Meta faces fresh questions about user privacy after a report claimed that contractors watched sensitive footage captured by its smart glasses. The devices, sold as Ray‑Ban Meta smart glasses, allow users to record video, take photos, and interact with AI features. The report suggests that some of that recorded material reached human reviewers who work on data labeling tasks for the company.
The investigation came from Swedish newspapers Svenska Dagbladet and Göteborgs‑Posten, working with Kenyan freelance journalist Naipanoi Lepapa. Their report draws on interviews with more than 30 workers connected to the outsourcing firm Sama. Sama provides data annotation services for several technology companies, including Meta Platforms.
Data annotation involves reviewing images, videos, or audio and tagging them so machine learning systems can learn from them. Companies use this process to train AI models and improve features such as speech recognition, image detection, and chatbot responses.
Several workers told reporters they had seen private moments recorded by users of the smart glasses. Some described videos that showed people changing clothes or walking out of bathrooms naked. Others said they saw recordings taken inside homes, including living rooms and bedrooms.
One worker described a video where a man left the glasses on a bedside table before leaving the room. His partner then entered and changed clothes. The worker said moments like that made them aware they were watching private parts of someone’s life.
Another employee said the work often created discomfort. Workers understood the footage came from real people who may not know that strangers could see the clips during AI training tasks. Still, the employees said they had to complete the annotation work as part of their job.
Meta Smart Glasses: Privacy, Policy, and the People Behind the Lens
Meta confirmed that it sometimes uses contractors to review certain data. In statements shared with BBC, the company said the practice helps improve products and user experience. Meta also said it filters the data first to protect privacy. For example, the company may blur faces in images before sending material to reviewers.
The company’s wearable privacy policy explains when media from the glasses may reach its servers. Photos and videos can upload when users enable cloud processing, interact with the built-in AI assistant, or share content through services like Facebook or Instagram.
Meta also stores audio and video from livestreams created with the glasses. Text transcripts and voice recordings produced by its AI assistant can also enter company systems. According to the policy, Meta uses machine learning and trained reviewers to process this information. The company may also share the data with outside vendors that help maintain and improve its products.
Users can view and delete recordings through the Meta AI mobile app. The policy also warns people not to share information they want to keep private when they interact with AI services.
Meta’s Smart Glasses Face Legal and Regulatory Scrutiny
Critics say the rules may still confuse many users. Some owners of the smart glasses may not read the detailed policies or understand how their recordings move through the system. Workers interviewed for the report said they sometimes saw clips that appeared to be recorded by accident. Examples included footage of people’s bank cards or explicit videos playing on screens.
The glasses show a small red light when they record. Critics argue that people nearby may miss the light or fail to understand what it means.
The report has caught the attention of regulators and privacy advocates as well. The UK data watchdog has contacted Meta about the claims. At the same time, the company is planning new features for its smart glasses. A report by The New York Times stated that Meta is planning to add the feature of facial recognition to its smart glasses.
The controversy has also reached the courts. A proposed class-action lawsuit has been filed against Meta and its parent company, EssilorLuxottica, which owns the Ray-Ban brand. The lawsuit claims that the marketing campaign for the glasses promised that it would provide good privacy protection, without disclosing that human reviewers could view the recorded content.
The lawsuit is seeking damages and changes to Meta’s way of doing business. Meta has not commented on the lawsuit.
