Seven million people bought Meta's Ray-Ban smart glasses. Now a class action lawsuit alleges that footage from those glasses - including videos of users undressing, having sex, and handling bank cards - was shipped to contract workers in Nairobi, Kenya for review. Users couldn't opt out.
The lawsuit, filed March 4 in federal court in San Francisco, is the sharpest example yet of a growing consumer revolt against AI devices designed to watch and listen around the clock. But Meta's glasses are just one front in a much wider backlash.
The Meta Smart Glasses Scandal
Swedish newspapers Svenska Dagbladet and Göteborgs-Posten broke the story: workers at data annotation firm Sama in Kenya were labeling visual content captured by Meta's glasses to train its AI models. The content included intimate moments users never expected anyone to see.
Meta's response was corporate boilerplate. The company said it "sometimes uses contractors to review this data for the purpose of improving people's experience" and that it "takes steps to filter this data to protect people's privacy." Given that contractors saw users on the toilet, those steps clearly weren't enough.
The UK's Information Commissioner's Office has since demanded answers from Meta about how the company meets its obligations under UK data protection law. The regulatory pressure is just starting.
Beyond Glasses: The Always-Listening Backlash
Meta isn't facing this alone. The broader category of always-on AI wearables is getting a hostile reception.
The Friend pendant - a $129 AI necklace that listens continuously through always-on microphones and sends you text messages based on what it hears - launched to a wave of public anger. New Yorkers vandalized subway ads with messages like "AI is not your friend" and "Make friends with something alive." The startup's terms reportedly allow captured data to be used for model training, and no independent security audit has been published.
The Humane AI Pin, which tried a similar always-on approach at $699 plus $24 per month, has effectively failed. The company is struggling, support may end soon, and the core product never worked reliably. Yet at CES 2026 in January, dozens of companies showed up with their own always-listening AI pendants, pins, and earbuds - apparently learning nothing from Humane's collapse.
One developer even published an app that detects when AI-enabled smart glasses are nearby, a kind of digital self-defense tool.
The Core Problem Nobody Has Solved
The fundamental tension is simple: these devices need to capture everything to be useful, but capturing everything means recording bystanders who never consented. In two-party consent states (like California, where many tech companies are headquartered), recording someone's conversation without their knowledge is illegal.
There's also the labor angle that the Meta lawsuit highlights. Training AI models on real-world footage requires human reviewers to label that footage. Those reviewers are typically low-paid contractors in countries with weaker labor protections, seeing content they never signed up to see. It's the same pattern that has plagued content moderation at social media companies for years, now extended to footage captured in people's bedrooms.
For anyone considering an always-on AI wearable right now, the practical advice is blunt: assume everything the device captures will be seen by strangers, stored indefinitely, and used to train models. If that's acceptable to you, go ahead. For most people, it shouldn't be.