Meta is developing smart glasses that record continuously, capturing audio and photos every few seconds. The prototype, called “super sensing” wearables, would let users ask Meta’s AI about the recorded data, though the raw footage might never be stored or accessible to the wearer.
How the glasses could work
The glasses would extract metadata from images and audio instead of storing full recordings. That data would be uploaded to Meta’s servers, where the AI could analyze it without keeping the original content. This method could ease privacy concerns by avoiding direct storage of raw footage.
It remains unclear whether the captured data would train Meta’s AI models. The company may also add similar features to existing glasses models in the future.
Meta has previously explored always-on recording. A March blog post about its Ray-Ban Meta glasses described a move toward a “more continuous, in-the-moment assistant” that doesn’t need manual prompts. CEO Mark Zuckerberg reinforced that goal during a Q1 2026 earnings call, calling the evolution of glasses into a “personal agent” a top priority.
Privacy concerns arise
The project emerges as Meta faces scrutiny over privacy and safety in its wearable devices. The company has already addressed reports of users filming women without consent and modders selling services to disable the glasses’ LED recording indicator. This week, Meta released an update that disables the camera if the LED is tampered with.
However, the “super sensing” mode introduces new issues. Meta plans to keep the LED off during this mode, arguing that constant activation could make users ignore the indicator. A July 2025 whitepaper stated the LED would only light up during “active capture” scenarios, like saving photos or videos, to prevent confusion about when recording occurs.
Without a visible indicator during passive recording, bystanders may not know they’re being filmed. Meta has faced legal challenges over facial recognition and data collection, and an always-on device would likely fuel those debates further.
A Meta spokesperson declined to comment on internal prototypes but stressed a focus on privacy. “Our approach has been to develop new technologies that help people throughout their day, with privacy built in from the start,” Dave Arnold said. He added that the glasses must work for both wearers and those around them.
Related: iRobot unveils manual floor cleaner
For now, the project is still in development. If released, the glasses could change how people use AI and how much of daily life is recorded.
Meta’s history with wearables shows a pattern of pushing boundaries. The Ray-Ban partnership already brought smart features to sunglasses, blending fashion with functionality. These new prototypes take that idea further, turning glasses into constant companions that document the world without explicit input.
Even if raw footage isn’t stored, metadata alone can reveal sensitive details about people’s movements, conversations, and environments. The lack of transparency about when recording happens only adds to the unease.
Meta’s update to block tampered LEDs is a small step. Yet it doesn’t address the core issue: how to balance convenience with respect for others’ privacy. The company’s whitepaper suggests confidence in its approach, but public trust may take longer to build.
The debate isn’t new. Wearable cameras have long sparked discussions about consent and control. What sets Meta’s project apart is the scale. With millions of potential users, even minor oversights could have major consequences.
Zuckerberg’s vision of a “personal agent” reflects a broader trend in tech. Devices are becoming more proactive, anticipating needs before users even ask. But that shift requires careful handling of data and clear communication about how it’s used.
Meta hasn’t shared a timeline for the glasses’ release. For now, the prototypes remain behind closed doors, where engineers refine the technology and lawyers assess the risks. When—or if—they reach the public, the reaction will shape the future of wearable AI.
