Meta’s latest push into smart glasses has run into the same problem that sank earlier face-worn gadgets: people do not like wondering whether the stranger across from them is recording. According to Victoria Song of The Verge, the company’s recent launch of cheaper smart glasses without Ray-Ban branding, promoted with Kylie Jenner, has rekindled a blunt privacy fight on Threads, Meta’s own social platform.
The argument is not only about Meta. Song, who covers wearables, describes a broader class of AI devices, including glasses, pins, pendants and rings, that can capture audio or video with little obvious warning. Their selling point is discretion. Their social problem is also discretion. Funny how that works.
Meta has been releasing smart glasses more aggressively since its Ray-Ban Meta glasses performed better than expected in 2023, according to Song. The company later introduced display glasses with a small screen in the right lens and a wrist-worn controller for gestures. Its newest non-Ray-Ban glasses arrived at a lower price point and with a celebrity campaign, which helped shove the privacy debate back into public view.
Some critics on Threads have cast the glasses as tools for creeps, Song reported. Author Namina Forna wrote in one widely liked Threads post, “We all agree that the Meta glasses are for perverts, yes?” Other posts cited by Song called for public shaming, grabbing glasses from wearers, or using smart glasses to record harassers back.
The technical reality is less cinematic than the panic. Song notes that Meta’s glasses are not capable of daylong video or audio surveillance in normal use. Continuous Live AI video, long calls, or roughly 10 3K videos can drain a full battery in under an hour, according to her testing. That limitation does not erase the problem: short clips, stealthy recording and hardware modifications are enough to make people uneasy.
Supporters have pointed to accessibility uses and hands-free family videos, Song reported. Others have compared smart glasses to phones, which can also record in public. That comparison only gets so far. A phone held up to film is visible. Glasses sit on someone’s face, pointed wherever that person looks, which is the sort of design choice that makes privacy lawyers reach for coffee.
The debate has also been sharpened by reporting from The New York Times and Wired, which Song said found Meta has considered facial recognition features for the glasses. That is a separate step from recording, and a more volatile one, because it could connect a live face to identity data.
Song also tested the Vocci AI note-taking ring, which starts recording with a button press and shows a recording light she described as insufficiently bright. She said the ring worked well for consensual interviews and voice notes during tech events. She also found it easy to record a spouse, friends and colleagues without them noticing, though she said she disclosed those tests and deleted the recordings afterward.
Apple’s AirTag offers one imperfect comparison. Song wrote that AirTags can be abused, but Apple added unwanted tracking alerts after criticism, including from domestic abuse advocates. Those alerts do not stop misuse. They do make stalking harder and less attractive.
AI wearables lack an equivalent warning system. LED recording lights can be overlooked, washed out in daylight or tampered with, Song reported. She points to louder shutter sounds, physical lens covers and removable camera modules as possible fixes. Xreal already uses a removable camera attachment on some glasses, though the design is bulkier.
Meta has considered modular hardware, Alex Himel, Meta’s vice president of wearables, told Song. He said the tradeoff is that glasses would be heavier, less integrated and worse-looking. Himel also said Meta knows people are trying to defeat privacy lights and that more robust privacy updates are coming.
Private venues are already reacting. Song cited examples of companies and locations restricting smart glasses or denying entry to Meta glasses wearers. Zenni Optical is also selling lenses marketed as anti-facial-recognition protection.
The uncomfortable lesson is that AI wearables do not need to be powerful enough for spy fiction to create real social friction. They only need to be cheap, ordinary-looking and good enough to make bystanders doubt what is happening in front of their faces.
This story draws on original reporting from The Verge.
