases turn a familiar problem — women approached, recorded and judged in public — into a sharper legal and political test for Europe.

A street encounter becomes a privacy case

The facts reported from Brussels fit a pattern seen elsewhere: a man approaches a woman, speaks as if the exchange is casual, records the interaction from eye level, then treats the footage as content. The woman may not notice the camera. She may not understand that the small light in the frame is a recording indicator. She may learn later that the encounter was uploaded, edited, captioned, monetised or used to sell confidence courses, dating coaching or social media influence. The harm is not only that a camera was present. The harm is that the social setting was misrepresented from the start.

Ray-Ban Meta glasses make that misrepresentation easier. Meta says the glasses use a capture LED to tell others when photos or videos are being taken, and its help pages say media can be captured through the frame button or Meta AI voice commands. The product is built around point-of-view recording. That is the appeal for users who want hands-free videos of travel, sport, family events or daily life. It is also the danger when the wearer is not recording scenery but another person’s face, voice, body language and discomfort.

The Brussels report also lands at a difficult moment for Meta and EssilorLuxottica. EssilorLuxottica said its AI-glasses category sold more than 7 million units in 2025, putting the product far beyond the experimental stage. A device that sells in the millions no longer raises niche gadget questions. It changes the social expectations of streets, bars, campuses, trains, shops and public squares.

Belgian Minister for Digitalisation Vanessa Matz has called for a wider European discussion on smart devices that can record or identify people in public spaces. That wording is careful. It does not only target one product or one company. It points to a category: camera-equipped wearables, AI-enabled glasses, discreet recording devices and future consumer tools that blur the border between personal technology and public surveillance.

Europe already has laws for personal data, online platforms, biometric systems and video surveillance. The question now is whether those laws work when the camera is worn on someone’s face, the interaction feels personal, the upload is commercial, and the subject did not meaningfully know that recording was taking place. The Brussels cases are not a side story about gadget etiquette. They are a test of whether privacy law can still protect ordinary social life when recording becomes wearable.

The Brussels allegations fit a wider pattern of gendered recording

The Brussels report does not stand alone. Investigations in Germany, the United Kingdom and the United States have described similar uses of smart glasses by pickup artists, prank creators and street-content accounts. Women have reported being filmed in public conversations, at nightlife spots, on beaches and in places where the interaction itself was framed as flirtation, challenge or entertainment. German public broadcaster Tagesschau, citing SWR reporting, described women being secretly filmed in private or intimate situations with smart glasses and linked the phenomenon to “rizz” content, dating coaches and pickup-artist culture.

This matters because the technology does not create the social behavior from nothing. Men filming women in public without consent predates smart glasses. Street harassment, “pickup” videos, nightlife walk-through channels, prank content and voyeuristic social media accounts already existed. The device changes the economics and stealth of the practice. A phone held up in someone’s face gives a visible signal. Glasses let the wearer maintain eye contact while recording. The recording device moves from the hand to the face, and that changes the social contract.

A phone also gives bystanders a familiar cue. People understand when a phone is pointed at them. They can ask whether they are being filmed. They can turn away, block the lens or challenge the person holding it. Smart glasses are harder to read. The wearer may look like any other person in sunglasses or thick-framed spectacles. The lens is small. The light is small. The body language is normal. The recording can appear as ordinary conversation.

The gender dimension is not incidental. Women are often the subjects because the content is built around performance: rejection, awkwardness, flirting, embarrassment, sexualised commentary or advice about approaching women. The filmed person is not a collaborator. She becomes evidence for the creator’s persona. If she smiles, the footage can be packaged as proof of charm. If she rejects him, the footage can be packaged as a lesson, a joke or a target for hostile comments. The woman’s role is assigned after the fact by editing and captions, not agreed before the interaction begins.

The Guardian’s reporting on non-consensual “nightlife content” in the UK showed how similar recordings can attract sexualised comments, reputational harm and fear even when filmed in public places rather than private rooms. Smart glasses intensify that problem by lowering the effort needed to capture first-person footage of strangers who may not know they are part of a media product.

The Brussels cases also show the limits of treating this as a simple manners problem. A creator may say the conversation happened in public. A platform may say users can report content. A device maker may say the wearer is responsible for obeying the law. Each statement may be partly true. None addresses the full harm. The harm sits across product design, platform incentives, privacy law, gendered harassment, commercial use and social deception.

The product was designed to make recording feel normal

Ray-Ban Meta glasses are not marketed as surveillance equipment. They are presented as everyday eyewear with a camera, speakers, microphones, voice control and AI features. Meta’s product materials describe hands-free photos and videos, voice commands, point-of-view capture and use through the Meta AI app. The official Meta camera page says recent models use a 12MP ultra-wide camera, record up to 3K video on some models, and store clips on internal memory before users review and share them through the app.

The design goal is social acceptability. The glasses must look like glasses, not a head-mounted camera rig. That is why the Ray-Ban partnership matters. Ray-Ban is not only a brand name; it is camouflage through fashion. The more normal the device looks, the easier it is to wear in ordinary spaces. For legitimate users, that removes friction. For people recording strangers without clear notice, it removes warning signs.

Meta says the capture LED lets others know when recording is active. Its help page states that the capture LED signals photo and video capture, and its privacy page tells wearers to show others how the LED works and to keep it visible. That design choice admits a basic problem: people near the wearer need some way to know that capture is happening. The LED is meant to serve as that social signal.

The question is whether the signal is good enough. A recording light on the front corner of glasses is not equivalent to explicit notice. A person being approached may not see it. They may not know what it means. Bright sunlight, movement, distance, the angle of the wearer’s head, the subject’s focus on conversation and the short length of some clips all reduce its practical value. A notice mechanism that only works for people who already know what to look for is weak notice.

The device also makes recording socially frictionless for the wearer. A creator does not need to interrupt the moment, unlock a phone, open the camera, frame the subject and press record. Recording can be triggered from the glasses. The user’s hands remain free. The face remains engaged. That makes the resulting video feel “natural,” which is exactly why it appeals to creators. But the same naturalness makes the subject less able to detect the act.

This is the central tension. Meta wants the device to blend into life. Privacy protection often requires the opposite: visible friction, clear signals, meaningful consent and social interruption before someone becomes the subject of capture. The better smart glasses become as effortless cameras, the worse they become as tools for bystander awareness unless safeguards improve at the same pace.

A tiny light is carrying too much weight

Meta’s privacy defense for Ray-Ban smart glasses rests heavily on the capture LED. The company describes it as the feature that lets other people know the wearer is recording. Meta’s privacy page says that if the LED is covered, the wearer is notified to clear it before taking a photo, recording video or going live. Ray-Ban’s own FAQ also identifies the capture LED as an external-facing light that signals when a photo or video is being taken.

The problem is that an LED is a brittle safeguard. It depends on hardware integrity, software checks, user honesty, environmental visibility and public understanding. If any one of those fails, the protection weakens. A person who does not know the light exists cannot interpret it. A person who notices it too late cannot make a meaningful choice. A person filmed during a brief street exchange may have no time to respond.

Recent reporting has made the problem more concrete. 404 Media reported on a modification that disables the privacy-protecting recording light on Meta Ray-Bans, and The Verge reported that people were paying to remove the recording indicator LED. TechRadar reported that modders were turning Meta Ray-Bans into “spy glasses” by disabling or removing the light, while noting that such tampering violates Meta’s terms and may raise legal risks depending on jurisdiction.

The existence of tampering tutorials and paid modifications changes the regulatory analysis. A manufacturer cannot be blamed for every misuse of a product, but regulators will ask whether predictable misuse was designed against seriously enough. The more public the workarounds become, the harder it is to describe them as unforeseeable edge cases. A safeguard that can be defeated cheaply, explained publicly and hidden from the people at risk may not satisfy the level of care expected from a mass-market camera worn in public.

The LED also has a deeper limitation: it is a signal, not consent. Someone may notice the light but not understand whether the recording is for private memory, live streaming, commercial content, AI processing, platform upload or later editing. GDPR, consumer protection and civil law questions often depend on purpose, sharing, storage and identifiability. A light cannot convey those details.

A phone camera at least carries a familiar cultural meaning. A lens held toward a person usually means capture. A smart-glasses LED must teach a new cultural signal at population scale. That education has not happened. Most people on a Brussels street are not reading Meta help pages before speaking to a stranger. They are living their day. Product design that puts the burden of recognition on non-users creates an unfair distribution of risk.

The difference between public filming and public processing

Some defenders of street filming argue that people in public should expect to be seen, photographed or recorded. That view contains a practical truth but misses the legal issue. Being visible in public is not the same as being processed as content. A passer-by may be seen by hundreds of people in a square. That does not mean a stranger can secretly capture her face and voice, upload the clip to a platform, use it to promote a business and expose her to comments, identification or harassment without any legal question.

