My View - ‘pervert glasses’ or assistive tech?

MY VIEW - My View – Sally Britnell, Fellow of HiNZ

Smart glasses have acquired an image problem. 

Critics have labelled them ‘pervert glasses’ following reports of people using apparently ordinary-looking eyewear to film women without their knowledge and publish the encounters online. The term has moved from social media into mainstream reporting, while courts, conferences and other venues have begun restricting camera-equipped eyewear. 

The concerns behind that reaction are real. 

New Zealand Privacy Commissioner Michael Webster has warned that smart glasses can make covert filming easier because they look like ordinary eyewear. He has also referred to reports of men using them for non-consensual filming and livestreaming of interactions with women.  

Overseas, the response is already moving beyond social disapproval. Courts in England and Wales and New York have prohibited smart glasses in court facilities. DEF CON banned recording-capable Meta-style glasses at its 2026 event. Norway has proposed broader restrictions in some public settings, while Australia has considered restrictions in government workplaces. 

These restrictions are not all equivalent. The UK Cinema Association, for example, has acknowledged privacy and piracy concerns alongside the potential benefits of wearable technology for people with accessibility requirements, and called for a proportionate response. 

That brings me to the side of this debate that risks being lost. 

I am Deafblind. 

For me, these glasses can be assistive technology. The same camera that understandably worries another person can help me access information I cannot see, connect me with another person for visual assistance, or give a guide-dog instructor another way to observe what is happening while we train. 

Is the problem the camera, the design of the technology, what companies do with the information, or what people choose to do with it? 

I waited years for this technology to work at home 

I bought my glasses overseas about four years ago. The functionality worked overseas, but not when I brought the glasses home to New Zealand. That was particularly frustrating because I already knew what the technology could potentially offer me. Geography determined whether I could access it. 

So I spent years waiting for New Zealand to catch up. 

Access is not simply about whether technology exists. It is also about where you live and whether the functionality you need is available there. 

For a mainstream consumer, delayed access to a feature might be inconvenient. For a disabled person who can use the same functionality as assistive technology, it can mean something quite different. 

When consumer technology becomes assistive technology 

Smart glasses were not created solely as disability devices. Current Meta AI glasses incorporate cameras, microphones, open-ear audio and AI functionality. Meta documents uses including questions about surroundings and visible text. When those questions involve vision, Meta says the glasses can capture a photograph and send it to Meta’s cloud for AI processing. Read Meta’s explanation 

For someone who is blind, Deafblind or has low vision, those functions can acquire an entirely different significance. A product does not have to be invented specifically for disabled people to become assistive technology. Our phones already demonstrate this. Smart glasses may represent another stage of that evolution. 

In ordinary life, the point is practical access to written information. I might ask the glasses to read a café menu, identify a street sign at an unfamiliar intersection, check a bus timetable to work out when the next service is due, or tell me the price on a supermarket shelf. In each case, I am not trying to record the people around me. I am trying to obtain information that sighted people can usually read at a glance. 

Sometimes the technology connects people 

The same access purpose can also involve another person. One of the most useful things I have already done with my glasses in New Zealand has nothing to do with replacing human assistance. 

I have used Be My Eyes through them. I was in Auckland. The volunteer who answered my call was in Rotorua. 

Someone in another part of Aotearoa could look through the camera in my glasses and provide visual assistance about what was immediately around me. Be My Eyes says its integration allows a volunteer to see through compatible Meta AI glasses and that users can switch between the glasses and mobile phone camera during a call.  

The published Be My Eyes country list I checked did not include New Zealand, despite my successfully using the integration here with a volunteer in Rotorua.  

The technology was not replacing another person. It was connecting two people. 

For me, independence is not about demonstrating that I never need anybody else. It is having greater choice and control over when, how and from whom I obtain information. 

I have also used them in guide-dog training 

I have used the glasses’ live video capability to allow a guide-dog instructor to walk behind me and see on their mobile phone what I was seeing through the glasses. 

I am not suggesting this is an established or validated method of guide-dog training. It is something I have used in my own training. In that context, the camera gave the instructor another view of what was happening while we worked together. 

Same camera. Very different purpose. 

