Meta's AI glasses under scrutiny as workers flag users' private footage

Two Swedish newspapers report that Meta's AI smart glasses are allegedly recording private and intimate footage, with contractors in Kenya tasked with reviewing the content to train the AI systems

Meta’s AI-powered smart glasses are allegedly collecting the daily lives of users, including highly intimate material, often without them realising it, according to reports by Swedish newspapers Göteborgs-Posten and Svenska Dagbladet.

 

The reports state that Meta, through its subcontractor Sama in Nairobi, employs data annotators to help train artificial intelligence systems. They label images, transcribe audio and assess how AI tools respond to users. These data annotators have described reviewing sensitive videos recorded through the glasses.

 

"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 Svenska Dagbladet.

 

The glasses allow users to activate an AI assistant with voice commands, process images and record short videos. 

 

"We see everything – from living rooms to naked bodies. Meta has that type of content in its databases. People can record themselves in the wrong way and not even know what they are recording. They are real people like you and me," another worker said.

 

Further elaborating on privacy concerns, a worker said some videos filmed through the device include people having sex. The workers also said that their own personal devices are banned inside the office to prevent leaks.

 

What does Meta say?

Meta’s terms state that "in some cases, Meta will review your interactions with AIs, including the content of your conversations with or messages to AIs, and this review can be automated or manual (human)".

 

According to a Meta spokesperson quoted by The Telegraph, "When people share content with Meta AI, like other companies, we sometimes use contractors to review this data to improve people’s experience with the glasses, as stated in our privacy policy. This data is first filtered to protect people’s privacy. We take the protection of people’s data very seriously and we’re continuously refining our efforts and tools in this area".

 

Former Meta employees also told the newspapers that faces appearing in annotation data are automatically blurred. However, Kenyan workers claim anonymisation tools do not always function properly.

 

"The algorithms sometimes miss. Especially in difficult lighting conditions, certain faces and bodies become visible," a former Meta worker said.

 

Data annotators said they feel uncomfortable reviewing such videos but continue the work to make ends meet.

 

"You understand that it is someone’s private life you are looking at, but at the same time you are just expected to carry out the work. You are not supposed to question it. If you start asking questions, you are gone," they said.