GDPR focuses on personal data, not only secrecy. The regulation defines personal data as information relating to an identified or identifiable natural person, and it applies to automated processing as well as some filing-system processing. A recognisable face, voice, location context, clothing, workplace badge, name exchanged in conversation or visible social media handle can turn a street clip into personal data. If the footage is uploaded and used beyond a purely personal setting, the legal analysis becomes more serious.

The household exemption is often misunderstood. GDPR does not apply to processing by a natural person in a purely personal or household activity. But the European Data Protection Board’s video-device guidance interprets that exemption narrowly. It gives the example of a tourist recording holiday videos for friends and family as household activity, while explaining that public accessibility and extension into public space can change the assessment.

That distinction is central to the Brussels allegations. A wearer recording a private memory for personal use is one scenario. A creator recording women during unsolicited approaches and posting clips to promote dating advice content is another. The second scenario involves publication, audience growth, commercial positioning and platform distribution. Once a person’s likeness becomes material for a business model, the argument that this is merely personal recording becomes far weaker.

European privacy law does not ban all public photography. It would be absurd and harmful if it did. Journalists, tourists, artists, activists, documentary makers, witnesses and ordinary people all rely on the ability to record public events. The issue is not whether any person can ever be filmed outdoors. The issue is whether covert or unclear recording of identifiable people, especially for commercial social media content, meets legal duties around fairness, transparency, lawful basis, proportionality and rights of objection.

Smart glasses complicate the boundary because they make purposeful recording look incidental. If a phone is pointed at a protest, a performer or a public event, the role of the camera may be understood. If a stranger in ordinary glasses approaches one woman and films her reaction for content, the recording is targeted even if the location is public. Public space does not erase the dignity interests of the individual subject.

GDPR gives regulators several routes to intervene

GDPR gives European regulators more than one route into smart-glasses misuse. Article 5 requires personal data to be processed lawfully, fairly and transparently, collected for specified purposes, limited to what is necessary and kept no longer than needed. Article 6 requires a lawful basis. Article 9 restricts processing of special categories of data, including biometric data used for unique identification.

A dating-content creator who records women without clear notice and posts the footage online may face questions under several principles. Was the subject told she was being recorded? Was the purpose clear? Was the footage necessary for that purpose? Was there a lawful basis for publishing identifiable footage? Was the person given a practical way to object before publication? Did the clip reveal sensitive details? Did the creator use the footage for commercial gain? Did the platform process or recommend it in ways that amplified harm?

Consent is the most obvious lawful basis, but genuine consent is hard in these encounters. Consent under GDPR must be informed, specific, freely given and unambiguous. A woman who only learns at the end of a conversation that she has been filmed has not consented before recording. A woman asked for permission after an awkward encounter may feel pressured. A woman who says nothing because she does not notice the LED has not agreed. Silence in front of a hidden or poorly understood wearable camera should not be treated as permission.

Some creators might claim legitimate interest. That argument is not automatically impossible, but it requires balancing the creator’s interest against the rights and freedoms of the filmed person. A public-interest journalist documenting wrongdoing has a stronger claim than a dating coach manufacturing social proof. A creator seeking engagement from clips of uncomfortable women has a weak privacy argument when the same business goal could be served through consenting participants, staged demonstrations or anonymised footage.

The Belgian and European data-protection authorities could also examine whether device design contributes to foreseeable unlawful processing. GDPR usually focuses on controllers and processors, but privacy-by-design duties under Article 25 matter for products that process personal data. If a company designs a mass-market wearable camera, advertises easy social media capture and relies on a small LED despite known tampering, regulators may ask whether design choices sufficiently reduce foreseeable risks.

The law will not produce one simple answer for every smart-glasses clip. A family video, a tourist panorama, a newsworthy street event and a targeted pickup video are not the same. The value of GDPR is that it forces purpose, transparency and proportionality into the analysis. That is exactly what covert street-content culture tries to avoid.

The legal pressure points around smart-glasses recording

The legal risk does not come from one fact in isolation. It grows when targeted recording, poor notice, identifiable subjects, upload, monetisation and platform reach combine in the same workflow.

Belgian law faces the same boundary problem as other countries

Belgium has privacy, data-protection and criminal-law tools, but smart-glasses filming exposes the same boundary problem seen across Europe: many laws distinguish sharply between private spaces and public spaces, while the harm of modern content can occur in public. Belgian rules on voyeurism and non-consensual intimate images can be severe when the footage concerns nudity, sexual acts or intimate material. Cornell’s Gender Justice database summarises Belgian criminal provisions on voyeurism as covering observation or recording of a nude person or someone engaged in sexual activity without consent.

The Brussels cases, as reported, involve unsolicited street interactions. They may be humiliating, invasive or commercially exploitative without always meeting the legal threshold for voyeurism or intimate-image abuse. That gap is familiar. UK debates around nightlife filming have shown how women may feel violated by public recordings that are sexualised by framing, captions and comments but do not easily fit older criminal categories.

Belgium’s Data Protection Authority can still matter because public footage may be personal data when people are identifiable. The Belgian DPA’s mission statement says it ensures compliance with fundamental data-protection principles. Past Belgian DPA activity on video images has included fines where camera positioning unlawfully captured personal data and breached data-protection-by-design principles. That history is not identical to smart glasses, but it shows the authority’s willingness to examine how cameras are aimed, justified and designed.

The hard part is enforcement against individual creators. A shop camera, employer surveillance system or municipal installation has a stable controller. A man filming women with smart glasses may operate through social accounts, shell brands, deleted videos or cross-border platforms. He may be in Belgium, target audiences elsewhere and use services hosted in another EU state or outside Europe. Wearable recording turns privacy enforcement into a moving-target problem.

Minister Vanessa Matz’s call for a wider European discussion makes sense because unilateral national action may be too narrow. Belgium can enforce its own laws. Brussels police can examine harassment or stalking complaints where facts support them. The Belgian DPA can handle data-protection complaints. But the product is sold across borders, the platform systems are multinational, and the design choices are made by Meta and EssilorLuxottica at scale.

A European conversation also reduces the risk of fragmented public-space rules that confuse users and victims. The point should not be to criminalise every tourist photo. The point should be to identify high-risk conduct: covert or disguised recording, targeted filming of individuals for humiliation or sexualisation, commercial publication without consent, LED tampering, biometric identification and platform monetisation of non-consensual encounters.

The Digital Services Act matters once the footage is uploaded

The Brussels allegations become a platform issue once footage moves from glasses to social media. The Digital Services Act requires online platforms to handle illegal content notices, protect fundamental rights and, for very large platforms, assess and mitigate systemic risks. The European Commission says the DSA strengthens users’ rights online and requires platforms to reduce risks of exposure to illegal and harmful content, including for children and young people.

A video of a woman filmed without consent may not always be illegal under every national law. That is part of the difficulty. But platforms still face policy and risk questions. Did the subject have a clear way to report the clip? Could she report as a non-user? Would the platform require her to reveal more personal information to prove harm? Would the platform remove reuploads? Would recommendation systems keep pushing the clip while a complaint is pending? Would the creator be allowed to monetise or advertise a dating course through the same account?

The DSA is not a magic removal button. It does not require platforms to pre-screen every video. It must be read alongside free expression, journalistic use and legitimate public recording. Yet the DSA can force platforms to build processes that do not leave filmed subjects helpless. A woman who discovers her face in a viral pickup video should not need to become a privacy lawyer to ask for removal.

The DSA also gives regulators a way to look at systemic incentives. If platforms reward street-interaction videos that feature uncomfortable strangers, then the harm is not only one user’s behavior. It is an attention system. Algorithms that boost confrontation, sexualised embarrassment or rejection content can create demand for more covert recording. The creator records because the platform rewards the format. The platform benefits because the format keeps users watching.

Meta is relevant twice. It makes the glasses and operates Instagram and Facebook. If Ray-Ban Meta glasses produce content for Meta platforms, the company cannot treat hardware and platform governance as entirely separate moral worlds. A device can capture. An app can import. A platform can recommend. Ads and creator businesses can monetise. The same ecosystem can turn a private encounter into a distribution event.

That does not mean Meta is automatically liable for every abusive video shot on its glasses. It means regulators can ask whether its product design, app defaults, reporting tools, platform rules and enforcement practices fit together. A coherent safety model would follow the footage from capture to upload to recommendation to takedown, instead of treating each stage as someone else’s responsibility.

The AI Act enters the debate through identification, not ordinary video

The EU AI Act is not the main law for every smart-glasses recording. A standard video clip shot on glasses is not automatically an AI system. But the AI Act becomes relevant when smart glasses move from capture to recognition, inference or automated decision support. European AI Act materials classify certain biometric uses as high-risk and restrict real-time remote biometric identification in publicly accessible spaces for law enforcement under strict conditions.

The distinction matters. A wearable camera records. An AI layer can identify, classify, search, summarise, translate, remember or connect that recording to other data. The privacy risk changes sharply when glasses do not merely capture a face but help the wearer know who the person is. Recording without consent is invasive. Recording plus identification can become street-level doxxing.

