Ask Onix
UK regulator investigates Meta's smart glasses data practices
The UK's Information Commissioner's Office (ICO) is demanding answers from Meta after reports revealed outsourced workers reviewed sensitive footage captured by the company's AI-enabled smart glasses. The watchdog described the allegations as "concerning" and confirmed it would seek clarification on Meta's compliance with data protection laws.
How content review works
Meta acknowledged that subcontracted workers occasionally review videos and images recorded by its Ray-Ban Meta smart glasses to enhance the product's AI capabilities. The company stated that such reviews are outlined in its privacy policy and that data is pre-filtered to protect user privacy, including measures like blurring faces in images.
However, sources interviewed by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten claimed these safeguards sometimes fail, allowing workers to see unblurred faces and highly personal content. Workers reported viewing footage of users in private settings, including bathrooms and intimate moments.
Worker testimonies raise alarms
The investigation revealed that data annotators employed by Nairobi-based firm Sama were tasked with labeling content to train Meta's AI. These workers described reviewing explicit material, such as users watching pornography or inadvertently recording others without consent. In one case, a worker recounted footage of a woman undressing, filmed by a man's glasses left recording in a bedroom.
Sama, which has previously faced criticism for its content moderation work, maintains strict workplace policies, including surveillance cameras and bans on mobile phones. The company has since discontinued content moderation services, citing ethical concerns.
Meta's response and user awareness
Meta emphasized that users must manually activate recording or use voice commands, and the glasses feature a visible light to indicate when the camera is active. The company advises users to inform others when recording and avoid capturing private spaces. Despite these warnings, critics argue that many users remain unaware their content may be reviewed by humans.
In its UK AI terms of service, Meta states that interactions with its AI may be reviewed either automatically or manually. However, when asked by the BBC to specify which sections of its terms address human review, the company did not provide a direct answer.
Regulatory scrutiny and broader concerns
The ICO stressed that devices processing personal data must prioritize user control and transparency. "Service providers must clearly explain what data is collected and how it is used," the regulator said. The watchdog's intervention follows growing unease over the privacy implications of AI-powered wearables, which have been linked to incidents of non-consensual recording.
Meta's smart glasses, launched in partnership with Ray-Ban and Oakley, are part of a wave of AI-driven wearables designed to interpret visual and audio inputs. While these devices offer accessibility benefits-such as assisting visually impaired users-their rapid adoption has sparked debates about misuse and data security.
What's next
The ICO's letter to Meta marks the start of a formal inquiry into the company's data handling practices. Meanwhile, the BBC has sought comment from Sama and EssilorLuxottica, the parent company of Ray-Ban and Oakley, but has yet to receive responses.
