Bank data, bedroom scenes: Meta AI glasses see all
What's the story
Meta, the tech giant founded by Mark Zuckerberg, has come under fire for its new AI glasses. The innovative device was unveiled in September 2025 as an all-in-one assistant capable of real-time translations and facial recognition. However, a recent investigation by Svenska Dagbladet and Goteborgs-Posten revealed that these glasses have been feeding a continuous stream of privacy-sensitive data into Meta's systems.
Behind-the-scenes
Data annotators describe their work
The development of Meta's AI glasses isn't just a Silicon Valley affair. In Nairobi, Kenya, thousands of data annotators work for Sama, a subcontractor to Meta. These workers review and annotate footage captured by the device, training Meta's systems to recognize and interpret the world. They describe images and label objects like cars and lamps to make smart glasses more intelligent.
Ethical dilemma
Annotators see deeply personal moments
The data annotators have raised ethical concerns over their work. They come across deeply private video clips from people using the glasses in their daily lives. Some describe videos showing bathroom visits or sexual encounters, raising questions about consent and privacy. One worker said, "When you see these videos, it feels that way. But since it is a job, you have to do it."
Market impact
Surge in sales, but data processing remains murky
Despite the privacy concerns, Meta Ray-Ban glasses have seen a surge in sales. In 2025 alone, sales tripled to seven million units. However, when asked about how data from these glasses is processed, the responses were mixed. While some salespeople assured customers that they could choose what data to share with Meta, others were uncertain about where the data goes or if it is even collected at all.
User control
Impossible to use Meta glasses without internet connection
When reporters at Svenska Dagbladet bought their own pair of glasses, it required connecting to an app called Meta AI. The app requested permission to share additional data with Meta for product improvement. However, when they attempted to use the glasses without an internet connection, the device was unable to interpret their surroundings. Instead, the glasses prompted the reporters to reconnect, and consistent data pings to Meta's servers were observed during use.
Data usage
Meta's terms of service reveal potential privacy risks
Meta's AI terms state that "in some cases; Meta will review your interactions with AIs including content conversations messages AIs; this review can automated manual (human)." It also states AIs may store use information shared them; user should not share information "that you don't want AIs use retain such as information about sensitive topics." This highlights the potential for sensitive data to be used in ways users may not fully understand or agree to.
Regulatory concerns
GDPR applies to Meta in EU, expert says transparency lacking
As Meta offers services in the EU, it is subject to the General Data Protection Regulation (GDPR), which requires transparency about how personal data is processed. Kleanthi Sardeli, a data protection lawyer at None Of Your Business (NOYB), says there is a clear transparency problem with these glasses. She believes explicit consent should be required when data is used to train artificial intelligence models.