WIRED reported in June 2026 that Meta had embedded face-recognition code, internally called “NameTag,” into the Meta AI companion app for smart glasses, though the feature was not enabled for consumers. WIRED’s analysis said the system could identify people captured by the glasses’ camera and alert the wearer when someone is recognised, with models for detecting, cropping and encoding faces into biometric data. Meta told WIRED that such a feature had not been released to consumers and that it would take a careful approach if it were to release one.

Even inactive code affects the policy debate because it reveals product direction. The fear is not science fiction. The device already has a camera and microphones. The app already connects to Meta services. Research and consumer software already show what face recognition can do. U.S. senators Markey, Wyden and Merkley demanded transparency from Meta in March 2026 over reported plans to integrate facial recognition into smart glasses. ACLU and partner organisations later warned that facial recognition in Ray-Ban and Oakley glasses would threaten privacy and liberty.

For Europe, the AI Act and GDPR will overlap. Biometric data used for unique identification receives special protection under GDPR. The AI Act adds product and deployment rules for AI systems. The hard case is consumer-to-consumer identification: not police scanning a square, but thousands or millions of private users wearing glasses that could recognise strangers. Existing laws were not written with that exact model at the center.

That is why the Brussels filming cases and the facial-recognition debate are connected. Today’s harm is covert recording for content. Tomorrow’s harm could be covert recognition for targeting, stalking, commercial profiling or harassment. Regulators who wait until recognition is fully deployed may find that public norms have already shifted.

Facial recognition would turn nuisance into infrastructure

Facial recognition in smart glasses would change the device category. Without recognition, the glasses are a discreet camera and audio device. With recognition, they can become a wearable identity interface. The wearer could move through a street, conference, bar or political event and receive names or saved profiles linked to people in view. Even if Meta designed such a feature around opt-in contacts, the pressure to expand recognition would be strong because identity is one of the most valuable context layers for AI assistants.

The most dangerous misuse cases are obvious. A stalker could identify a target who changed social circles. An abuser could recognise someone in a crowd. A political extremist could identify activists. A predatory creator could attach names and profiles to women filmed in public. A fraudster could use real-time identity cues to make social engineering more convincing. The moment glasses connect faces to names, the wearer gains asymmetric power over people who may not know they are being scanned.

Supporters sometimes point to accessibility. Face recognition can help blind or visually impaired people identify friends, colleagues or people who have consented to be recognised. That use is real. Research on smart glasses and accessibility has explored tools that assist visually impaired users with scene description, text recognition and other daily tasks. The ethical mistake would be to use accessibility as a blanket justification for mass consumer facial recognition without strict limits.

A safer accessibility model can be narrow: recognition only for enrolled contacts, no unknown-person identification, on-device processing where possible, no central face database, clear audible or visible notices, strong deletion rights and independent audits. A risky model is broad: background recognition, weak consent, cloud processing, unclear databases and vague user controls. The difference is not philosophical; it is architectural.

Meta’s history makes trust harder. WIRED noted that Meta previously announced the shutdown of Facebook’s face-recognition system in 2021 and agreed to major biometric settlements in Illinois and Texas. The company will argue that new systems can be designed differently. Critics will answer that the burden of proof belongs to the company seeking to put recognition inside everyday eyewear.

Europe’s concern should not be limited to Meta. Once one major company normalises recognition glasses, competitors may follow. Smart eyewear is an industry race involving Meta, EssilorLuxottica, Snap, Google, Samsung, Apple-adjacent ecosystems and many smaller AI hardware makers. The regulatory question is broader than one brand: should consumer eyewear be allowed to identify strangers in public at all, and if so under what conditions?

Commercial dating advice changes the consent analysis

The alleged use of Brussels footage to promote dating advice businesses is not a minor detail. It changes the privacy and consumer-protection analysis. A private individual recording a conversation for personal memory is one thing. A creator recording a woman’s reaction to sell a method, course, persona or coaching service is another. The subject is no longer incidental; she becomes unpaid commercial material.

Commercial use increases the need for clear consent. In advertising, influencer content and promotional materials, recognisable people usually require stronger permissions because their image is being used to support someone else’s business. The creator may not call the video an advertisement, but if the account drives sales, builds a funnel, promotes paid coaching or establishes authority in the dating-advice market, the woman’s likeness is part of the commercial asset.

This is especially troubling in “approach” content because the woman is often the proof. Dating coaches sell confidence, technique, scripts, status and social success. A video of a real woman reacting on the street is more persuasive than a staged example because it feels authentic. The business value comes from her realness and her lack of preparation. The same lack of consent that makes the clip ethically suspect can make it more marketable.

The power imbalance is sharp. The creator controls the camera, the edit, the caption, the upload timing, the comments, the deletion decision and the business link. The woman controls almost nothing unless she discovers the clip and successfully demands removal. Even then, copies, stitches, duets, screenshots and commentary videos can remain. A short unsolicited interaction can outlive the moment by years.

Platforms also benefit commercially. Viral pickup videos drive views, comments, shares and ad inventory. A creator’s business may sell coaching, while the platform sells attention. Device makers benefit if the content demonstrates a popular use case, even when the use case is socially corrosive. This chain of incentives is why voluntary etiquette is not enough. No single actor has a strong financial reason to slow the content unless law, platform rules or public backlash make the cost higher than the revenue.

Regulators should look at the entire commercial pipeline. Was the video used to advertise a paid service? Were affiliate links or courses promoted? Did the platform allow monetisation? Did the subject have a removal path? Did the account repeatedly film women without permission? Repetition matters because one clip might be defended as poor judgment; a business model built on such clips points to systematic exploitation.

Smart glasses make bystander privacy a product-design issue

Bystanders have always been the neglected people in consumer technology. Phones, drones, doorbell cameras, dashcams, livestreaming apps and smart speakers all collect data about people who did not buy the device and may not even know it is active. Smart glasses bring the bystander problem into the most intimate public setting: face-to-face interaction.

The wearer has controls. The bystander has guesses. The wearer can read privacy settings, disable or enable cloud features, manage uploads and delete clips. The bystander may not know the glasses are smart. They may not know which company made them. They may not know whether recording is local, cloud-based, live-streamed or AI-processed. The person at greatest privacy risk may be the person with no interface.

Product designers often focus on user privacy because users are customers. They ask whether the wearer can control data sharing with Meta, whether the wearer can delete voice recordings, whether the wearer can manage cloud media. Those controls matter. But they do not solve the Brussels problem, where the subject of the footage is not the customer. A privacy dashboard for the wearer does not give a filmed stranger control over her image.

A serious bystander-safety model would ask different questions. Can non-users recognise recording at a glance? Can they request deletion easily? Can the app detect and discourage repeated close-range recording of strangers? Can uploads of targeted interpersonal footage trigger stronger consent prompts? Can platforms require blur tools before publication? Can a disabled LED make capture impossible at hardware level rather than merely producing a warning? Can retailers, repair shops and marketplaces be prevented from offering LED-removal services?

Some of those ideas may be technically hard or legally contested. But the starting point must shift. Meta’s current public materials place responsibility on the wearer: be respectful, follow the law, show the LED, avoid sensitive spaces. That is necessary but weak. Safety cannot depend mainly on the goodwill of the person choosing to record.

The strongest product interventions are those that do not rely on user virtue. A car seatbelt works even when the driver is careless. A payment card chip reduces fraud without asking every merchant to be a cryptographer. Smart glasses need equivalent thinking for bystander privacy. The LED was one attempt. The evidence of tampering, confusion and misuse suggests it is not enough.

Europe has seen this story before with Google Glass

The backlash against Ray-Ban Meta glasses recalls Google Glass, but the comparison has limits. Google Glass was visually obvious, expensive, socially awkward and ahead of its consumer market. It triggered the term “Glasshole” because people felt uneasy around a head-mounted camera and display. Ray-Ban Meta glasses are subtler, cheaper, more fashionable and tied to creator platforms that did not exist in the same form during the Google Glass era.

Google Glass failed partly because people noticed it. Ray-Ban Meta glasses may succeed partly because people do not. That is the central historical difference. A bulky futuristic device invites social resistance. A familiar Ray-Ban frame can pass through the same spaces as ordinary eyewear. The privacy risk rises when the social warning signs disappear.

The second difference is AI. Google Glass was a wearable computer and camera. Ray-Ban Meta sits inside the current AI race: voice assistants, multimodal models, real-time translation, visual question answering, object recognition and possible memory features. Research published in 2026 on always-on AI agents through smart glasses describes systems that continuously perceive real-world context and connect perception with task execution. That direction makes glasses less like cameras and more like sensors for ambient computing.

The third difference is distribution. Google Glass never became a mainstream eyewear product. EssilorLuxottica’s 2025 disclosure that AI glasses sold more than 7 million units in one year means the category has crossed into mass adoption. A social norm that might have been negotiated slowly among early adopters is now being forced quickly into public life.

