A privacy lawsuit against Meta Platforms over its AI-powered Ray-Ban smart glasses has widened significantly, with an amended complaint filed August 31 in the U.S. District Court for the Northern District of California adding a new class of plaintiffs: people who never purchased or wore the glasses at all, but say they were recorded by someone else’s pair without their knowledge.
How the case started
The lawsuit, originally filed March 4, 2026 by plaintiffs Gina Bartone of New Jersey and Mateo Canu of California, names both Meta Platforms and Luxottica of America, the eyewear giant that co-produces the Ray-Ban Meta line. The complaint draws heavily on a February 2026 investigation by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten, which reported that workers at Sama, a Kenya-based data-labeling subcontractor, reviewed video clips captured through the glasses as part of Meta’s AI training pipeline. According to that reporting and the court filing, the reviewed footage reportedly included bathroom visits, nudity, sexual activity and other intimate moments captured inside users’ homes.
The core allegation: marketing versus practice
Plaintiffs argue Meta marketed the glasses as “designed for privacy, controlled by you,” while in practice routing captured recordings to third-party contractors who could view and label the material to improve Meta’s AI systems. The complaint frames this as a mismatch between privacy-forward advertising and the actual data pipeline underneath the product — a gap that, if proven, could expose Meta to liability under California’s consumer-protection and privacy statutes.
Why the August amendment matters
The original suit was limited to people who bought and wore the glasses. The amended complaint filed August 31 broadens that scope to bystanders never given a chance to consent — a legally significant expansion because it shifts the case from a product-disclosure dispute (did buyers understand what they signed up for?) to a wiretapping-style claim (can someone be recorded and analyzed by AI without ever being a customer at all?). Legal commentators tracking the case, including coverage from Fortune, have described the amendment as an attempt to test whether Meta’s data practices violate the privacy rights of the general public, not just its own users.
Meta’s defense
Meta has pushed back on the characterization, stating that unless a user chooses to actively share captured media with Meta or others, that footage stays on the user’s device. The company says that when people do share content with Meta AI, contractors are sometimes used to review data to improve the product experience, but that Meta takes steps to filter that data to protect people’s privacy. Meta has not disputed that the Sama-based review process existed, but disputes the plaintiffs’ characterization of how systematically sensitive footage was retained and viewed.
A second front: privacy advocates and regulators
Consumer-privacy organizations have seized on the case as evidence that camera-equipped wearables need clearer, real-time recording indicators and stricter limits on what footage can be routed into AI training sets. The Clarkson Law Firm has separately pursued class-action claims on similar grounds, arguing that Ray-Ban and Oakley-branded Meta glasses owners were not adequately informed about how their recordings could be used. The dispute also lands amid a broader regulatory mood shift: Norway’s government said this month it is drafting rules that could restrict or ban camera-enabled wearables entirely, citing many of the same bystander-consent concerns now at the center of the Meta litigation.
What it means and what’s next
The case is still in its early stages, and Meta is expected to move to dismiss or narrow the amended claims before any trial date is set. But the litigation already illustrates a structural tension facing every company selling AI-powered wearables: the more useful the glasses become — translating conversations, identifying objects, answering questions about what the wearer sees — the more raw footage of the world, and the people in it, the underlying AI needs to process and often store. Whether courts ultimately treat that footage the same way they treat a phone camera’s photos, or hold AI-training pipelines to a higher consent standard, will shape not just Meta’s product line but the entire emerging category of camera-based AI wearables now being built by Snap, Google and others.
Unlike a smartphone, which a subject can usually see being pointed at them, glasses-mounted cameras are built to look ordinary, and the recording indicator — a small LED near the lens — is easy to miss in a crowded room or dim lighting. Privacy engineers who have reviewed the Meta case note that even a perfectly compliant on-device policy cannot resolve the deeper issue: once footage of a bystander is captured, the question of who reviews it, where it is stored, and whether it ever reaches a human labeler thousands of miles away depends entirely on backend practices that are invisible to everyone in the room, including the person wearing the glasses.
Photo: planet_fox / PIXABAY via Pixabay
