The people in the footage did not know
Investigations by Svenska Dagbladet and Göteborgs-Posten reported that annotators working for Sama in Nairobi reviewed footage captured by smart glasses. The material reportedly included people undressing, using bathrooms, having sex, and entering card details.
The people shown in those images had no reason to expect that a worker in another country might inspect the footage. The distance between wearer and reviewer does not soften that breach of expectation. It makes the chain of access harder for the subject to see.
Automatic blurring is not a complete safeguard
Former Meta employees told reporters that faces are blurred automatically. The annotators described cases in which anonymity failed and faces remained visible. The gap matters because automated privacy protections are often treated as if they settle the question before a human sees the material.
A safeguard needs an audit trail: when it ran, when it failed, who could access the original material, and what happens after an exception. Without that, a bystander has no reliable way to understand where a moment captured by glasses can travel.
The investigation makes the hidden work visible
Kenya's Office of the Data Protection Commissioner opened an investigation after a petition from The Oversight Lab, according to the supplied reporting. The story brings a neglected part of wearable AI into view: people who label or review data are part of the product's privacy model.
The question for smart-glasses makers is not confined to what a wearer can record. It also concerns where footage goes, who is able to view it, and whether a person caught on camera can ever know that review happened.