The workers behind Meta's smart glasses can see everything
7 hours ago
- #AI Ethics
- #Privacy Concerns
- #Data Exploitation
- Meta's new AI smart glasses are marketed as an all-in-one assistant with features like live translations and facial recognition.
- A hidden workforce in Nairobi, Kenya, manually annotates sensitive data, including private videos showing intimate moments, bank details, and more.
- Workers in Kenya describe discomfort with handling deeply private content, including bathroom visits, sex scenes, and naked individuals unaware of being recorded.
- Meta's subcontractor, Sama, employs data annotators who label and describe images to train AI, often involving sensitive and private data.
- Despite sales assurances, data from the glasses is shared with Meta servers, contradicting claims of full user control over privacy.
- Meta's privacy policy allows for human review of interactions with AI, including sensitive conversations, without clear user consent.
- Experts highlight transparency issues under GDPR, questioning whether users truly understand the extent of data collection and processing.
- Former Meta employees reveal that sensitive data can still end up in training datasets, despite intended anonymization measures.
- Meta's response to inquiries is vague, referring to terms of use without directly addressing concerns about privacy violations.
- The investigation raises ethical concerns about the global data economy, exploitation of low-income labor, and user privacy risks.