The old lesson still applies: people resist technology that makes them feel watched without consent. But the response cannot be nostalgia for a world without cameras. Phones already made cameras ubiquitous. The new challenge is subtler. Smart glasses make recording persistent, face-level and socially disguised. They also merge hardware with platforms and AI. That fusion requires more precise rules than the blunt “ban it” debates that followed Google Glass.

Europe should learn from that earlier backlash without repeating its vagueness. The point is not to stigmatise all smart-glasses users. Some use the devices for accessibility, safety, travel, hands-free work or family memories. The point is to prevent a small but visible set of abusive practices from defining the category and harming people who never chose to participate.

The market is growing faster than public understanding

EssilorLuxottica’s 2025 results put the business stakes in plain view. The company reported full-year revenue of €28.491 billion and said AI glasses sold more than 7 million units during the year. That volume matters for regulation because risk scales with adoption. A privacy issue affecting a few thousand early adopters is one enforcement problem. A privacy issue built into millions of wearable cameras is a public-space governance problem.

The eyewear industry also gives smart glasses a retail channel that earlier wearables lacked. These products can be sold through optical stores, fashion channels and prescription-lens pathways. A buyer may not think of them as surveillance-adjacent electronics. They are glasses. That normalisation is commercially powerful. It also means public education will lag behind adoption.

Most people do not know how to identify Ray-Ban Meta glasses. They do not know which side contains the camera. They do not know what the capture LED looks like. They do not know whether voice commands can start recording. They do not know how to complain if footage appears online. A technology can become socially widespread before society learns its basic signals. That gap is where misuse thrives.

Creators understand the device faster than the public does. A person building social media content has strong incentives to learn recording gestures, LED behavior, editing workflows, upload formats and platform tactics. The people being filmed are ordinary passers-by. The knowledge imbalance favors the recorder.

The market’s growth also raises a competition problem. If privacy safeguards are too weak but the product sells well, rivals may copy the weak model. A stronger LED, louder capture sound, visible recording band or consent workflow might protect bystanders but make the product less “cool.” Without regulation or industry standards, companies may fear that stronger signals will hurt adoption. The market may reward discretion even when discretion creates social harm.

That is why voluntary design principles can be insufficient. Companies often support privacy in the abstract, but the most protective choices can conflict with product appeal. A regulator can reset the competitive field by requiring minimum standards for all camera-equipped glasses. If every brand must provide visible recording indicators, tamper resistance and bystander reporting mechanisms, no company is punished alone for doing the safer thing.

The LED-tampering problem demands a hardware answer

LED tampering deserves its own regulatory focus because it is the clearest sign that current safeguards can fail. A wearer who disables the recording light is not merely forgetting etiquette. They are removing the feature designed to alert others. The intent is usually obvious: to record without being noticed. That conduct should be treated differently from ordinary recording.

Meta says covering the LED triggers a notification and blocks capture until the obstruction is cleared. Reports of physical LED removal or workarounds suggest that software checks may not be enough. Once a user can drill out, mask, replace or otherwise defeat the light, the device’s public-facing privacy promise weakens.

A hardware-level response could include stronger tamper detection, secure component checks, capture shutdown when LED integrity fails, repair locks, warranty invalidation tied to reporting, retailer obligations and marketplace bans on LED-disabling services. Regulators may also examine whether offering LED-removal services should be treated like selling covert recording modifications. When the safety signal is deliberately removed, the product is no longer operating as represented to the public.

The challenge is that right-to-repair principles also matter. Users should not be blocked from legitimate repairs. But privacy-sensitive components require special handling. A phone camera can be repaired without turning the phone into a hidden camera. Smart-glasses LED repair or removal affects non-users directly. That gives regulators a stronger reason to require certified repair pathways for the capture indicator.

Tamper resistance also needs public accountability. Companies should not simply say the LED cannot be disabled if independent reporting shows workarounds. They should publish enough information for regulators and researchers to assess whether the safeguard works. Independent testing could examine visibility in different lighting, recording distances, skin tones, frame colors, head angles and modified devices.

A weak LED does not mean smart glasses must be banned. It means the LED cannot be the only line of defense. Capture sounds, stronger visible indicators, app-level consent prompts, automatic face blurring, platform upload checks and strict penalties for tampering can work together. Safety needs redundancy because no single safeguard will survive every misuse.

Platforms must stop treating non-consensual street content as ordinary engagement

The social media layer is where private discomfort becomes public harm. A woman may feel uneasy during an unsolicited approach, but the larger injury often begins when the clip is uploaded. Comments sexualise her, mock her, identify her or debate whether she was “rude.” The creator’s audience turns a short interaction into a permanent spectacle. The platform’s systems may then recommend the video to more viewers because conflict performs well.

Platforms have policies against harassment, bullying, non-consensual intimate content and privacy violations. The problem is fit. A street video may not be intimate in the narrow sense. It may not contain threats. It may not reveal a home address. The creator may claim it is comedy, dating advice or public interaction. The content can be harmful precisely because it lives between existing policy boxes.

Platforms should create clearer rules for targeted recording of non-consenting individuals in public when the content is used for humiliation, sexualisation, harassment or commercial promotion. Such rules need careful exceptions for journalism, public-interest documentation, political speech and recording of misconduct. But the difficulty of drafting exceptions is not a reason to ignore the harm.

A useful platform rule would look at context rather than one visual element. Is one identifiable person the focus? Was the interaction initiated by the creator? Is the subject unaware of recording? Does the caption sexualise or mock them? Are comments predictably abusive? Is the account repeatedly filming similar subjects? Is the video promoting paid services? Did the subject ask for removal? Those factors are familiar to trust-and-safety teams.

Platforms also need faster non-user complaint channels. A filmed person may not have an account. They may not want to send identity documents to a platform that already failed to protect them. They may fear retaliation if the creator learns they complained. A good reporting process would allow privacy claims with minimal data, offer emergency review for viral clips and reduce reupload risk.

Monetisation should be the first lever. Even where removal is legally contested, platforms can deny ad revenue, affiliate promotion or algorithmic boost to non-consensual street-content accounts. If creators cannot profit from ambushing strangers, the format loses part of its appeal. The fastest way to shrink abusive content is often to remove its reward.

Public-space privacy is not the enemy of journalism

Any proposal to regulate smart-glasses recording must protect journalism, documentation of police conduct, evidence gathering, protest coverage and ordinary public-interest filming. Authoritarian governments often misuse privacy language to restrict scrutiny. Europe should not create rules that make it harder to film abuse of power. The line between harmful recording and public-interest recording must be drawn with care.

The Brussels cases are different from journalism in several ways. The reported purpose was not exposing wrongdoing, informing the public or documenting a newsworthy event. It was allegedly content for social media and dating advice businesses. The subjects were not public officials acting in official roles. They were women approached in public. The value of the clip came from their unprepared participation.

Journalism also has professional duties: editorial purpose, minimisation, right of reply in some cases, public-interest balancing, newsroom accountability and legal review. Amateur creators do not automatically lose speech rights, but they cannot claim journalistic protection merely because they upload video. Calling something “content” does not give it the public-interest status of reporting.

Regulators should avoid rules that ban recording based solely on location. Public recording is sometimes necessary. Better rules would target covert recording of identifiable individuals for commercial, sexualised, humiliating or harassing distribution, especially where the subject is the main focus and has no meaningful notice. That approach focuses on conduct, purpose and harm rather than banning cameras in public.

There is also a difference between incidental capture and targeted capture. A tourist filming Grand Place may capture hundreds of faces in the background. A creator approaching one woman with smart glasses and using her reaction as the content has targeted her. GDPR and platform policies can reflect that distinction.

Public-interest exceptions should be explicit. So should anti-abuse provisions. Europe does not need to choose between a surveillance free-for-all and a ban on public documentation. It needs rules that recognise why filming a police assault, filming a street performer and secretly filming a woman for dating-content promotion are different acts.

Police and regulators need usable complaint pathways

Victims of covert smart-glasses filming face practical barriers. They may not know who filmed them. They may not know the account name. They may have only a screenshot from a friend. They may feel embarrassed. They may worry that reporting will spread the video further. They may be told that filming in public is legal, even when upload, monetisation or harassment could raise separate issues.

Belgium and other EU countries need clear guidance for police, data-protection authorities and equality bodies. A person should be able to understand which pathway fits the harm: harassment, stalking, voyeurism, non-consensual intimate imagery, image rights, GDPR complaint, platform report or civil claim. Without guidance, victims are bounced between institutions.

Police need training on smart glasses as recording devices. Officers should know how to identify the hardware, ask whether footage exists, preserve evidence, document platform links and assess whether LED tampering or repeated targeting is involved. This does not mean every complaint becomes a criminal case. It means the first response should not dismiss the harm because the camera was worn rather than held.

Data-protection authorities need accessible templates for complaints about public videos used commercially. A GDPR complaint should capture key facts: who recorded, where, whether the person was identifiable, whether they knew, where it was posted, whether it promoted a business, whether deletion was requested, and whether the platform responded. The easier it is to file a precise complaint, the easier it is for regulators to detect patterns.

