The launch of the first Meta-branded pair of smart glasses — a lower-priced sequel to the Ray-Ban Meta collaboration that has been on the market since 2024 — arrives at a moment the operator would prefer we forget. The previous two years produced a documented pattern: women recorded in airports, on beaches, at bars, and in the unremarkable business of walking down a public street. The footage, in most cases, was posted to the very platforms that built the device. The subjects, in nearly all cases, did not consent and did not know.
Meta's response, then and now, is the privacy safeguard. There is, the company has repeatedly explained, a small LED indicator that illuminates when the device is recording. This safeguard is the centre of the firm's defence, and we wish to spend a moment on it.
The safeguard, examined
The LED, in the operator's own framing, is the answer to the question of consent. It is, in our reading, not an answer at all. It is a deflection.
An LED is not a request. It is not a question. It does not ask the subject whether they wish to be recorded. It informs them, after the operator of the device has already decided to record, that the recording has begun. The subject's only remedy, on this architecture, is to observe the device, identify the light, interpret its meaning, and object in time to prevent the capture. Most subjects do not know the device is recording-capable. Many of those who do cannot see the light at conversational distance, in daylight, or at certain angles. An investigation by two Swedish newspapers confirmed that the lights are easily covered by a small piece of tape, which is to say the safeguard is a courteous suggestion that a determined user may decline.
The operator's privacy posture rests on a hardware indicator that an end-user can defeat in approximately two seconds, using a material available at any office supply store.
This is not a privacy safeguard. It is a privacy performance — a visible gesture that allows the operator to say, in the next round of press enquiries, that consent architecture was provided. The fact that the architecture is, by design, defeatable does not trouble the script.
What the marketing actually says
The marketing of the new device leans on two pillars. The first is the influencer — in this case Kylie Jenner, whose collaboration with the brand is the explicit pretext for the launch. The second is the descriptor "smart", which has by now done enough work in the industry to mean whatever the press release requires it to mean on the day. In this campaign, "smart" appears to mean: a wearable camera, an open-ear audio system, an AI assistant, and a small light.
What the marketing does not say is that the operator is, according to multiple reports, considering the integration of facial recognition into a future software update. More than seventy civil liberties organisations in the United States have already written to the operator's chief executive to urge that this not happen. The letter's language is restrained, which is the language of organisations accustomed to being ignored; it is worth quoting in full.
"Facial recognition technology built into inconspicuous consumer eyewear represents a serious threat to privacy and civil liberties for every member of our society, and particularly for historically marginalised and vulnerable groups. People should be able to move through their daily lives without fear. Our concerns reflect the fundamental danger of the technology itself." — Open letter to Mark Zuckerberg, April 2026
That a coalition of seventy organisations felt obliged to write this letter in plain English tells you what kind of organisation the operator has become.
The framing of consent
There is a defence of the device, common in the commentariat, that we should name in order to dismiss it. It runs: if you do not wish to be filmed, do not leave the house. This is the argument of a person who has not thought the matter through. The argument assumes that the burden of consent lies with the subject of the recording rather than the operator of the device. It assumes that the public street is, by default, a film set. It assumes that the women filmed at airports and beaches were negligent in failing to anticipate that a stranger might approach them with a hidden camera and post the footage to a network of 1.3 million followers.
This is, plainly, the inverse of consent. Consent is the question asked before the recording begins, of the person being recorded, in terms they understand, with the option of refusal that is meaningful. None of these conditions are met by a small light on a stranger's face.
The legal posture, as filed
The operator's defence, where it engages the law at all, is that recording in a public space is permissible in most Australian states and US jurisdictions. This is correct in a narrow sense. It is not, however, the end of the matter.
New South Wales' Surveillance Devices Act provides for penalties of up to five years' imprisonment for the recording of a private conversation without the consent of all parties. Other states rely on a Human Rights Act that includes an explicit right to privacy. The fact that these laws have, in the recent cases, resulted in platform bans rather than criminal charges is a feature of the enforcement gap, not the legal architecture. The platform ban, in most cases, took the form of a permanent suspension of the user's account — a remedy the operator applied long after the footage had been viewed, shared, and absorbed into the algorithmic recommendation stack.
Operator states that the device is "designed with privacy in mind." Observed practice: the device is designed so that its privacy features may be defeated by the end-user, without consequence, in seconds.
The pattern, in summary
What we have, then, is a familiar pattern. A platform announces a privacy safeguard that is technically present and practically defeatable. An influencer is paid to demonstrate the device in public, in contexts where the subjects of the demonstration are not asked. A press cycle follows in which the operator's communications team repeats the word "privacy" at sufficient volume to make the matter feel settled. A coalition of civil liberties organisations writes a letter. The letter is acknowledged, if at all, with a thank-you and a forward-looking commitment to "ongoing engagement with stakeholders." The product is shipped. The device ships with a small light.
This is, in our reading, the gap between the press release and the postmortem — the gap our Practice was built to measure. The press release is that the device is private. The postmortem, when it comes, will be that it was not.
What the firm recommends
We do not, in our practice, recommend products. We do not consult for the firms we monitor, and we do not endorse the platforms that build the devices. We will note the following, however, for the public record.
- The operator's privacy safeguards, as currently configured, are not a substitute for affirmative consent. They are a courtesy. Courtesy is not consent.
- The platform's own enforcement record — account suspensions issued weeks or months after the offending footage has circulated — is a lagging indicator of harm, not a leading one.
- Regulators in the relevant jurisdictions should not require the operator to retrofit consent onto a device whose design precludes it. They should require the operator to ship a different device.
- The "smart" descriptor, applied to a wearable that films by default and announces itself by light, is doing more work than it is licensed to do.
None of the above requires the operator to apologise. It requires the operator to change the product. We are not, on the public record, optimistic that it will.