What Happened
On March 5, 2026, Clarkson Law Firm filed a class action lawsuit against Meta in the US District Court for the Northern District of California, San Francisco Division. The suit, filed on behalf of Meta AI Glasses users, alleges that Meta deliberately deceived consumers about the privacy protections of its Ray-Ban smart glasses.
The core allegation: workers at Sama, a Nairobi-based data subcontractor, were manually reviewing footage captured by users' AI glasses to train Meta's AI systems. That footage included people using the bathroom, changing clothes, credit card numbers, and explicit sexual activity. Swedish media first reported workers' accounts of the content they were exposed to.
Plaintiffs Gina Bartone of New Jersey and Mateo Canu of California allege false advertising and privacy law violations. Meta sold over 7 million pairs of the glasses in 2025, and the suit claims buyers were misled by Meta's privacy marketing. Luxottica of America, the glasses' manufacturing partner, is also named in the lawsuit.
Why It Matters
This lawsuit sits at the intersection of two major issues in AI: the hidden human labor behind "AI-powered" features, and the privacy implications of always-on cameras in consumer products.
When you use Meta's AI glasses, your interactions get processed to improve the system. That processing is not purely algorithmic. Real people in outsourced facilities look at your data. This is how most AI training works - human annotators label data so models can learn - but most consumers buying smart glasses do not understand that their personal footage may end up on a screen in a data labeling facility.
The 7 million units sold number makes this potentially massive in scope. If the class is certified, it could represent one of the largest consumer privacy actions related to AI hardware.
For the broader AI industry, this case could set precedent on disclosure requirements. Every major AI company uses human data annotators. The question the court will address is whether companies need to explicitly tell users that humans - not just algorithms - will review their data, and whether marketing materials that emphasize privacy create a false impression.
Our Take
This was predictable and, frankly, overdue. The AI industry has a transparency problem with human review. When you upload a document to an AI tool, send a voice message to an assistant, or walk around wearing AI-enabled glasses, there is a reasonable chance a human will eventually see that data. Most companies bury this in terms of service that nobody reads.
Meta's situation is worse because of the form factor. A phone app that processes text is one thing. Glasses that record video of your daily life - including moments you would never intentionally share with anyone - are fundamentally different. The expectation of privacy is higher, and the potential for capturing sensitive content is built into the product's design.
For anyone using AI tools with camera, microphone, or screen-sharing features, this is a reminder to check what data gets retained and who reviews it. That includes meeting transcription tools, AI assistants with screen access, and any product that processes visual or audio data.
The practical advice: assume anything you send through an AI system could be reviewed by a human. If you would not want a stranger seeing it, do not let a camera controlled by a tech company record it. That is not how privacy should work in 2026, but it is how privacy does work right now.