Platforms should be required to preserve evidence once a privacy complaint is filed. If a creator deletes a video after complaint but continues the practice, regulators may need records. Preservation must respect rights and limits, but total disappearance benefits repeat offenders.

Civil-society organisations can help by creating victim guides. Women’s rights groups, digital rights groups and consumer organisations understand different parts of the problem. The Brussels cases sit at their intersection. A strong response will require privacy expertise, gender-based violence expertise, platform governance knowledge and local legal support.

Brussels is a symbolic location for this debate

Brussels is not just another city in this story. It is Belgium’s capital and the de facto capital of the European Union. A privacy dispute on Brussels streets naturally becomes a European policy question because the institutions that write and enforce digital rules are nearby. The symbolism is hard to miss: a device shaped by Silicon Valley and European eyewear manufacturing meets women’s everyday safety in the same city where GDPR, the DSA and the AI Act define Europe’s digital identity.

The EU often presents itself as the jurisdiction that protects fundamental rights in technology markets. GDPR became a global reference point. The DSA and AI Act are part of the same regulatory identity. Smart glasses test whether that identity applies outside websites and apps. The next privacy frontier is not only online. It is the sensor layer that feeds online systems from the physical world.

Brussels also represents urban density. Smart glasses are most privacy-sensitive where people are close together: metro stations, cafés, nightlife districts, shopping streets, campuses, tourist areas and political gatherings. Dense public life depends on partial anonymity. People can be visible without being turned into searchable content. They can talk to strangers without becoming a clip. They can reject an approach without becoming an example for an audience.

Wearable cameras put that partial anonymity under pressure. Not because every wearer is malicious, but because the cost of recording drops so low. Public life changes when anyone might be filming from their face, especially if the subject cannot tell. People may become more guarded, less open, less willing to interact, less willing to attend sensitive events or more suspicious of strangers.

That chilling effect is difficult to measure but real. Women already adapt their behavior in public because of harassment risk. If smart glasses make unsolicited approaches recordable and publishable by default, the burden grows. A technology sold as frictionless for the wearer adds friction to everyone else’s public life.

Brussels can therefore become a useful policy case. The city’s experience can help EU institutions ask grounded questions rather than abstract ones. What do people notice? What do they miss? How do victims discover footage? How do platforms respond? What evidence do police need? How do creators monetise? What product changes would reduce harm without banning legitimate use?

The device has legitimate uses that deserve protection

A serious critique of Ray-Ban Meta misuse must not erase legitimate uses. Smart glasses can help people capture hands-free memories, communicate while moving, translate signs, listen to audio, ask AI about their surroundings and support accessibility. Meta’s help pages describe capture, calls, messaging, audio and AI features. Research has explored smart glasses for blind and visually impaired users, including visual assistance through scene description, OCR and object recognition.

These uses matter because regulation that ignores benefits will fail politically and practically. A person with a disability may rely on wearable cameras to understand spaces. A cyclist may use hands-free recording for safety. A parent may record a child’s birthday from eye level. A worker may need remote expert support. A traveller may translate a menu. The goal should be to protect those uses while narrowing the space for covert exploitation.

The problem is not the existence of a camera on glasses. It is the combination of discreet capture, weak bystander notice, social media incentives, commercial use of strangers, LED tampering and potential biometric identification. A good policy response should be specific enough to address that combination.

Accessibility also strengthens the case for careful design. People who rely on smart glasses should not lose access because companies ignored privacy risks and triggered public bans. If Meta and rivals want smart eyewear to become accepted, they need bystander trust. Public backlash can harm disabled users too by making venues ban devices indiscriminately.

A better industry approach would separate high-trust features from high-risk features. Basic capture with clear indicators, strong tamper resistance and easy deletion rights is one category. Always-on recording, unknown-person identification and commercial publication of targeted encounters are another. The second category deserves stricter controls.

Legitimate users also need social clarity. If people understand the recording signal and trust that tampered devices are rare, responsible wearers face fewer confrontations. Weak safeguards create suspicion toward everyone wearing smart glasses. Privacy protection is not anti-innovation; it is the condition for social acceptance.

The smart-glasses industry needs standards before scandals set them

Industries often resist regulation until scandals produce harsher rules. Smart glasses still have a chance to set standards before public anger hardens. The Brussels cases, German reports, U.S. investigations and facial-recognition debate are warnings. They show predictable misuse while the category is still young enough to change.

A credible standard for camera-equipped glasses should include visible recording indicators that are brighter, harder to confuse and tested in real conditions. It should require hardware tamper detection. It should make recording impossible if the indicator fails. It should ban designs that mimic ordinary eyewear too closely while hiding cameras and lights. It should require clear packaging and retail education so buyers understand duties to others.

App standards matter too. The app could warn users before first use that recording identifiable strangers for upload may require consent or another lawful basis. It could present special warnings before sharing close-range footage of people. It could provide automatic face-blurring tools. It could log whether the LED was functioning during capture. It could detect repeated uploads of targeted street interactions and require stronger review.

Platform standards are equally important. Instagram, TikTok, YouTube and Facebook should adopt shared rules for non-consensual wearable-camera content. If only one platform cracks down, creators migrate. A cross-platform norm would reduce evasion. The DSA could support this by pushing very large online platforms to treat the issue as a systemic risk where evidence supports it.

Retailers and repair markets need rules. Selling smart glasses is not like selling ordinary sunglasses. Customers should receive a short privacy notice, not just a product manual. Repair services should not be allowed to disable capture indicators. Online marketplaces should remove listings that advertise covert-recording modifications.

Independent auditing would make standards credible. Companies should not self-grade their own privacy lights. Regulators, consumer groups and technical labs can test devices. Results should be public. A product that records the public should be accountable to the public, not only to its buyer.

Privacy by design must include people outside the transaction

GDPR’s privacy-by-design principle is often discussed in terms of data minimisation, security and default settings. Smart glasses demand a wider reading. The design must protect people who did not buy, install or agree to anything. This is harder than user privacy because non-users have no account relationship. But it is not optional if the device collects their personal data.

The Brussels allegations illustrate a failure mode: the filmed person is not in the transaction. Meta sells to the wearer. EssilorLuxottica manufactures and distributes. The creator may monetise through platforms. The audience consumes. The woman supplies the content involuntarily. The person whose rights are at stake is economically absent from the system.

Privacy by design should correct that absence. At minimum, it should make recording obvious. Better, it should make targeted recording of strangers harder to upload without reflection. Stronger still, it should give filmed people practical remedies after publication. Design can shape behavior by adding friction at the right moments.

Friction is not always bad. A confirmation prompt before uploading a close-range clip of a stranger is friction. A default blur option is friction. A warning that commercial use may require permission is friction. A brighter recording indicator is friction. For creators who film with consent, these steps are manageable. For creators who rely on ambush, friction is the point.

Companies may argue that automated detection of strangers is imperfect. True. But perfection is not required to improve defaults. Platforms already use imperfect systems for copyright, nudity, violence, spam and child-safety signals. Privacy risks deserve similar effort, especially when the product’s intended workflow is capture and share.

Privacy by design also means avoiding future features that make bystander risk worse. Always-on “memory” modes, persistent environmental scanning, unknown-person identification and automatic social profile lookup should face strict review before launch. A company should not deploy first and ask society to adapt later.

The Brussels cases expose a consent gap, not a confusion gap

Some technology companies respond to privacy criticism with education campaigns: teach people what the LED means, tell users to behave responsibly, publish safety tips. Education helps, but the Brussels cases are not mainly caused by confusion. They are caused by a consent gap. The subjects did not choose to be recorded for someone else’s content.

A better-informed public might spot more smart glasses. But asking women to monitor every stranger’s frames is unreasonable. Public safety should not require constant device literacy. People should not need to scan for lens placement before answering a question in the street. A consent system that works only when the subject detects the device is not a consent system.

True consent in street-content scenarios would require the creator to ask before recording or at least before using identifiable footage. “I’m making a video for my dating advice channel; are you comfortable being recorded and posted?” That sentence would ruin many ambush videos because honest notice changes the interaction. That is precisely why it matters.

Post-recording consent is weaker. A creator who reveals the recording after a conversation may pressure the subject socially. The woman may agree just to exit the situation. She may not understand where the clip will appear, how large the audience is, whether it promotes a business or whether comments will be moderated. Consent must cover use, not merely capture.

Some jurisdictions allow filming in public without consent under many circumstances. But even where capture is lawful, publication and commercial use can trigger separate duties. That distinction should be central to public guidance. A person may be allowed to record a public scene yet still violate privacy, image rights or data-protection rules by posting a targeted, identifiable clip for profit.

Education should focus on creators too. The message should be blunt: if your content depends on strangers not knowing they are content, your model is legally and ethically fragile. Wearable cameras do not exempt you from consent.

