Meta is facing a US privacy lawsuit alleging that recordings captured by its artificial intelligence glasses were sent to human contractors in Kenya and used to help train its AI systems without adequate disclosure.
The amended complaint, filed in federal court in California on August 31, expands the case beyond people who bought or wore the glasses to include bystanders who say they were recorded without their knowledge or consent.
The claim alleges that Meta marketed the devices as “designed for privacy, controlled by you”, while failing to explain that material captured when AI features were activated could be transmitted to the company’s servers and reviewed by third-party workers.
It argues that even people who noticed the glasses and agreed to be recorded could not have consented to their images and conversations being combined with other data, stored, exploited and inspected overseas.
Meta AI glasses privacy claims
The glasses combine cameras and microphones with Meta’s AI assistant, allowing users to ask questions about what they see, identify objects and landmarks, translate information and take photographs and videos hands-free.
The lawsuit alleges that imagery and audio collected while those features were being used were routed through Meta’s systems and ultimately fed into the development of its AI models. It also claims that Meta’s facial-anonymisation tools did not always conceal people’s identities.
The allegations followed a February investigation by Swedish newspapers Svenska Dagbladet and Göteborgs-Posten, which reported interviews with workers at Sama, a data-annotation company operating in Nairobi.
Those workers allegedly described reviewing material recorded by Ray-Ban Meta glasses, including footage showing people changing clothes, using bathrooms and engaging in intimate activity. The reports also referred to visible financial and other personal information.
Meta has since ended its work with Sama. In a statement reported by the Swedish newspapers, the company said it had paused the partnership while investigating the allegations and had decided to terminate it because Sama did not meet its standards.
Meta said at the time that photographs and videos were private to users and that human review of AI content was undertaken to improve product performance where the company had obtained user consent.
Ryan Clarkson, founder and managing partner of the Clarkson Law Firm and co-lead counsel in the litigation, said the case reflected a wider push by major technology companies to obtain more material for AI development.
“I think there is a huge push to improve AI tools by every corporation, especially the largest corporations in the world,” he said. “And the way that they believe is the best way to improve their AI is to find more signal, to find more content, more recordings that they can feed into this machine, feeding the beast.”
He said the firm had heard from people who had never purchased or used the glasses but believed they had been captured by them in everyday situations.
“It even resulted in this term called perv glasses,” Mr Clarkson said.
Meta disputes the allegations and said it would defend the case. A spokesperson said the glasses were designed to let people use AI hands-free to answer questions about the world around them, adding that data might be reviewed to improve products and users’ experiences “the same way as many other companies”.
The company said it took steps to filter the data, remove identifying information and protect people’s privacy.
Meta’s published guidance says the glasses have a capture light which signals when content is being recorded, and that users can manage or delete photographs, videos and voice interactions through its software. The company has also said that its privacy approach is intended to take account of people who do not use the technology.
The case remains at the litigation stage and the allegations have not been tested or proven in court.
