Glasses that see too much: How smart glasses are not only changing the future, but also privacy

Aditi Chandanani, who owns a pair of Meta glasses, always makes it a point to tell people when she is about to start recording; (right) Data privacy strategist Ruchi Gosalia’s concern about Meta glasses is the lack of privacy as it comes with AI face recognition tech

When content creator and mental health coach Aditi Chandanani wears her Meta glasses, she has developed a habit of telling people when they are switched on. She is not always recording, but she knows people cannot easily tell the difference. "People kept looking at my glasses before they said something," she says, recalling a trip to Dubai with friends. "I would tell them, ‘I'll tell you when I switch it on.'"

That uncertainty is at the centre of the debate around smart glasses. While devices like Meta's Ray-Ban smart glasses are marketed as a way to capture moments hands-free, critics argue that their biggest concern is not the wearer's privacy, but everyone around them. Unlike a phone camera, which requires a person to physically lift a device and point it, smart glasses sit naturally on the face. A tiny LED light indicates when recording is happening, but privacy experts argue you can switch the light off. Even if it is on, many people may not notice it or even know what it means.

Meta launched its second-generation Ray-Ban Meta glasses globally in 2023, adding AI features in 2024, including the ability to answer questions, identify objects, translate languages, and interact through voice commands. For Chandanani, who creates content using the glasses, the responsibility lies with the user.

She recalls filming a taxi ride where she was recording the view outside the window. During the ride, her boyfriend and the driver began discussing the Jeffrey Epstein files. Intrigued by the conversation, she turned her head, capturing them in the frame. Before uploading the clip, she asked for permission. "I told them I had recorded it. Their faces weren't visible, but their voices and opinions were there. If someone is uncomfortable, you can blur it or explain your intention," she says.

However, she believes certain boundaries should be obvious. "Don't shoot babies," she says, adding that while people may be comfortable being recorded casually, children and unaware strangers need more consideration. Chandanani also points out that the glasses have some built-in safeguards. The glasses stop recording after three minutes, limiting accidental extended recordings. But she keeps communication at the forefront.

Data privacy strategist Ruchi Gosalia believes the issue goes beyond social etiquette. "The problem is the small microphone and camera chip," she says. "A lot of people are not aware that they won't be able to make out if someone wearing those glasses is recording them."

According to Gosalia, the concern grows when smart glasses are combined with AI tools. A recorded image or video could potentially be processed by AI systems for tasks like identifying people or analysing surroundings. "Once face recognition is done, it can identify anyone's name and address," she says, highlighting fears around how biometric information could be misused. She also raises questions about where AI-related data travels after collection. Many AI systems rely on human reviewers and external processes for improving models, meaning user-generated information can sometimes move beyond the direct interaction between a person and a tool.

"The information is not only between us and AI but to other coutnries where the data is going. They have access to all the recordings so there is a higher chance of misuse and lack of safety," she says, "The question is what about the bystanders? I understand someone wearing the glasses has read Meta's privacy policy, but where do we get consent from the people whose photos or videos are captured?" This is what separates smart glasses from traditional cameras, she argues. "But through glasses, no one can make out someone can record," she says. "How can they question you or even stop you?"

The technology itself may not be the problem, but the invisible nature of it creates a new challenge; consent in a world where recording can happen without people knowing. Chandanani's solution is simply awareness. "Just tell people when you switch it on," she says, "People should know where to look to see if it's on or not."

Because with technology that asks others to trust what they cannot see, the question remains - how much trust should we place in a pair of glasses?