Meta's Smart Glasses Scandal: Workers Who Reported Seeing Users' Sex Acts Lost Their Jobs — Then 1,108 More Were Cut

In late February 2026, Swedish newspapers Svenska Dagbladet and Göteborgs-Posten published a joint investigation, conducted in collaboration with Kenyan journalists, that exposed a chain of events Meta had not disclosed to the public [1][2]. Footage captured by users of Meta's Ray-Ban smart glasses — including videos of people undressing, using the bathroom, engaging in sexual activity, and handling sensitive financial documents — was being transmitted to servers and then routed to contract workers at Sama, a data annotation firm in Nairobi, Kenya [3][4]. Those workers were tasked with manually labeling the footage to train Meta's AI systems. Some who objected to the material they were forced to review lost their jobs [5]. Six weeks later, Meta terminated the Sama contract entirely, and 1,108 workers received redundancy notices [6][7].

The fallout has produced a class action lawsuit in U.S. federal court, a separate $1.6 billion labor suit in Kenya, a formal inquiry from the UK's Information Commissioner's Office, and questions from multiple members of the European Parliament [8][9][10].

What the Workers Saw

According to the Swedish investigation, Sama employees in Nairobi described reviewing footage that users had opted to share with Meta AI for product improvement. The clips included people in their bathrooms, people changing clothes, credit card and bank account details, users watching pornography, and in multiple cases, footage of sexual activity recorded by glasses left on bedside tables or worn during intimate encounters [2][3].

"In some videos you can see someone going to the toilet, or getting undressed. I don't think they know, because if they knew they wouldn't be recording," one worker told the investigators [4]. Another said faces that Meta claimed were blurred in review pipelines were sometimes clearly visible [11].

Meta had marketed the Ray-Ban smart glasses with slogans including "designed for privacy, controlled by you" and "built for privacy" [12]. The company's public position was that media captured by the glasses stays on the user's device unless the user chooses to share it. When users did opt into sharing content with Meta AI, the company acknowledged using contractors to review the data "for the purpose of improving people's experience" and said it took "steps to filter this data to protect people's privacy" [3][13].

The gap between that marketing and the workers' accounts became the foundation of the lawsuits that followed.

The Firings and Contract Termination

The complaint filed in Kenyan courts by 185 former Sama content moderators alleges that workers who raised concerns about the sensitive content they were reviewing and about their working conditions were fired in retaliation [5][14]. One worker quoted in the Swedish reporting put it bluntly: "You are not supposed to question it. If you start asking questions, you are gone" [4].

The labor dispute between Sama moderators and Meta predates the smart glasses scandal. Daniel Motaung, a South African recruited to Sama's Nairobi hub in 2019 to moderate Zulu-language content on Facebook, filed a lawsuit in May 2022 after he was terminated following an attempt to unionize workers [15][16]. Motaung described nine-hour shifts reviewing videos of violence, rape, and dismemberment, with minimal psychological support [15]. In April 2026, the Employment and Labour Relations Court in Kenya ruled in favor of 187 former moderators in their case against Meta and Sama [17].

Then came the contract termination. On April 16, 2026, Sama announced redundancies affecting 1,108 employees at its Nairobi delivery center after Meta formally ended their partnership [6][7]. The layoffs represent one of the largest single workforce reductions in Kenya's digital economy. The moderators' legal representatives allege that ending the contract was retaliation for organizing and speaking out, not a legitimate business decision [5][14]. Meta and Sama have not publicly confirmed or denied this characterization; Sama said it "engaged with Meta to try to save the work" but that those efforts "have not been successful" [6].

A Product With 9 Million Users

The scale of the privacy exposure tracks with the product's commercial success. Meta and EssilorLuxottica sold over 7 million Ray-Ban smart glasses in 2025 alone, tripling the previous year's sales and bringing cumulative units sold to approximately 9 million [18][19]. Meta captured over 70% of the global smart glasses market in the first half of 2025 [20]. EssilorLuxottica has announced plans to increase annual production capacity to 10-30 million units [18].

Source: Road to VR / UploadVR

Data as of Apr 1, 2026CSV

Each pair of those glasses has a forward-facing camera capable of recording video and taking photos. The Electronic Frontier Foundation noted in a March 2026 advisory that the glasses automatically import media to Meta's mobile app by default, that AI features require sending footage to Meta's servers with no local processing option, and that conversations with Meta AI are saved by default [21].

The "Pervert Glasses" Problem

Beyond the contractor review issue, a BBC investigation found dozens of male influencers using Meta's smart glasses to secretly film women in public and post the videos online without consent, earning the product the nickname "pervert glasses" [22][23]. The glasses are designed to look like ordinary Ray-Bans, and multiple reviews have noted that friends and bystanders did not realize the glasses contained cameras [24]. Meta's glasses have a small LED indicator light that activates during recording, but the BBC documented multiple videos showing how it can be covered or disabled [22].

Meta's official position is that users should not engage in "harmful activities like harassment, infringing privacy rights, or capturing sensitive information" [22]. Critics note that this places the compliance burden entirely on the user while selling a device specifically designed to make recording inconspicuous.

Google Glass vs. Ray-Ban Meta: Lessons Not Learned

The privacy controversy around wearable cameras is not new. Google Glass, launched to "Explorer" users in 2013 and briefly sold to the U.S. public in 2014, generated immediate backlash [24][25]. Bars and restaurants banned the device. Critics dubbed wearers "glassholes." Google pulled Glass from the consumer market in January 2015, and the entire experiment was formally shuttered in 2023 [24].