The label responds to a real harm, but can create another 

The term ‘pervert glasses’ did not emerge without cause. Reporting has documented people using smart glasses to record women without their knowledge, including interactions later distributed as social-media content. An RNZ analysis places this within wider patterns of gendered harm involving covert filming and subsequent sexualisation or distribution of footage.  

That behaviour deserves condemnation. But when the label moves from condemning what somebody has done to describing the object they are wearing, every wearer can begin to look suspect. 

For a consumer, that stigma might make an expensive gadget embarrassing to wear. For a Deafblind person, it could discourage the use of something functioning as assistive technology. 

This does not make disability exclusion equivalent to the harms experienced by women who are covertly filmed. Nor should accessibility be used to minimise gendered harassment. But we should be capable of holding two truths simultaneously: 

People deserve protection from covert recording and online exploitation. 

Disabled people should not automatically become objects of suspicion for using camera-based assistive technology. 

So what makes glasses different from holding up a phone? 

There is an important difference. Someone holding a phone towards another person often provides an obvious social cue that photography or filming might be occurring. Smart glasses can look like conventional eyewear. That is precisely why New Zealand’s Privacy Commissioner has warned they can make covert filming easier. 

But less visible does not automatically mean malicious. 

Consider two people wearing identical glasses. One deliberately records a private encounter without the other person’s knowledge. The other is Deafblind and asks the glasses to read a street sign, menu, supermarket price label or bus timetable. The hardware might be identical. The purpose, intention, information sought and ethical context are not. 

What the glasses do not tell me is interesting too 

In my experience, the glasses seem reluctant to give detailed descriptions of people. They do not appear to remember people around me and then identify them when I encounter them again. That is my observation as a user. I cannot say this was specifically designed as a privacy safeguard unless Meta documents it. 

I initially thought of a visual interaction as a temporary screenshot. Meta’s documentation describes a more complex process. A photograph can be sent to its cloud for AI processing, and the request, image and response can be reviewed later in the mobile app.  

If I ask what a menu, street sign, supermarket price label or bus timetable says, I do not want the photograph. I want the information written there. The image is a means of obtaining accessible information, but that does not make the privacy implications of capturing and processing it disappear. 

I have confronted the same tension as a researcher 

My PhD involved designing computer vision within medical technology. 

Medical technology and consumer smart glasses are not the same. But both show why the ethical analysis of computer vision cannot stop at the discovery that there is a camera. Computer vision needs visual information in order to interpret something about the world. 

The point is not that good intentions erase privacy risk. It is that privacy risk has to be assessed alongside purpose, design and safeguards. 

We need to ask what information a system requires, what is captured, where processing occurs, what is retained, who might have access, whether collection is proportionate, and whether the same benefit could be achieved while collecting less. 

That tension is now personal as well as professional: I understand the need to scrutinise computer vision, and I also know what it can make possible when it is used for access. 

Responsibility exists at several levels 

The wearer is responsible for how the technology is used. People should be accountable for deliberately recording or exploiting others. 

Responsibility also exists at the level of design. Meta’s glasses include an outward-facing capture indicator for conventional photography and video. In 2026, Meta strengthened anti-tampering measures so camera functionality can be disabled when interference with the indicator is detected.  

There is also the less visible question of what happens to information within company systems. Meta has described a ‘Private Processing’ architecture using confidential computing for some AI workloads.  

Taken together, these issues show why privacy cannot be reduced to whether somebody notices a recording light. 

Blanket bans could create another problem 

Some restrictions will be justified. A courtroom is not a supermarket. A changing room is not a footpath. A confidential clinical consultation is not a railway station. 

But imagine I enter a supermarket wearing my glasses to read shelf prices or product information. Someone notices the camera and complains. The supermarket tells me to remove the glasses. It may believe it has resolved a privacy problem, but it may also have removed my means of accessing the information needed to shop independently. 

A policy saying ‘no unauthorised recording in this area’ targets conduct. A policy saying ‘no smart glasses’ targets a device. 

There will be environments where excluding recording-capable equipment is justified. Elsewhere, organisations should ask whether the wider restriction is necessary and what its accessibility consequences will be. 

