Meta’s AI-powered smart glasses could beryllium sending delicate footage to quality reviewers successful Nairobi, Kenya, according to an investigation by the Swedish outlets Svenska Dagbladet and Göteborgs-Posten. The report, which was published past week, claims Meta contractors successful Kenya person seen videos captured pinch the smart glasses that show “bathroom visits, activity and different friendly moments.”
So far, astatine slightest 1 proposed people action lawsuit accusing Meta of violating mendacious advertizing and privateness laws has emerged successful consequence to Svenska Dagbladet’s reporting, citing the company’s declare that its smart glasses are designed for privacy:
By affirmatively claiming that the Glasses were designed to protect privacy, Meta assumed a work to disclose worldly facts that would pass a reasonable consumer’s determination to acquisition the product. Instead, Meta hid the alarming reality: that usage of the AI features results successful a alien halfway astir the world watching the astir backstage moments of a person’s life.
The Nairobi-based contractors interviewed by Svenska Dagbladet are AI annotators, meaning they explanation images, text, aliases audio, pinch the extremity of helping AI systems make consciousness of the information they’re training on. “We spot everything — from surviving rooms to naked bodies,” 1 worker says, according to Svenska Dagbladet. “Meta has that type of contented successful its databases.”
A erstwhile Meta worker reportedly tells Svenska Dagbladet that faces successful note information are blurred automatically, though workers successful Kenya opportunity this “does not ever activity arsenic intended,” and immoderate faces are still visible. Another personification reportedly tells the outlet that a wearer’s slope cards are sometimes seen successful the footage they reappraisal arsenic well.
Meta’s Ray-Ban and Oakley smart glasses travel pinch a built-in AI adjunct tin of answering questions astir what a personification tin see. The glasses person soared successful fame successful caller years, contempt increasing concerns complete privateness and surveillance.
EssilorLuxottica, the eyewear elephantine that Meta useful pinch to create the camera-equipped glasses, sold complete 7 cardinal of the AI-powered glasses successful 2025 — more than tripling its sales successful 2023 and 2024 combined. Last year, Meta made immoderate changes to its privateness policy that support Meta AI pinch camera usage enabled connected your glasses “unless you move disconnected ‘Hey Meta.’” It besides stopped allowing wearers to opt retired of storing their sound recordings successful the cloud.
As reported by Svenska Dagbladet, the Kenya-based AI reviewers activity pinch transcriptions arsenic well, ensuring Meta AI provides the correct reply to the questions users inquire aloud. In a connection to The Verge, Meta spokesperson Tracy Clayton says media captured by its smart glasses “stays connected the user’s device” unless they take to stock it pinch different group aliases Meta.
“When group stock contented pinch Meta AI, we sometimes usage contractors to reappraisal this information for the intent of improving people’s experience, arsenic galore different companies do,” Clayton says. “We return steps to select this information to protect people’s privateness and to thief forestall identifying accusation from being reviewed.”
The UK’s Information Commissioner’s Office has questioned Meta astir the claims successful Svenska Dagbladet’s reporting. Privacy advocates person raised concerns astir Meta’s alleged goals to build facial nickname into its smart glasses arsenic well, pinch the Electronic Privacy Information Center calling it a “grave consequence to privacy, safety, and civilian liberties.”
Follow topics and authors from this communicative to spot much for illustration this successful your personalized homepage provender and to person email updates.
English (US) · 
Indonesian (ID) ·