The critical difference: Google Glass was unmistakably a piece of technology. Its distinctive prism display made it obvious the wearer had a camera-equipped device. Meta's Ray-Ban glasses, by contrast, are deliberately designed to be indistinguishable from ordinary eyewear [24]. This design choice addressed the social stigma that killed Google Glass — but by solving it, Meta created a more acute privacy problem. Where Google Glass wearers were identifiable and could be asked to stop recording, Meta smart glasses wearers blend in.

Meta did add the LED recording indicator that Google Glass also used. But unlike the Google Glass era, when the product sold in small quantities to a tech-enthusiast niche, Meta has placed 9 million recording-capable devices into the general population, creating a surveillance surface area that Google never approached [18][19].

The Legal Landscape

U.S. Class Action

On March 4, 2026, plaintiffs Gina Bartone of New Jersey and Mateo Canu of California, represented by the Clarkson Law Firm, filed a class action suit against Meta Platforms and Luxottica of America in the U.S. District Court for the Northern District of California [3][12]. The complaint alleges violations of federal and state privacy laws and false advertising, arguing that Meta's privacy-centric marketing was contradicted by the undisclosed transmission of footage to human contractors overseas.

Kenya Labor Suit

The 185 former Sama moderators are pursuing a $1.6 billion lawsuit against Meta, Sama, and Majorel in Kenyan courts, alleging unfair dismissal, exploitative working conditions, inadequate mental health support, blacklisting from similar roles with other contractors, and wages as low as approximately $1.46 per hour after taxes [5][14]. In April 2026, the court ruled in favor of the moderators on preliminary matters [17].

Source: NPR / TechCrunch / Kenyan Wall Street

Data as of Apr 30, 2026CSV

Precedent: The $52 Million PTSD Settlement

Meta has been down this road before. In 2020, Facebook agreed to pay $52 million to settle a class action brought by more than 10,000 content moderators who developed PTSD from reviewing graphic content [26][27]. The settlement provided $1,000 per moderator, with additional damages of up to $50,000 for those diagnosed with psychological conditions related to their work [26]. The company also agreed to provide counseling and to assess candidates' emotional resiliency before hiring [27].

The current Kenyan claim of $1.6 billion dwarfs that settlement by a factor of 30. If even a fraction of the claimed amount is awarded, it would represent one of the largest labor judgments in the tech industry's history.

Regulatory Response

United Kingdom

The Information Commissioner's Office confirmed on March 5, 2026 that it was writing to Meta to formally request information about how the company was meeting its obligations under UK data protection law [9][10]. The ICO has the authority to impose fines of up to £17.5 million or 4% of global annual turnover, whichever is higher [10].

European Union

Multiple members of the European Parliament have submitted questions to the European Commission asking whether Meta's data transfer practices — specifically, sending European users' footage to contractors in Kenya — are compatible with GDPR requirements [10]. Under GDPR Article 9, processing of biometric data and data concerning a person's sex life is subject to heightened protections. The EU's AI Act, which began phased implementation in 2025, could also apply to the AI training pipeline that the footage feeds into, though specific enforcement actions related to smart glasses had not been announced as of late April 2026.

United States

No federal legislation specifically targeting ambient-recording wearables has been introduced, though the controversy has renewed calls for comprehensive federal privacy legislation [28]. Individual states with two-party consent recording laws — including California, where Meta is headquartered — may provide additional legal avenues for affected users [28].

Whistleblower Protections and the Retaliation Question

The question of whether the fired workers qualify for whistleblower protections depends on jurisdiction and the specifics of their employment arrangements. In California, Labor Code §1102.5 prohibits employers from retaliating against employees who report suspected violations of law. Federal OSHA anti-retaliation provisions offer similar protections. However, the Sama workers were based in Kenya, employed by a Kenyan entity, and subject to Kenyan labor law — which has its own protections but a different enforcement apparatus [5][14].

The Kenyan court's April 2026 ruling in favor of the moderators suggests that the workers' claims of retaliatory termination have legal traction under Kenyan law [17]. Whether parallel claims under U.S. whistleblower statutes could be brought — potentially by arguing that Meta, as the ultimate beneficiary of the labor, bears responsibility — remains an open legal question.

As for whether there is a legitimate non-retaliatory explanation: Meta could argue that the Sama contract ended for standard business reasons, including the company's broader workforce restructuring that cut approximately 8,000 jobs across the organization in early 2026 [29]. Sama's statement that it tried to preserve the contract but was unsuccessful is consistent with either interpretation — retaliation or routine business decision. The timing, however, is notable: the contract was terminated weeks after the Swedish investigation went public and amid active litigation.

The Core Tension

Meta has built a product that 9 million people have purchased specifically because it makes recording seamless and invisible. The company's AI ambitions require training data, and the glasses generate a unique stream of first-person footage that no other Meta product captures. Training AI on that footage requires human reviewers to label the data. Some of that data will inevitably include intimate moments, because the glasses are worn during intimate moments.

Each step in this chain follows logically from the preceding one. The result is a system where contract workers in Nairobi are paid roughly $1.46 an hour to watch strangers' private lives unfold on their screens — and those who object are, according to the workers' own accounts and the Kenyan court's findings, removed [4][5][17].

The Electronic Frontier Foundation's March advisory summed up the underlying design problem: "Meta designed the glasses' data pipeline to route through its servers, and therefore to require users to trust Meta's handling of intimate data" [21]. The Swedish investigation showed what happens when that trust is tested, and the Sama layoffs showed what happens to the people caught in the middle.

Meta did not respond to specific questions about the relationship between the Swedish investigation and the subsequent contract termination. The company has previously said it "takes steps to filter this data to protect people's privacy" [3].

The glasses remain on sale.