Workers report watching Ray-Ban Meta-shot footage of people using the bathroom

Meta's privacy practices for its AI-powered Ray-Ban smart glasses face intense scrutiny following a Swedish investigative report revealing that contractors have watched highly sensitive user footage. The report, based on interviews with over 30 employees at Kenya-based subcontractor Sama, details how workers annotating video, image, and speech data for Meta's AI systems have viewed private moments captured by the glasses—including people changing clothes, using bathrooms, and engaging in sexual activity. Meta confirmed to the BBC that it "sometimes" shares user content with contractors to improve AI experiences, implementing privacy filters like face blurring, but the report suggests these measures may be insufficient for the intimate nature of first-person wearable footage.

Meta's privacy policy states that photos, videos, and audio from Ray-Ban Metas are sent to the company when users enable cloud processing, interact with Meta AI, or upload media to Facebook/Instagram. The policy explicitly notes that "trained reviewers" may process this data, including conversation transcripts with the AI chatbot. While Meta claims users can disable these features and delete recordings, the default activation of "Meta AI with camera" via voice commands and complex privacy documentation likely leaves many users unaware their private moments could be reviewed. This incident highlights the critical tension between AI training data needs and personal privacy, especially as always-on wearable devices become more common.