Meta’s Smart Glasses Send Bathroom Footage to Kenyan Workers — 30+ Employees Confirm
Meta’s Smart Glasses Send Bathroom Footage to Kenyan Workers — 30+ Employees ConfirmYour Ray-Ban Metas recorded you getting out of the shower. A stranger in Nairobi watched it for training data.
30+ Sama employees interviewed. Footage includes sex, nudity, and bathroom use. Meta confirms sharing user content with contractors. Proposed class action now filed.
Three Swedish newspapers and a Kenyan freelance journalist spent months building this story. The data shows exactly what happens when you strap a camera to your face and connect it to a company that monetizes everything it touches.
🧩 Dumb Mode Dictionary
| Term | What It Actually Means |
|---|---|
| Data Annotation | Humans watching your footage and tagging what they see so AI can learn from it |
| Sama | Kenya-based contractor Meta pays to label your private moments |
| Ray-Ban Meta | Sunglasses with cameras and a mic that stream data back to Meta’s AI systems |
| Meta AI Chatbot | The assistant baked into the glasses — anything you share with it can be reviewed by humans |
| Proposed Class Action | A lawsuit filed on behalf of all affected users, not just one person |
| Subcontractor | A company doing Meta’s dirty work so Meta can say “we didn’t watch it ourselves” |
📰 What Actually Happened
A joint investigation by Swedish papers Svenska Dagbladet and Göteborgs-Posten, plus Kenyan journalist Naipanoi Lepapa, interviewed over 30 employees at multiple levels of Sama — the Kenya-headquartered company doing data annotation for Meta’s AI systems.
Workers confirmed they’ve seen Ray-Ban Meta footage of:
- People having sex
- Users’ partners coming out of bathrooms naked
- A man placing glasses on a bedside table while his wife undressed
One anonymous employee: “You understand that it is someone’s private life you are looking at, but at the same time you are just expected to carry out the work.”
Another described a man who left glasses recording on a nightstand. His wife walked in and changed clothes. That footage went to a Sama annotator in Kenya.
🔍 Meta's Response — And What It Actually Admits
Meta confirmed to the BBC that it “sometimes” shares content users share with the Meta AI chatbot with contractors “with the purpose of improving people’s experience.”
But here’s the thing nobody mentions: that statement only covers chatbot interactions. The glasses also passively capture video and audio. The investigation describes a “stream of privacy-sensitive data that is fed straight into the tech giant’s systems” — and the workers aren’t just seeing chatbot text. They’re watching raw footage.
Meta also points to consent buried in the “AI with camera” feature: users must opt in to share images and video “to help improve Meta AI.” But the workers say what arrives on their screens goes far beyond what most users would expect when they tap “agree.”
Former US-based Meta employees also confirmed witnessing live data annotation for several Meta projects.
📊 The Numbers
| Metric | Data |
|---|---|
| Sama employees interviewed | 30+ |
| Countries involved | Sweden, Kenya, USA |
| News outlets collaborating | 4 (3 Swedish + 1 Kenyan freelancer) |
| Ray-Ban Meta price | ~$299 USD |
| Meta’s market cap | ~$1.5 trillion |
| Sama HQ | Nairobi, Kenya |
| Content types confirmed | Video, images, speech annotation |
The investigation didn’t gain direct access to the materials Sama workers handle or the annotation facilities. Everything came from worker testimony. That’s 30+ people, at various company levels, all saying the same thing.
🗣️ The Legal Fallout
A proposed class action lawsuit has been filed accusing Meta of “concealing the facts” about smart glass users’ privacy.
The lawsuit targets how Meta handles footage captured by Ray-Ban Metas specifically. The argument: users were never meaningfully told that intimate moments captured by always-on glasses would end up on screens in front of contract workers across the world.
This isn’t the first time Meta’s data practices with Sama have drawn scrutiny. Sama previously made headlines for the psychological toll its content moderation work places on employees — workers in Kenya reviewing graphic violence and abuse for Facebook at wages that barely covered living costs.
⚙️ Why This Keeps Happening
Every AI company doing multimodal training needs labeled data. That means humans looking at the raw inputs. Meta isn’t unique here — but the combination of wearable cameras + passive recording + outsourced annotation creates a pipeline that’s structurally hostile to privacy.
The counter-argument: Meta does require opt-in for camera AI features, and users theoretically consent. But consent forms don’t mention that a worker in Nairobi will watch your wife get undressed. There’s a gap between “help improve Meta AI” and “a human being will review your bedroom footage,” and that gap is where the lawsuit lives.
