Overview
- A Svenska Dagbladet and Göteborgs-Posten probe, echoed by international outlets, reports that Nairobi-based contractors have reviewed clips showing nudity, toilet use, sexual activity, and visible credit card details from Meta’s Ray-Ban AI glasses.
- Meta did not dispute human review and referred questions to its AI Terms of Service and Privacy Policy, which state that interactions with AIs may be reviewed by automated systems or humans and that users must accept these terms to use AI features.
- Contractors working for Sama say they feel compelled to annotate disturbing footage and fear losing their jobs if they refuse, highlighting labor and ethical concerns tied to large-scale wearable data annotation.
- Key technical details remain unclear, including what triggers transfers to Meta’s servers, whether manual recordings are ever included, and how long such data is retained, with reporters noting frequent app contacts to Meta servers in Sweden and Denmark.
- Privacy experts cited in the reporting say cross-border transfers and limited transparency raise potential GDPR issues, and the scale of exposure is heightened by reports that more than seven million pairs were sold in 2025.