Meta's AI Smart Glasses' Intimate Footages Reviewed by Actual Humans in Kenya, Privacy Lawsuit Reveals

Somewhere in Nairobi, contract workers sat in a monitored office drawing boxes around objects in video footage. Some of that footage showed people undressing. Some showed sex. Some showed bathroom visits that the wearers of Meta's Ray-Ban AI smart glasses almost certainly never intended anyone else to see.

A class action lawsuit filed on Wednesday in federal court in San Francisco now puts those claims at the centre of a legal fight over what Meta actually meant when it marketed the glasses as 'designed for privacy, controlled by you.'

The Clarkson Law Firm brought the complaint on behalf of Gina Bartone of New Jersey and Mateo Canu of California. Both allege Meta violated privacy laws and engaged in false advertising, according to TechCrunch. The suit names Meta and glasses manufacturing partner Luxottica of America as defendants.

More than seven million pairs were sold in 2025. Footage from those devices feeds into a data pipeline for human review, and users cannot opt out.

How A Swedish Investigation Blew The Lid Off Meta's Review Process

The lawsuit traces back to a joint investigation published on 27 February by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten. Reporters found that data annotators employed by Sama, a subcontracting firm in Nairobi, had been labelling images and videos captured by the glasses to train Meta's AI systems, Decrypt reported.

The contractors did not mince words. One told the Swedish papers they had seen people using the toilet and getting undressed, and that users probably had no idea the footage was being watched. Others described viewing credit card numbers and explicit sexual content filmed by wearers.

Meta has said it uses AI to blur faces before footage reaches the annotation teams. Workers at Sama disputed that. The blurring did not consistently work, they said, particularly in low light. A former Meta employee backed that up, telling the Swedish outlets the anonymisation algorithms sometimes failed.

Footage gets flagged for review when users interact with Meta AI. Asking the glasses to identify an object or describe a scene can route video straight to human contractors. Meta's US AI terms of service do acknowledge that interactions 'may be automated or manual (human)', but the complaint argues that language was buried deep in the fine print.

UK Watchdog Writes To Meta Over Smart Glasses Privacy Concerns

The UK's Information Commissioner's Office called the allegations 'concerning' and confirmed it has written to Meta demanding information on how the company meets its obligations under UK data protection law, The Register reported. 'Devices processing personal data, including smart glasses, should put users in control and provide appropriate transparency,' the ICO spokesperson said.

Meta spokesperson Christopher Sgro responded to the broader issue but not the lawsuit itself. He said media stays on a user's device unless they choose to share it. 'When people share content with Meta AI, we sometimes use contractors to review this data for the purpose of improving people's experience, as many other companies do,' he said, Engadget revealed.

Sama's History With Meta And The Workers Left Holding The Footage

Sama is not new to controversy. The Nairobi firm previously handled content moderation for Meta but walked away from that work in 2023 after reports about worker trauma. It pivoted to computer vision data annotation, which is precisely the kind of work now tied to the smart glasses, Help Net Security noted.

Workers told the Swedish journalists they operated under strict confidentiality agreements. Offices were monitored by cameras. Personal phones were banned. Several said the material made them uncomfortable, but the pay kept them there.

The Clarkson Law Firm, which has previously brought major cases against Apple, Google, and OpenAI, is seeking monetary damages and injunctive relief. The complaint seeks to hold Meta accountable for what it calls 'affirmatively false advertising' and a failure to disclose the true nature of its AI data collection pipeline.