Meta Ai Glasses: Your Private Moments May Be Unveiled Overseas

Meta AI Glasses: Unseen Watchers

What you thought was private might not be so private after all. According to a recent revelation by Swedish outlet Svenska Dagbladet, Meta Ray-Ban AI smart glasses are sending a surprising amount of user-recorded footage to offshore contractors for data labeling. This process, essential for training AI models, involves human reviewers who might witness some of your most intimate moments.

These glasses are undeniably popular, with sales skyrocketing to over seven million pairs in 2025. They offer the tech-savvy a way to capture first-person video and audio, with the added bonus of analyzing surroundings using Meta's AI tech. Yet, with great innovation comes greater controversy. Critics are buzzing about potential facial recognition features and the tech giant's notorious past with privacy issues.

The Human Element in AI Training

Though AI models are often portrayed as fully automated, the reality is quite different. Human contractors play a vital role in training these systems by reviewing and annotating footage. A recent joint investigation revealed that contractors in Nairobi, Kenya, are handling some extremely sensitive data from these AI glasses.

“In some videos you can see someone going to the toilet, or getting undressed,” shared a contractor from Sama, highlighting the depth of personal exposure involved.

Alarmingly, other footage includes bank card images, adult content views, and even "sex scenes." The uncomfortable truth is that users might never suspect that their private moments could be seen by human eyes.

Unavoidable Data Sharing

Meta's AI terms of use vaguely hint at this privacy breach by stating that the company might "review your interactions with AIs," both through automation and human oversight. However, it seems many users aren't fully aware of these terms.

What’s more, users have no choice but to agree to data sharing if they wish to use AI features. "Once the material has been fed into the models, the user in practice loses control over how it is used," explained data protection lawyer Kleanthi Sardeli. After two months of silence, a Meta spokesperson referred to the AI Terms of Service, yet the concerns remain unresolved.