Meta is working on prototype smart glasses that could listen continuously and take photos every few seconds, according to the Financial Times. The reported feature, called “super sensing,” would let a wearer ask Meta AI questions about what the glasses have recently heard or seen.
That is the useful version of the pitch: a computer on your face with enough context to answer questions about your day. The less tidy version is also obvious: people near the wearer could be swept into an audio and image collection system without doing anything except existing in range of Meta hardware.
How the reported system would work
The Financial Times reported that one proposed design would avoid saving raw footage or audio at Meta and would not give the wearer direct access to those recordings. Instead, the system would extract metadata from the images and sound, upload that derived information to Meta’s servers, and let Meta AI search it later.
That distinction matters, but it is not magic privacy dust. Metadata can still describe what happened, who or what may have been present, and when the device detected it. According to the Financial Times, supporters of the proposed design argue that using extracted metadata rather than retaining raw recordings would reduce privacy risks.
The prototypes have not been announced as a shipping product. The Financial Times described them as work in progress, and the reported features could change or never be released. Still, the mechanism being considered is clear enough: persistent sensing feeds an AI assistant, and the assistant answers based on a processed memory of the wearer’s surroundings.
The recording light problem is not theoretical
The report lands while Meta’s existing smart glasses are already under scrutiny. The company has faced attention over reported facial-recognition work for its glasses, according to earlier coverage cited by The Verge. CNN has also reported on people filming women while wearing the glasses.
Meta has separately had to deal with modders offering paid services to remove the LED recording indicators from its Ray-Ban Meta glasses, according to The Verge. On Tuesday, Meta announced an update that will disable the camera if the glasses detect tampering with that privacy light.
The Financial Times reported that Meta’s current plan for “super sensing” mode would leave the LED recording indicator off. That would be a sharp policy choice for a product category already asking bystanders to trust a tiny light on somebody else’s face. If Meta ships this, the fight will not be about whether AI glasses can remember things. It will be about who gets noticed, who gets queried, and who gets no say in either.
This story draws on original reporting from The Verge.
