Meta's camera-free 'Luna' glasses pivot away from privacy backlash

Intimate and explicit videos of users were accessed by subcontractors without user awareness, including personal financial information.
A voice-first device cannot record video of bystanders
Luna's camera-free design fundamentally changes what smart glasses can do to others.
Mark

So Meta is just removing the camera entirely? That seems like a pretty dramatic reversal from Ray-Ban Meta.

Mimi

It does, but the privacy scandal left them little choice. When subcontractors in Kenya were accessing intimate videos without users knowing—including explicit content and financial information—it became clear that even technical safeguards like the LED light weren't enough to contain the damage.

Luke

Wait, let's be precise about what we know. The Swedish investigation found that subcontractors had access to these videos. But do we know if that access was intentional on Meta's part, or was it a failure of their contractor vetting?

Mimi

The reporting suggests it was a failure of oversight. Workers were reviewing videos as part of data labeling to train AI models. Many appeared to have been recorded without user awareness. But you're right—we don't have confirmation of whether Meta knowingly allowed this or simply didn't catch it.

Mark

So Luna is the answer—no camera, no videos to leak, no problem?

Mimi

In theory, yes. Six microphones instead, voice-activated AI. You can't secretly record someone with a microphone the way you can with a camera.

Luke

But that's a different kind of recording. Audio has its own privacy implications. And we don't know yet how users will actually respond to a smart glasses device without a camera. Will anyone buy it?

Mark

That's the real gamble, isn't it? Meta is betting that voice AI is enough to make smart glasses useful.

Mimi

Exactly. They're launching at their developer conference in late September, with sales starting in October. We'll see pretty quickly whether the market agrees.

Luke

And if it doesn't? If Luna flops because people actually want cameras in their glasses?

Mimi

Then Meta has a much bigger problem than privacy—they have a product problem.

  • A scandal broke when Swedish journalists revealed that Meta subcontractors in Kenya had reviewed thousands of intimate and explicit videos — many captured without the wearers' knowledge — as part of AI training work.
  • The exposure shattered confidence in Meta's existing privacy protections, including the LED recording indicator that some users had physically drilled out to secretly film others.
  • Meta is now pivoting with Luna, a smart glasses device stripped of all camera capability, relying instead on six embedded microphones and a voice-activated AI assistant.
  • Luna is set to debut at Meta's developer conference on September 23–24 and reach consumers in October, with the company betting that hands-free voice AI is a compelling enough offering on its own.
  • The deeper tension remains unresolved: whether consumers will embrace smart glasses that cannot photograph or record, and whether voice-only interaction can carry the product category forward.

In the wake of revelations that subcontractors accessed intimate videos recorded unknowingly by wearers of its Ray-Ban smart glasses, Meta is introducing Luna — a camera-free device that trades visual capture for voice-activated AI. The move reflects a deeper reckoning the technology industry faces: that the tools designed to augment human perception can, without careful restraint, become instruments of surveillance. By removing the lens entirely, Meta is asking whether trust can be rebuilt not through better safeguards, but through deliberate subtraction.

Meta is developing a new pair of smart glasses codenamed Luna — and their most defining feature is what they lack: a camera. The decision is a direct response to a privacy crisis that surfaced earlier this year, when Swedish media revealed that subcontractors working for Meta in Kenya had reviewed intimate and explicit videos recorded by Ray-Ban Meta glasses wearers, often without those users knowing they were being filmed. Personal financial information was also among the content accessed.

Luna will instead embed six microphones in the frame, allowing users to summon Meta's AI assistant through a button press and interact entirely through voice. There will be no photo or video capability — and therefore no footage of bystanders, no intimate recordings, nothing to end up in the hands of data labelers halfway around the world.

