Ray-Ban Meta camera smart glasses, at the center of a privacy report about contractors reviewing footage
Image: Meta

People who wear Meta’s camera glasses may not be the only ones who see what those glasses capture. A joint investigation published Friday by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten reports that workers at a Meta contractor in Nairobi, Kenya, have reviewed images and video recorded by the glasses, including footage of people in some of the most private moments of their lives.

The report focuses on Sama, a data annotation company that Meta uses to label content for training its AI systems. According to the newspapers, footage that reaches the workers can include scenes recorded inside homes, bathrooms and bedrooms. None of the workers’ accounts has been independently confirmed by Meta, and the company has not addressed the specific claims in detail.

What the investigation found

Annotators at Sama told the Swedish journalists that their job is to describe, tag and categorise objects and actions in images and video so that Meta’s AI can learn to recognise them. In the course of that work, they said, they have seen clips of people undressing, using the toilet and having sex while the glasses were recording. “We see everything,” one worker is quoted as saying in the English version of Svenska Dagbladet’s report.

Workers also described seeing bank cards caught on camera, people watching pornography, and transcripts of conversations touching on crimes, protests and sexual topics. They said they work under strict non-disclosure agreements, in offices with cameras and without personal phones, and that raising concerns could cost them their jobs.

One recurring claim concerns face blurring. Former Meta employees in the United States told the newspapers that faces are blurred automatically before review. The Kenyan annotators said the blurring does not always work, particularly in poor lighting, and that faces and bodies are sometimes visible.

Meta’s response

According to the reporting, Meta did not answer the journalists’ specific questions. The company’s position was that data processing follows its terms of use and privacy policy for Meta AI. Sama did not respond to the newspapers.

Meta markets its glasses as designed “with your privacy in mind.” Its policies allow it to use content that people share with Meta AI to improve its products, and those policies say data can be transferred, stored and processed around the world.

The questions raised by the investigation are less about whether Meta reviews AI data, which many tech companies do, and more about whether owners understand what they agree to when they ask the glasses’ assistant to look at something. Many people may not expect a photo of their kitchen, or of whoever is in it, to end up in front of a human reviewer on another continent.

Why it matters for smart glasses

Camera glasses are now the fastest-growing corner of the XR market. Earlier this month, EssilorLuxottica, which makes Ray-Ban Meta and Oakley Meta glasses with Meta, said it sold about seven million units in 2025, more than three times its 2024 total. That scale turns a niche privacy question into a mainstream one, both for owners and for the people around them who never agreed to be recorded.

The story is also likely to draw attention from European regulators. The EU has no adequacy decision covering Kenya, the mechanism the bloc uses to approve transfers of personal data to countries outside it, so any footage of Europeans sent to Nairobi would need to rely on other legal safeguards. Meta has not said whether footage from European users is part of the material reviewed there.

Sama has drawn scrutiny before. In 2023 it ended its content moderation work for Meta after reports of worker trauma and disputes over labour practices, and shifted toward computer vision annotation.

What owners can check in settings

Meta changed its smart glasses privacy policy last April, as TechCrunch reported at the time. Since then, Meta AI with camera use stays switched on unless the owner turns off “Hey Meta” voice commands entirely, and voice recordings are stored by default. Users can delete recordings in the companion app, but they cannot opt out of storage while still using voice features.

Meta has said photos and videos taken with the glasses stay on the user’s phone unless they are shared with Meta AI or another service. In practice, that means the most direct way to keep footage away from reviewers is to avoid asking the AI assistant to analyse anything sensitive, and to review the AI and voice settings in the Meta AI app.

Owners who are worried can also go through the app’s history and delete stored interactions. None of those steps changes what may already have been reviewed, and Meta has not said how long reviewed clips are kept or how many people have access to them.

Meta has not announced any changes to its review process in response to the report.

Related coverage on Virtual Reality Times

Sources

https://virtualrealitytimes.com/wp-content/uploads/2026/02/ray-ban-meta-sales-2026-02-11-600x338.jpghttps://virtualrealitytimes.com/wp-content/uploads/2026/02/ray-ban-meta-sales-2026-02-11-150x90.jpgVirtual Reality TimesAR HeadsetsMetaPhilosophy and EthicsMeta,Meta AI,privacy,Ray-Ban Meta,smart glassesPeople who wear Meta's camera glasses may not be the only ones who see what those glasses capture. A joint investigation published Friday by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten reports that workers at a Meta contractor in Nairobi, Kenya, have reviewed images and video recorded by the glasses,...VR, Oculus Rift, and HTC Vive News - Cryptocurrency, Adult, Sex, Porn, XXX