Gender-based digital violence is moving into wearable form

The Brussels allegations belong within a broader discussion of gender-based digital violence. Non-consensual filming, sexualised comments, doxxing, deepfake abuse, image-based harassment and platform pile-ons are connected by the same pattern: women’s images and reactions are taken out of their control and used for attention, humiliation or power.

Smart glasses add a capture tool to that chain. The abuse does not need to start with explicit images. A normal street conversation can become sexualised through captions, comments or the creator’s framing. A rejection can become content for misogynistic audiences. A face can become a target for identification. A location clue can reveal routines. The harm begins with recording but expands through interpretation and distribution.

Belgian and EU policy debates on online violence against women should include wearable recording. If law focuses only on deepfakes or intimate images, it will miss the earlier stages of exploitation. Many harmful clips are not nude, not explicit and not fabricated. They are ordinary images weaponised by context.

This does not mean every unwanted recording is gender-based violence. Men can be filmed without consent too. Public officials, workers, children, migrants and LGBTQ+ people can all be targeted. But the dating-advice and pickup-content pattern has a clear gendered structure. Women are approached as subjects, evaluated as proof, and exposed to audiences that may sexualise or insult them.

Policy should therefore connect privacy law with equality law and harassment frameworks. A data-protection complaint may remove a video. A harassment framework may address repeated targeting. A platform policy may reduce amplification. A gender-equality body may identify patterns that pure privacy analysis misses. These tools should reinforce each other.

The Brussels cases also show why “just don’t engage” is a poor answer. Women cannot avoid every street approach. They should not have to treat ordinary conversation as a recording risk. Public life belongs to them too.

Children and young people raise even sharper concerns

Smart-glasses recording of minors raises higher stakes. Teenagers and children may not recognise recording indicators, understand platform distribution or assert boundaries with adults. They may be filmed in schools, parks, shopping streets, sports venues or public transport. Even when the Brussels cases concern adult women, the same device risks apply to younger people.

GDPR treats children’s data with added care in many contexts, and the DSA highlights protection of minors as part of platform risk management. A smart-glasses clip of a minor can expose location, school uniform, social circle, voice and vulnerability. If the clip becomes part of prank or dating-style content, the harm can be severe.

Schools and youth venues will likely face pressure to ban or restrict camera glasses. Some bans may be justified in classrooms, changing areas, toilets, medical rooms and safeguarding contexts. Meta’s own privacy guidance tells users to turn off glasses in sensitive spaces such as doctors’ surgeries, changing rooms, public toilets, schools and places of worship. That is a strong admission that context matters.

But bans are a blunt tool. A child with visual impairment might benefit from smart glasses. A school media project might use wearable cameras responsibly. Policies need room for supervised, consent-based use while prohibiting casual recording of students. The default should be no recording in child-centered spaces unless there is clear permission and purpose.

Platforms should also treat clips of minors filmed through smart glasses with extra caution. A minor may not file a complaint. Parents may not see the video. School staff may be unaware. Automated and human review systems should prioritise targeted clips of identifiable minors in public-interaction content, especially when comments are sexualised or mocking.

For regulators, minors provide the strongest case for proactive standards. Waiting for individual complaints is too slow. By the time a parent discovers a viral video, the damage may be done.

The workplace and service-sector angle is underplayed

Women approached on streets are one group at risk. Workers in public-facing jobs are another. Retail staff, baristas, receptionists, transport workers, hospitality employees and security guards can be filmed by smart glasses while doing their jobs. They may have less freedom to walk away. If a customer is recording with glasses, the worker may not know or may fear escalation if they object.

Prank content already targets workers because they are accessible and constrained. Smart glasses make it easier to capture reactions without visible setup. A worker’s awkwardness, frustration or politeness becomes material. If the clip goes viral, the worker may face embarrassment, employer scrutiny or customer harassment.

Employers have duties to protect staff, but they cannot control every customer’s eyewear. Venues may need policies: no recording staff without consent, smart glasses off in certain areas, visible signage, refusal of service for covert recording. Such policies will be easier to enforce if the law and industry standards make recording indicators clear.

There is also a data-protection issue for businesses. If a customer records inside a shop, they may capture payment terminals, staff names, customer faces, security layouts or confidential conversations. Meta’s privacy guidance warns users not to capture sensitive information such as PIN codes. That warning reflects a real risk. Wearable cameras can record more than the intended subject.

Workers should not carry the burden alone. Platforms should not reward content that harasses employees. Device makers should make covert capture harder. Employers should train staff on how to respond. Regulators should clarify when businesses can require smart glasses to be turned off.

The service-sector angle also broadens public support. This is not only about dating videos. It is about the right to work, shop, commute and socialise without being secretly turned into content.

The role of Meta is bigger than user responsibility

Meta’s standard position is that users are responsible for obeying laws and using the glasses respectfully. WIRED quoted a Meta spokesperson saying the company’s terms require users to comply with applicable laws and not use Ray-Ban Meta glasses for harmful activities such as harassment, privacy infringement or capturing sensitive information; Meta also pointed to the LED light as a recording signal.

That position is partly fair. A company cannot control every user. Phones are used for unlawful filming too. Cars are used dangerously. Knives can be misused. Personal responsibility matters. But product makers still have duties when misuse is foreseeable, scalable and connected to design features.

Meta chose a discreet form factor. Meta chose social media integration. Meta chose voice and button capture. Meta chose the LED design. Meta chose app flows and privacy defaults. Meta operates platforms where some of the resulting content may be posted. The company is not a passive observer of a misuse pattern its ecosystem helps enable.

EssilorLuxottica also has a role. It is not merely licensing a logo. It manufactures and distributes the eyewear category and reports AI-glasses sales as a business growth driver. Eyewear design choices affect detectability, comfort, social acceptance and repair pathways. The privacy problem sits inside the hardware partnership.

A serious corporate response would go beyond terms of service. It would include public reporting on abuse cases, LED tamper research, stronger repair controls, platform enforcement metrics, victim complaint channels, independent audits and consultation with women’s safety groups, disability advocates and data-protection experts. It would also draw red lines around facial recognition and always-on recording.

Meta has an opportunity to avoid the fate of Google Glass backlash. But trust will not come from saying “users must follow the law.” Trust will come from design choices that protect people who never agreed to be part of the product’s world.

The platform economy rewards exactly the wrong behavior

Why do creators film strangers with smart glasses? Because the format works. First-person footage feels intimate and authentic. Viewers like awkwardness, rejection, flirtation and confrontation. Algorithms often reward watch time and comments, not dignity. Dating advice businesses can turn viral clips into leads. The result is a market for social discomfort.

This economy changes how public encounters are staged. A creator is no longer only speaking to the person in front of him. He is performing for the audience he expects later. The woman may think she is having a conversation with one stranger. In reality, she is interacting with a future edit, caption, comment section and sales funnel. The audience is hidden at the moment when consent would matter most.

Smart glasses are ideal for that hidden-audience model because they preserve the illusion of one-to-one interaction. The wearer can behave as if the moment is private while recording for public distribution. The device’s strongest creator feature becomes its weakest ethical point.

Platforms often treat this as user-generated content until a complaint arises. But if the format becomes common, reactive moderation is too slow. Recommendation systems can identify patterns: repeated street approaches, recurring non-consenting subjects, sexualised captions, dating-course links, frequent privacy complaints. Platforms should use those signals to reduce distribution before harm scales.

The business incentives are not neutral. If a video of a woman rejecting a stranger receives millions of views, more creators copy it. If platforms recommend copies, the format spreads. If courses sell, the creator professionalises. If LED tampering improves footage stealth, tampering services grow. This is a feedback loop.

Breaking the loop requires changing rewards. Demonetise non-consensual targeted street recordings. Reduce recommendation of repeat ambush accounts. Require consent evidence for commercial use of identifiable strangers. Penalise accounts that hide recording or ignore removal requests. These steps would not end all misuse, but they would change the economics.

Europe should distinguish recording, recognition and recall

Smart-glasses policy needs three separate categories: recording, recognition and recall. Recording is capture of photo, video or audio. Recognition is identifying or classifying people, objects or scenes through AI. Recall is storing and retrieving past encounters, potentially through memory features. Each layer raises stronger privacy issues than the last.

Recording can be legitimate with notice, consent where needed and limits on publication. Recognition needs stricter rules because it can turn anonymous public life into identified public life. Recall may be the most sensitive because it could let users search past encounters: “the woman I met outside the station,” “the person wearing a badge at the conference,” “the activist I saw at the protest.” Even if such features sound speculative, AI development is moving in that direction.

Research on wearable AI agents already points toward continuous perception and delegated action through smart glasses. Meta and other companies are racing to make glasses more context-aware. Context awareness is useful for assistance, but dangerous when it builds records about bystanders.

Policy should therefore avoid treating smart glasses as simple cameras. A camera captures a scene. An AI wearable can interpret the scene. A memory system can store it for later search. The privacy risk is cumulative. The law should not wait until all layers are active before drawing boundaries.