Meta had attempted to address earlier concerns through hardware safeguards: a white LED that flashes during recording, and newer designs that disable the glasses entirely if that light is covered. But after discovering users were drilling out the indicator to film others covertly, the company hardened the design further. Meta's head of Instagram, Adam Mosseri, acknowledged publicly that while Meta cannot stop people from filming others, it can make recording visible — and work to prevent the footage from being weaponized on its platforms.

Those measures, it seems, were not enough. Luna represents a more radical conclusion: that the only reliable privacy protection is elimination of the camera itself. Meta plans to unveil the device at its developer conference on September 23–24, with a consumer launch expected in October.

The open question is whether the market will follow. Smart glasses have long been defined by their ability to capture the world visually. Meta is now wagering that voice-first AI access — hands-free, screenless, ambient — is a sufficient draw on its own, and that consumers wary of surveillance will find the trade-off worthwhile.

Meta is betting that the future of smart glasses doesn't need to see anything at all. The company is developing a new device codenamed Luna—a pair of glasses stripped of the one feature that has made its existing smart eyewear so controversial: a camera. The move represents a direct response to a privacy crisis that emerged earlier this year when Swedish media outlets revealed that Meta subcontractors in Kenya had accessed intimate and explicit videos recorded by Ray-Ban Meta glasses wearers, often without those users knowing they were being filmed.

The Luna glasses will instead rely on six microphones embedded in the frame and voice-activated artificial intelligence. Users will be able to summon Meta's AI assistant with a button press on the side of the glasses, accessing information and services through spoken commands rather than visual recording. The device is also designed to work with Meta's consumer AI agent called Muse. There will be no photo or video capture capability—no camera to record the world around the wearer, and critically, no camera to record other people without their knowledge.

Meta plans to unveil Luna at its developer conference on September 23 and 24, with the glasses expected to reach consumers in October. The timing is deliberate. The company has spent months defending the privacy protections built into its existing Ray-Ban Meta glasses, which include a white LED light that flashes when recording begins. Newer versions are designed to stop functioning entirely if that light is covered or tampered with. Ray-Ban's own documentation states the LED remains active throughout any recording session, and users can control recording through physical buttons or voice commands.

Yet these safeguards proved insufficient to contain the damage. The Swedish investigation detailed how workers for a Meta subcontractor reviewed thousands of videos as part of a process called data labeling—the work of training AI models to recognize and categorize images. Many of the videos they examined appeared to have been captured without the wearer's awareness. The content included not just intimate moments but also personal financial information like bank account details. Workers described reviewing explicit material that users had no idea was being recorded and later accessed by strangers.

Meta's head of Instagram, Adam Mosseri, acknowledged the company's struggle with the problem in a public statement. He noted that while Meta cannot prevent people from filming others, it can make the recording visible through the LED indicator. He also explained that the company had hardened the glasses against tampering after discovering users were drilling out the light to disable it. The company views this as a form of harassment prevention—an attempt to stop people from secretly recording others and then posting the footage on Meta's platforms.

But the Luna pivot suggests Meta has concluded that the best way to address privacy concerns is not to build better safeguards around cameras, but to eliminate them entirely. A voice-first device cannot record video of bystanders. It cannot capture the intimate moments that ended up in the hands of Kenyan subcontractors. It shifts the interaction model from visual documentation to audio interaction, fundamentally changing what the glasses can do and, by extension, what they can do to others.

The question now is whether consumers will accept smart glasses that cannot take pictures or record video—whether the appeal of AI-powered eyewear survives the removal of its most obvious use case. Meta is betting that voice is enough, that the convenience of hands-free AI access outweighs the loss of a camera. The developer conference will reveal whether the company's vision for a privacy-respecting smart glasses future can actually compete with the camera-equipped alternatives already on the market.

We can't stop people from filming other people, but we can try to make sure that that light is on if they are filming. We now actually have the glasses basically stop working if you drill that light out.
— Adam Mosseri, Meta head of Instagram
Quer a matéria completa? Leia o original em Livemint ↗
Fale Conosco FAQ