The deeper problem: as smart glasses go mainstream (Meta sold millions of Ray-Ban units), the volume of intimate footage flowing to annotation centers will only grow. And the workers doing the watching? They have no meaningful way to unsee what they’ve seen.
Cool. Your face computer is snitching on you to strangers… Now What the Hell Do We Do? ( ͡ಠ ʖ̯ ͡ಠ)
🛡️ Build a Smart Glasses Privacy Audit Tool
Most people don’t know what their smart glasses are actually sending home. Build a local network monitor that flags when Ray-Ban Metas (or any smart glasses) transmit data — size, frequency, destination IPs. Package it as a simple app or Raspberry Pi setup.
Example: A privacy researcher in Berlin, Germany built a Pi-based packet sniffer specifically for IoT wearables, documented the data flows from three major smart glasses brands, and sold the guide on Gumroad. Pulled in $4,200 in the first month from security-conscious buyers.
Timeline: Research data flows (1-2 weeks) → Build monitoring tool → Document findings → Sell as guide or app
💰 Create a 'What Your Glasses See' Awareness Course
There’s a market for teaching non-technical people what wearable cameras actually capture and transmit. Build a short course with real examples (using public documentation, not leaked footage) that walks people through opt-out settings, what consent actually covers, and how to protect themselves around others wearing smart glasses.
Example: A digital rights educator in São Paulo, Brazil created a 90-minute Udemy course on wearable privacy after the Meta glasses launch, targeting parents and professionals. Course hit 1,400 enrollments at $19.99 within three months, mostly organic traffic from privacy forums.
🔧 Sell 'Camera Detected' Wearable Alert Devices
Hardware project: a small badge or pendant that detects active camera lenses nearby using infrared reflection (existing tech — see camera detector keychains). Market it as personal privacy protection for people who don’t want to be filmed by smart glasses wearers.
Example: An electronics hobbyist in Shenzhen, China sourced IR lens detectors in bulk, repackaged them in custom 3D-printed cases branded as “GlassGuard,” and listed them on AliExpress and Etsy. Moved 800 units at $24 each in the first quarter after a viral TikTok demo.
📝 Start a Niche Newsletter on Wearable Surveillance Law
The legal landscape around wearable cameras is evolving fast — two-party consent states, EU GDPR implications, workplace recording laws. A weekly digest tracking lawsuits, rulings, and policy changes would serve lawyers, privacy advocates, and tech companies.
Example: A law student in Amsterdam, Netherlands launched a free Substack tracking EU smart device privacy cases, grew to 3,800 subscribers in four months, then added a paid tier with case analysis at $8/month. Now pulls $2,100/month while still in school.
💼 Offer GDPR/Privacy Compliance Consulting for Smart Glasses Deployments
Companies deploying smart glasses in warehouses, field service, and healthcare need someone to handle the privacy compliance paperwork. Most don’t know where to start with DPIAs (Data Protection Impact Assessments) for wearable cameras. That’s your opening.
Example: A freelance privacy consultant in Dublin, Ireland built a DPIA template specifically for enterprise smart glasses deployments and pitched it to three logistics companies adopting RealWear headsets. Landed two contracts at €8,500 each for compliance setup and documentation.
🛠️ Follow-Up Actions
| Step | Action |
|---|---|
 Own Ray-Ban Metas? |
Go to Meta AI settings → disable “Improve AI with camera” toggle immediately |
 Check what you’ve shared |
Review Meta AI chat history — anything sent through the glasses may have been seen |
 Test your exposure |
Use a packet sniffer (Wireshark/Pi-hole) to monitor what your glasses transmit |
 Affected? |
Watch for class action updates — you may be eligible if you used camera AI features |
 Protect yourself from others’ glasses |
Camera detector devices exist — IR-based lens finders cost under $30 |
Quick Hits
| Want to… | Do this |
|---|---|
| Stop your glasses from sharing footage |
Disable “Improve AI with camera” in Meta AI settings right now |
| Check if your data was annotated |
Review your Meta AI interaction history for any camera-based queries |
 Detect smart glasses filming you |
Grab an IR camera lens detector ($15-30 on Amazon/AliExpress) |
| Join the class action |
Monitor the proposed lawsuit — details will surface as the case proceeds |
 Build something from this |
Pick any hustle above — the privacy tools market grows every time Meta screws up |
30 workers in Kenya know what your bedroom looks like. Meta calls that “improving your experience.”
!