Meta’s Ray Ban AI glasses have grown in popularity in recent years and so have their sales seven crore pairs This is more than the two million sold in 2025 and 2023 and 2024 combined.
While smart glasses have gained huge popularity among consumers, allowing them to record first-person footage through an integrated camera and microphone array and analyze the world around them through Meta’s AI models, the hardware has sparked a heated debate. Critics say enabling facial recognition in glasses’ software can have dangerous effectsEspecially given the militarization of law enforcement and Meta’s poor track record when it comes to ensuring users’ privacy.
And regardless of the wearer’s intent, much of the footage being recorded by the glasses is being sent to offshore contractors for data labeling, a widely used preprocessing step in training new AI models in which human contractors are asked to review and annotate the footage. This is a laborious and highly resource-intensive process that tech companies often overlook when discussing the power of their latest AI models.
Reality may be messed up. Meta contractors based in Nairobi, Kenya told Swedish newspapers Svenska Dagbladet And goteborg-posten In a recently published joint investigation They are being asked to review highly sensitive and intimate data.
“In some videos you can see someone going to the toilet or taking off clothes,” said a contractor with a company called Sama. “I don’t think they know, because if they knew they wouldn’t be recording.”
“I saw a video in which a man puts glasses on the bedside table and walks out of the room,” one data annotator told newspapers. “After a while his wife comes and changes her clothes.”
Other footage included photos of people’s bank cards, users viewing porn, or even filming entire “sex scenes”.
One employee said they felt forced to watch and comment otherwise risk losing their job.
“You understand it’s someone’s private life you’re looking into, but at the same time you’re just expected to get the job done,” the employee said. “You shouldn’t question it. If you start asking questions, you’re gone.”
buried in Meta’s AI Terms of UseThe Company reserves the right to “review your interactions with AI, including the content of your interactions or messages with the AI, and this review may be automated or manual (human).”
The document also warns that users should not share information that “you do not want the AI to use and retain, such as information about sensitive topics.”
But given the kind of information data interpreters are being asked to review, it seems that many users are not aware of that last part of the advice.
Worst of all, owners of Meta’s AI glasses don’t have the option to use AI features without agreeing to share the data they share with Meta’s remote servers. And once the data is sent, it is often too late.
“Once content is put into the model, in practice the user loses control over how it is used,” explained Clenthy Sardelli, data protection attorney at the nonprofit Nun of Your Business. Svenska Dagbladet And goteborg-posten.
After receiving no response for two months, a Meta spokesperson contacted two Swedish newspapers about its terms of use and privacy policy.
“When Live AI is being used, we process that media in accordance with the Meta AI Terms of Service and Privacy Policy,” the spokesperson said in a brief statement.
It’s not just Meta using offshore data annotators to train its AI models in countries like Kenya, Colombia and India. As Agence France-Presse As reported last year, workers have often had to review gruesome crime scene images and even dead bodies.
reminds of trend social media content moderationA practice that has relied on exploitative labor in developing countries for many years.
But with the advent of AI and wearable technology, which can be used to easily record high-resolution footage by simply tapping a capacitive button next to your temple, the hidden human cost of data labeling has taken on a new meaning.
It’s a reality that Meta would prefer to bury lengthy terms of service that only a handful of people will take the time to read.
One commentator told newspapers, “You’d think that if they knew about the extent of data collection, no one would dare use the glasses.”
More on Meta: Meta’s top AI scientist is stepping down amid Zuckerberg’s spending spree
