2 SEP 2026 — Norway's digital minister Karianne Tung is considering a ban on camera-enabled wearables, naming Meta and Snap glasses, and will convene an expert group. Her own words describe a rule about conduct — preventing the equipment being used to monitor people in public. A product ban and a use restriction are different instruments, and only one of them is straightforward for an EEA state.
What has been proposed
Tung said the aim is to prevent this equipment being used to monitor other people in public places. The scope under consideration covers devices combining artificial intelligence with cameras and microphones built into everyday objects, and may include prohibiting the use of facial recognition on other people in public spaces.
Meta and Snap smart glasses are named as examples. An expert group will be established to advise on stronger privacy protections. No draft legislation has been published, no timetable given, and no response from either manufacturer appears in the reporting.
The phrase circulating in the debate, used by civil liberties advocates rather than by the minister, is pervert glasses.
Banning a product and banning a behaviour are not the same job
Norway is in the European Economic Area, which means it applies EU single-market rules on goods. A member state prohibiting the sale of a product lawfully placed on the EU market has to justify that under a narrow set of grounds and expect it to be tested.
Restricting what people may do with a device carries none of that difficulty. Norway already regulates covert filming and data protection, and adding a rule about recording identifiable people in public is ordinary domestic lawmaking.
Tung's quote is squarely in the second category. The word ban in the coverage is doing more work than the minister's own description. The expert group is how that gap will most likely be resolved, in favour of the achievable option.
Some of this may already be illegal
The facial recognition element in particular has recent law sitting on top of it. The EU AI Act restricts real-time remote biometric identification in publicly accessible spaces, and the GDPR treats biometric data used to identify a person as a special category requiring an explicit legal basis.
An individual pointing recognition software at strangers in the street is not obviously covered by the household exemption that protects ordinary personal photography, because the exemption falls away once processing is systematic and directed at people outside your own affairs.
So a national prohibition may be adding clarity rather than new law. A rule people can understand without reading two European regulations has its own enforcement value, and it is a smaller change than the headline suggests.
Enforcement is where a use rule gets difficult
A conduct rule is easier to pass but harder to police, because the offence is invisible at the moment it occurs.
Nobody can tell from across a street whether a pair of glasses is recording, whether the footage was kept, or whether recognition software ran against it. A complaint therefore arrives after the fact, usually because the footage surfaced somewhere, and by then the evidence sits on the wearer's own device.
Which points at where such rules actually bite. They are enforced against people who publish, because publication is the visible act, and they do very little about the person who records continuously and keeps it. That does not make the law pointless. It means the law will change behaviour at the margins rather than stop the core activity that worries people.
The recording indicator is the part that already failed
Manufacturers answered this problem with a light. Every camera-enabled pair has an indicator that illuminates when recording, on the theory that bystanders can see when they are being filmed.
The theory requires the bystander to be looking at the wearer's face, in adequate light, at the moment recording starts, and to know what the light means. We reported that Meta had to address the capture indicator being obstructed, which is the failure mode you get when a safety feature depends on remaining visible on a device people wear on their face.
That is the strongest argument for regulating conduct rather than the device. A consent mechanism that relies on a stranger noticing a small light is not a consent mechanism, however it is engineered.
Three answers to one problem
Norway is the third jurisdiction-level response to a question that has so far been handled privately. We reported that a pub chain wrote its own rule about filming with smart glasses on its premises, which is the venue answer: a private operator sets conditions of entry and enforces them at the door.
The device answer is the indicator light, which has now been shown to be fragile. The statutory answer, which Norway is now considering, is the only one that binds someone walking down a public street where no operator sets conditions.
None of the three has been tried at scale in this region. Singapore's PDPA, Malaysia's PDPA and Indonesia's PDP Law all carry personal-use or household exemptions written before wearable cameras were common, and none of them addresses a private individual recording strangers continuously. A venue here can post a sign; there is currently nothing else.
View author profile → · Editorial policy
