Meta's 'Super-Sensing' Smart Glasses Would Record Everything — Without Telling Anyone

What happens when a pair of glasses listens to everything you say and captures a photo every few seconds, without a blinking light to warn you? Meta is about to find out — and the answer may determine the future of wearable computing.

The company is developing a prototype internally called “super-sensing” AI smart glasses, according to multiple people familiar with the project. The device can keep audio recording running continuously and automatically snap photos at intervals of just a few seconds, feeding the captured data to Meta’s AI for real-time queries.

What makes the project particularly contentious: Meta’s executive team plans to leave the frame-mounted LED recording indicator dark while these features are active. Bystanders would have no obvious way to tell the glasses are collecting data.

The feature set could eventually reach existing Ray-Ban Meta glasses already on the market. People familiar with internal discussions said Meta is considering pushing the super-sensing capabilities to current hardware through an over-the-air software update. So the privacy implications aren’t hypothetical — they could land on millions of faces.

Meta has publicly addressed the LED question. In a 2025 policy document, the company argued that a constantly blinking LED would train people to ignore it, reducing situational awareness rather than improving it. Under the proposed approach, the LED lights up only during “active capture” scenarios — when the user manually saves a photo or video — but stays off during “AI feature” usage.

One internal proposal would avoid storing raw footage altogether. Instead, the glasses would extract metadata from audio and video streams, upload only anonymized data to servers, and let Meta AI respond to queries without retaining the originals. Company executives are also debating whether to use the collected data for training Meta’s AI models.

Dave Arnold, a Meta spokesperson, declined to comment on the prototype but told The Verge: “We’re committed to getting glasses right — they need to be loved by the people wearing them and the people around them. Our approach is to build privacy in from the start and develop new technology that helps people throughout their day.”

CEO Mark Zuckerberg framed the ambition during Meta’s Q1 2026 earnings call: “I’m really excited to see glasses evolve from something that answers questions into a personal agent that’s with you all day, helping you remember things and achieve goals.”

This isn’t Meta’s first brush with smart glass controversy. In March, the company wrote that its Ray-Ban Meta glasses would evolve from a tool that requires active prompting into a “more continuous, ambient, always-on assistant.”

The existing hardware has already drawn scrutiny. In February, Kenyan content moderators told media outlets they were shown explicit footage captured by the glasses — including users having sex and using the bathroom. Last month, Wired discovered embedded code for an unreleased facial recognition system inside the Meta smart glasses platform; the code was removed within a day. Third-party modification services already offer to physically remove the LED indicator from the glasses for a fee.

In a separate move Tuesday, Meta pushed an update that automatically disables the camera when the glasses detect the LED indicator has been tampered with — a tacit acknowledgment of the modding trend.