The UK's Information Commissioner's Office (ICO) is contacting Meta following reports that outsourced workers viewed sensitive content captured by the company's AI smart glasses. Meta stated that subcontracted workers may review footage, including films and images, to improve the AI experience. Investigations by Swedish newspapers claim workers in Kenya reviewed videos of intimate moments, including instances of sexual activity and private domestic scenes.
Meta asserts that data is filtered to protect privacy, with measures like blurring faces. However, sources suggest these filters sometimes fail. Users activate recording manually or via voice command, and may not realize their data is human-reviewed, as outlined in Meta's privacy policies.
The ICO stated that devices processing personal data must offer users control and transparency, clearly explaining data collection and usage. The watchdog will seek information from Meta regarding its compliance with UK data protection laws.
These workers, identified as data annotators for Meta's AI, also reviewed transcripts of AI interactions. While workplace privacy measures were in place, the content reviewed was often highly sensitive. Meta's smart glasses feature a light indicating recording is active, and the company advises users to be mindful of recording in private spaces.