New Zealand has another gap to confront: accessibility law 

New Zealand has disability-rights protections and the concept of reasonable accommodation. Guidance describes reasonable accommodation as necessary and appropriate adjustments that do not impose a disproportionate or undue burden and enable disabled people to exercise rights equally.  

That is different from having enforceable accessibility standards across society. 

The Accessibility for New Zealanders Bill was introduced in 2022. It proposed an Accessibility Committee and sought to strengthen government accountability, but it was discharged before its second reading and did not become law.  

The proposal was criticised for not going far enough. Parliament’s first-reading summary records objections that it focused on advice and data gathering rather than meaningful standards and enforcement mechanisms.  

The New Zealand Disability Strategy 2026–2030 includes accessibility as a common theme and guides government agencies. But a strategy is not legislation establishing enforceable accessibility standards.  

Could we regulate risk faster than we protect access? 

Overseas, governments, courts and organisations are already deciding where camera-equipped glasses should be restricted. Meanwhile, New Zealand does not have comprehensive accessibility legislation establishing enforceable cross-sector standards. 

We could become increasingly sophisticated about deciding where disabled people cannot use emerging technology before becoming equally sophisticated about protecting their access when they cannot use it. 

If a venue prohibits camera-enabled glasses, could the privacy risk be managed without removing the device? If not, what accessible alternative will be available? Will it provide equivalent information when it is needed? Who carries the burden of negotiating this? 

Reasonable accommodation remains important and is considered case by case. But responding after a disabled person encounters a barrier is different from designing policies that anticipate accessibility from the outset. 

Assistive technology may no longer look ‘assistive’ 

A white cane is readily understood as assistive technology. So is my guide dog. The mobile phone in my hand may be less obvious, yet it can perform important access functions. The same is now true of the glasses on my face. 

Increasingly, what makes something assistive technology may be less about what the device looks like and more about what the disabled person is using it to do. 

For me, disability is not only about what I can or cannot see or hear. It is also about what happens when the world is not set up for me to get the same information as everyone else. A menu I cannot read, a street sign I cannot identify, or a timetable I cannot check becomes less disabling when technology helps me access it. The same technology becomes disabling again if other people respond to it only with suspicion. 

Blanket product bans struggle with that distinction. 

What could stronger accessibility legislation change? 

Comprehensive accessibility legislation would not need to give me an unrestricted right to use a camera wherever I wanted. That would simply replace one person’s rights with another’s. 

Instead, stronger rules could shift responsibility upstream. Organisations could be expected to consider disabled people’s access when policies are designed, rather than waiting until someone is excluded and then treating it as an individual exception. 

For smart glasses, this could mean asking whether a restriction is necessary and proportionate, whether harmful conduct can be regulated without prohibiting an entire product category, and what reasonable alternative will be available if the technology genuinely cannot be used. 

These are policy principles I am proposing, not requirements contained in current New Zealand law. But they would add a vital question to the privacy debate: 

If we restrict this technology, whose access are we removing and what will replace it? 

We need something more sophisticated than ‘ban them’ or ‘trust them’ 

I am not arguing that society should simply trust smart glasses. Disability access does not provide a blanket justification for recording other people. Both positions are too simple. 

There will be environments where camera-enabled technology must be restricted, and others where restrictions should accommodate legitimate disability access. Manufacturers should reduce unnecessary collection and strengthen privacy protections. Users should be responsible for their behaviour. Organisations should regulate proportionately. 

My glasses have connected me in Auckland with a volunteer in Rotorua. They have supported how I work with a guide-dog instructor. They can help me read menus, street signs, timetables and prices that I cannot otherwise access visually. None of this excuses another person using the same technology to covertly record or harass somebody. But their misconduct should not determine my access. 

Privacy is a right. 

Access to information is also a right. 

That is why the question is not simply whether smart glasses are risky. It is whether we can respond to risk without treating access itself as suspicious. 

The challenge for Aotearoa is not deciding which right should give way first. It is developing technology, law and institutional practice sophisticated enough to protect both. 

Author’s note: AI was used to assist with editing and to make the article’s language and structure consistent with the conversation. It was not used to formulate the article’s arguments or examples. 

Read more VIEWS

Return to eHealthNews.nz home page