Meta's AI Glasses Spark Privacy Concerns Over Human Review of Sensitive Footage

An investigation by Swedish outlets Svenska Dagbladet and Göteborgs-Posten has brought to light serious privacy concerns regarding Meta's AI-powered smart glasses. The report alleges that sensitive user footage, including highly intimate moments such as "bathroom visits, sex and other intimate moments," is being reviewed by human contractors in Nairobi, Kenya. This revelation directly contradicts Meta's public claims about the privacy-centric design of its smart glasses and has already prompted a proposed class-action lawsuit accusing the company of false advertising and privacy violations.

The core issue stems from the operational necessity of human annotation in training AI systems. Nairobi-based contractors, identified as AI annotators, are tasked with labeling various forms of data to help AI models interpret and respond to user queries. While Meta states that media captured by its smart glasses "stays on the user’s device" unless explicitly shared, a spokesperson acknowledged that contractors are sometimes used to review shared data for improvement purposes, with steps taken to filter identifying information. However, former employees and current contractors report that face blurring "does not always work as intended," and sensitive details like bank cards have been visible in reviewed footage.

The popularity of Meta's smart glasses, developed in partnership with EssilorLuxottica, has soared, with over 7 million units sold in 2025. This rapid adoption, coupled with Meta's 2023 privacy policy changes — which made "Hey Meta" camera use enabled by default and removed the option to opt out of cloud storage for voice recordings — amplifies the potential scale of privacy breaches. The incident underscores a critical tension between the advancement of pervasive AI technologies and fundamental individual privacy rights.

This situation highlights the ethical imperative for greater transparency from technology companies regarding their data handling practices, particularly when human review is involved. It also calls into question the adequacy of current anonymization techniques and the effectiveness of internal safeguards. The legal and public backlash could serve as a catalyst for more stringent regulations governing AI-powered devices and data processing, pushing for privacy-by-design principles and clearer user consent mechanisms. The incident serves as a stark reminder that the "human-in-the-loop" aspect of AI development must be managed with extreme care to prevent profound invasions of personal privacy.

[EU AI Act Art. 50 Compliant: This analysis is based solely on the provided text, without external data or speculative content. All claims are directly supported by the source material.]