A workable framework might allow basic recording subject to notice and existing law, tightly restrict unknown-person recognition, and prohibit or heavily regulate searchable memory of non-consenting bystanders. It should also require clear user interfaces showing when each layer is active. A person nearby should not need to guess whether the glasses are merely taking a photo or running identity analysis.

The EU AI Act can cover some recognition systems. GDPR covers personal data. The DSA covers platform distribution. Consumer safety and product rules can cover hardware indicators. The policy challenge is coordination. Without coordination, each regulator sees only one layer while the product works across all of them.

The public needs clearer social norms, not only legal rules

Law moves slowly. Social norms move every day. People need practical expectations for smart glasses now. A simple norm should be promoted: do not record identifiable strangers in close-range social interactions without telling them first, and do not post them without permission unless there is a clear public-interest reason.

That norm is stricter than what some public-filming laws require. It is also reasonable. Face-to-face conversation carries an expectation of presence, not broadcast. A person approached on the street should not have to assume they are being recorded for a stranger’s audience. The norm should protect the interaction before it becomes evidence.

For wearers, responsible use means making recording obvious, asking before filming targeted interactions, stopping when someone objects, avoiding sensitive spaces, not capturing children without permission, not hiding the LED, not using footage for commercial purposes without consent, and blurring people who are not central to a legitimate purpose.

For bystanders, practical advice is harder because the burden should not be on them. Still, public education can help people recognise smart-glasses cues: thick frames, small lens placement, capture LEDs, tapping gestures and voice commands. People should feel entitled to ask, “Are you recording me?” and “Do not post that.” Venues should back them up.

For creators, the norm should be explicit: consent is part of production. If asking for consent ruins the video, the video likely depends on deception. Creators who want real street interviews can build consent into the format. Many journalists and documentary makers already do.

For platforms, the norm should be that non-consensual targeted footage is not ordinary entertainment. It carries heightened review duties, especially when gendered harassment, minors, workers or commercial promotion are involved.

Brussels should push for an EU wearable-camera code

Minister Matz’s call for a European discussion could become a practical policy project: an EU wearable-camera code for consumer devices. The code could be voluntary at first, then tied to product safety, consumer protection or harmonised standards if needed. It should cover smart glasses, camera earbuds, pendants, pins and other discreet wearable devices.

The code should set minimum recording-signal standards. Indicators should be visible from typical interaction distance, tested in daylight and low light, and difficult to disable. The code should require hardware integrity checks and prohibit capture when indicators fail. It should require clear physical markings that identify camera-equipped eyewear.

The code should set user-notice standards. Packaging, setup flows and retail materials should explain legal duties in plain language. Users should be warned that recording and posting identifiable strangers, especially for commercial use, may require consent or another lawful basis. Apps should provide easy blur and deletion tools.

The code should set bystander-remedy standards. Device makers and platforms should provide reporting paths for people filmed by smart glasses. Users should be able to find out how to request removal without knowing the exact device model. Platforms should accept privacy complaints from non-users and act quickly on targeted non-consensual footage.

The code should set high-risk feature rules. Unknown-person facial recognition, always-on recording, persistent memory and biometric inference should be off by default, subject to strict legal review and, where necessary, prohibited for consumer use. Accessibility exceptions should be narrow, consent-based and audited.

Such a code would not solve every problem. But it would give regulators, companies, users and victims a shared baseline. The absence of a baseline lets each company define “responsible use” in its own marketing language.

The product choices regulators may examine

The safest future is not one where smart glasses disappear. It is one where bystanders can tell what is happening and have rights that do not depend on buying the device.

The Brussels controversy is also a consumer trust problem

Consumers who buy smart glasses may not want to be associated with harassment. If the product becomes known as a tool for creeps, ordinary users lose social permission to wear it. WIRED reported that pickup artists and pranksters have already helped give Meta Ray-Bans a hostile nickname in some circles. That reputational risk is not abstract. Google Glass became socially toxic long before the technology matured.

Meta and EssilorLuxottica should treat bystander privacy as brand protection. Ray-Ban’s value depends on social confidence. People wear the brand because it signals style, not suspicion. If Ray-Ban frames become associated with covert filming, the brand pays a cost. The same applies to retailers and opticians selling smart eyewear to mainstream customers.

Trust can erode quickly. A few viral stories of women secretly filmed, LED lights drilled out, or face recognition hidden in apps can define public perception. The product’s legitimate uses may be drowned out by fear. This is why safety critics are not necessarily enemies of the category. Many are warning that without stronger safeguards, public backlash could become broad and indiscriminate.

Consumer trust also depends on clarity about data flows. Users need to know when their own photos, videos and voice recordings go to Meta. Bystanders need assurance that footage of them is not automatically used for AI training or reviewed by contractors. The Verge and TechCrunch have reported on Meta privacy-policy changes affecting AI glasses and data use, including voice recordings and AI training concerns. EFF has urged users to review settings and turn off cloud media where possible.

If users themselves struggle to understand the settings, bystanders have no chance. A trustable product must make the most privacy-sensitive states visible and understandable. Hidden complexity is not compatible with a device that records public life.

Data flows matter after capture

The moment of filming is only the first privacy event. The next questions are where the data goes, how long it stays there, who can access it, whether it trains AI models, whether humans review it, whether it is linked to accounts and whether deletion removes all copies. Smart glasses create data that feels casual but may be rich: faces, voices, locations, routines, social ties, objects, screens, documents and background conversations.

Meta’s help materials say captured media can be reviewed and shared through the Meta AI app, and EFF notes that users can disable some sharing options such as “additional data” and cloud media. These controls are user-facing. They do not directly answer what happens to a bystander’s data when a wearer chooses cloud features or uploads a clip.

The bystander issue becomes sharper with AI processing. If a user asks Meta AI about what they see, the system may process images containing strangers. If cloud processing is involved, data may leave the local device. If training or quality review is involved, people who never interacted with Meta may be present in the material. The public needs clear separation between user-controlled capture, cloud AI processing and platform publication.

Data minimisation should apply at every stage. Do AI features need to send full-resolution images of bystanders to the cloud? Can faces be blurred before upload when identity is irrelevant? Can processing happen on-device? Can retention be shorter by default? Can sensitive-location recording trigger stronger warnings? These are design questions, not only policy documents.

The most privacy-preserving smart-glasses systems will process as much as possible locally, avoid storing bystander-identifiable data where not needed, provide automatic redaction and make sharing deliberate. The least privacy-preserving systems will treat the world as raw input for AI improvement. Europe’s legal framework should push the industry toward the first model.

Enforcement will require evidence, not outrage

Public anger can start a policy debate, but enforcement needs evidence. Investigators will need to know whether specific women were identifiable, whether they were told, whether the footage was uploaded, whether it promoted a business, whether deletion was requested, whether the account repeated the practice, whether the LED was functioning and whether platform systems amplified the content.

Victims and bystanders should preserve links, screenshots, account names, timestamps, comments, business pages and any messages with the creator. They should avoid engaging in ways that expose more personal data. If the video includes threats, stalking, sexualisation, minors or private information, urgency rises. For GDPR complaints, the difference between private storage and public commercial posting matters.

Regulators should build cases that distinguish legitimate public recording from exploitative wearable-camera content. A weak case can create bad precedent. A strong case would involve repeated conduct, clear identifiability, poor notice, commercial use, refusal to remove and platform reach. The most useful first enforcement actions will target business models, not accidental background capture.

Evidence from device design also matters. Independent experts can test LED visibility. Researchers can document tampering markets. Consumer groups can survey public understanding. Women’s organisations can collect victim reports. Platform researchers can track recommendation patterns. Together, this evidence can move the debate from anecdote to policy.

The EDPB’s video-device guidance already warns against normalising a lack of privacy through expanding video use. Smart glasses are a new form of the same pressure. Evidence will help regulators adapt existing principles without overreaching.

Enforcement should also include soft tools: guidance, warnings, retailer notices, platform policy changes and industry commitments. Not every solution must be a fine. But fines and orders must remain available when companies or creators ignore clear risks.

A ban would be tempting but too blunt

Some people will call for banning smart glasses in public. That response is understandable when women are secretly filmed. But a broad public ban would be hard to enforce, easy to overreach and potentially harmful to legitimate users. It could also push covert devices into less accountable forms. The better response is targeted restriction.

Specific bans make more sense in sensitive places: toilets, changing rooms, schools, medical settings, courts, secure workplaces, shelters, gyms and certain nightlife venues. Meta’s own privacy guidance tells users to turn off glasses in several sensitive spaces. Venue-level bans can be clear and enforceable when privacy expectations are high.

For streets and public squares, targeted conduct rules are better. Ban or penalise disabling recording indicators. Restrict commercial publication of targeted identifiable footage without consent. Require platforms to act on privacy complaints. Prohibit unknown-person facial recognition through consumer glasses. Mandate minimum indicator standards. These measures address the harmful uses without outlawing all wearable cameras.

A ban-first approach could also distract from platform incentives. If smart glasses disappeared tomorrow, phones would still be used for non-consensual content. The policy goal should be broader: stop turning strangers into monetised humiliation or sexualised material without permission. Smart glasses are the current pressure point, not the only tool.

The EU’s best regulatory style is often risk-based. Apply stronger rules where the risk is higher: covert recording, minors, sexualised framing, commercial use, biometric identification, sensitive spaces and repeated targeting. Keep lower-risk uses possible with safeguards. That approach fits GDPR, the DSA and the AI Act better than a symbolic ban.

The next version of smart glasses will be harder to govern

Today’s Ray-Ban Meta glasses already raise privacy concerns. Future models will raise harder ones. Displays, better batteries, live AI assistance, improved microphones, wider fields of view, on-device models and memory functions will make glasses more useful and more intrusive. Meta has already expanded the AI-glasses portfolio with EssilorLuxottica, and the broader market is moving quickly.

A display changes the social dynamic. The wearer can receive information while maintaining eye contact. If that information includes AI-generated prompts, identification cues or conversation advice, the other person may be interacting not only with a human but with a hidden assistant. Dating advice businesses will notice. Salespeople will notice. Political canvassers will notice. Fraudsters will notice.

Better AI also changes what counts as recording. A system may not store a full video but may extract facts: gender presentation, age estimate, emotion cues, clothing, location, text on badges, names spoken, relationship hints. Some of those inferences may be personal data. Some may be sensitive. The privacy debate cannot focus only on saved video files.

Wearable AI memory could create a private archive of public encounters. Even if the archive is user-only, it affects others. A person who chatted briefly with a stranger may not expect to be searchable in that stranger’s AI memory months later. If such memory syncs to cloud systems, the risk grows.

The regulatory window is now. Once advanced features become normal, removing them becomes politically and commercially harder. Europe should set boundaries before the next generation arrives, not after.

A practical European response could start with five moves

Europe does not need to solve every wearable-computing question at once. It can start with practical steps tied to the Brussels cases.

First, regulators should issue guidance on smart-glasses recording under GDPR. The guidance should distinguish personal use, public-interest recording, commercial social content and targeted recording of identifiable strangers. It should explain when the household exemption is unlikely to apply.

Second, consumer authorities should examine marketing claims about privacy safeguards. If companies present the LED as a privacy-protecting feature, authorities can ask whether the feature is visible, understood and tamper-resistant enough for that claim.

Third, the European Commission and national digital regulators should ask very large platforms how they handle non-consensual wearable-camera content under the DSA. Reporting pathways, non-user complaints, monetisation and recommendation should be part of the review.

Fourth, EU AI Act authorities should prepare guidance on consumer smart glasses and biometric identification before facial recognition features are launched. Waiting for deployment would be poor governance.

Fifth, governments should fund support and reporting tools for victims. Legal rights mean little if the person filmed cannot get a video removed or preserve evidence.

These moves are modest. None bans the product. None prevents legitimate recording. Each targets the gap revealed in Brussels: people being recorded and used as content without meaningful knowledge or control. Good regulation should make abusive uses harder while leaving responsible uses intact.

The burden should move away from women in the street

The most troubling part of the Brussels story is how familiar it feels. Women are approached in public, placed in an unwanted social situation, recorded, judged and made responsible for managing the aftermath. Technology changes the tool but not the burden. Smart glasses risk making that burden heavier.

A fair response moves responsibility back to the people and companies with power. Creators should ask before recording and posting. Device makers should design safeguards that do not depend on bystander vigilance. Platforms should stop rewarding non-consensual street content. Regulators should clarify rights and enforce against repeat offenders. Police and support organisations should take complaints seriously.

The public should resist the idea that being outdoors means surrendering all control over one’s image. Public life requires visibility, but it also requires trust. A city where every interaction might be secretly recorded for content is not more open. It is more guarded.

Ray-Ban Meta glasses may become useful everyday tools. They may also become a symbol of a privacy failure if companies treat bystanders as background data. The Brussels cases show which path is forming. Europe still has time to steer it.

The central question is social, not technical

The technical question is whether smart glasses can record discreetly. They can. The legal question is whether existing laws can respond. Often they can, though with gaps. The market question is whether millions of people will buy them. They already have. The deeper question is whether society will accept face-worn cameras as normal in close human interaction.

Acceptance will depend on trust. People may accept smart glasses when used openly, in appropriate spaces, with clear signals and respect for objections. They will resist them when used to ambush strangers, film women for dating content, hide recording indicators or prepare facial recognition. The future of smart glasses will be decided less by camera quality than by whether non-users feel protected.

Meta, EssilorLuxottica and their rivals should treat Brussels as a warning, not a public-relations inconvenience. The same device can support accessibility, creativity and communication, or it can erode the privacy of public life. The difference lies in design, defaults, platform rules and enforcement.

Europe’s task is to protect the ordinary person who did not buy the device, download the app or agree to appear in someone’s content. That person is the missing stakeholder in much of wearable AI. Brussels has put her back at the center of the debate.

Questions readers are asking about smart glasses, privacy and the Brussels cases

Author:
Jan Bielik
CEO & Founder of Webiano Digital & Marketing Agency

This article is an original analysis supported by the sources cited below

Women in Brussels filmed without their knowledge by men wearing Meta smart glasses
The Brussels Times report on women allegedly filmed in Brussels with Meta smart glasses and Belgian Minister Vanessa Matz’s call for a European discussion.

Ray-Ban Meta AI glasses
Meta’s product page for Ray-Ban Meta AI glasses, used for product context and feature positioning.

Capture photos and videos with AI glasses
Meta help page explaining capture methods and the role of the capture LED.

Privacy settings for Ray-Ban Meta AI glasses
Meta privacy guidance for Ray-Ban Meta glasses, including advice on sensitive spaces and the capture LED.

Ray-Ban Meta FAQ
Ray-Ban’s FAQ explaining device identification, capture LED function and user controls.

EssilorLuxottica Q4 and full year 2025 results
Official EssilorLuxottica results reporting more than 7 million AI-glasses units sold in 2025.

Regulation EU 2016/679 General Data Protection Regulation
Official GDPR text used for data-protection principles, lawful processing and biometric-data context.

EDPB guidelines on processing personal data through video devices
European Data Protection Board guidance on video devices, household exemption and GDPR application.

Digital Services Act
European Commission page explaining the DSA’s role in platform accountability and user protections.

AI Act article 5 prohibited AI practices
EU AI Act Service Desk text on prohibited AI practices and restrictions around real-time remote biometric identification.

AI Act biometrics guidance
EU AI Act Service Desk material on biometric use cases and high-risk AI systems.

The rise of the Ray-Ban Meta creep
WIRED reporting on pickup artists, prank creators and privacy concerns around Ray-Ban Meta glasses.

Meta silently added face-recognition code for its smart glasses to millions of phones
WIRED investigation into inactive face-recognition code in the Meta AI companion app for smart glasses.

Think twice before buying or using Meta’s Ray-Bans
Electronic Frontier Foundation analysis of privacy settings, cloud media and bystander risks.

Neue Form digitaler Gewalt? Heimlich gefilmt mit Smart Glasses
German public-broadcast reporting on women secretly filmed with smart glasses and the “rizz” content pattern.

The disturbing rise of nightlife content
The Guardian’s reporting on non-consensual public filming, sexualised comments and legal gaps around nightlife footage.

A $60 mod to Meta’s Ray-Bans disables its privacy-protecting recording light
404 Media investigation into modifications that disable the recording indicator on Meta Ray-Bans.

People are paying to get rid of the recording light on their Meta Ray-Bans
The Verge report on services removing the recording indicator LED from Meta smart glasses.

Meta tightens privacy policy around Ray-Ban glasses to boost AI training
The Verge coverage of Meta privacy-policy changes affecting Ray-Ban Meta glasses and AI data practices.

If you own Ray-Ban Meta glasses, double-check your privacy settings
TechCrunch reporting on privacy-setting changes and data-use concerns for Ray-Ban Meta users.

Markey, Wyden, Merkley demand transparency from Meta on facial recognition technology in smart glasses
Official U.S. Senate press release demanding answers from Meta over reported facial-recognition plans for smart glasses.

ACLU and 75 organizations sound alarm on Meta’s plans to add facial recognition technology
ACLU-led coalition statement warning against facial recognition in Ray-Ban and Oakley smart glasses.

Belgian Data Protection Authority citizen page
Belgian DPA public page describing its mission to protect personal data and enforce data-protection principles.

Belgian DPA fine for unlawful processing of video images
EDPB summary of a Belgian DPA decision involving unlawful video-image processing and data-protection-by-design concerns.

VisionClaw always-on AI agents through smart glasses
Research paper on always-on wearable AI agents using smart glasses and continuous egocentric perception.

WhatsAI transforming Meta Ray-Bans into an extensible generative AI platform for accessibility
Research paper on accessibility uses of Meta Ray-Bans for blind and visually impaired users.

Imaging for all-day wearable smart glasses
Research paper on imaging constraints and design issues for all-day wearable smart glasses.