Meta is developing prototype smart glasses that would record nearly everything you see and hear. The “super sensing” glasses snap photos every few seconds while continuously capturing audio, then let wearers query Meta AI about what they’ve experienced. But here’s the twist – users might never actually see the raw footage themselves, according to sources speaking to the Financial Times. Instead, the system would process everything into metadata, raising fresh questions about privacy and who really controls your memories when AI does the remembering.

Meta is making its play for the always-on AI wearable market. The company’s working on prototype smart glasses that would constantly record your surroundings, capturing photos every few seconds while continuously listening to audio, according to sources who spoke with the Financial Times. It’s Meta’s answer to a question the tech industry keeps asking: what if your AI assistant could see and hear everything you do?

The “super sensing” glasses build on Meta’s existing Ray-Ban partnership, which already produces smart glasses with cameras and AI features. But this prototype takes things much further. Instead of manually triggering recordings, these glasses would passively capture your entire day, creating a continuous stream of visual and audio data that Meta AI could then parse and answer questions about.

Here’s where it gets interesting – and controversial. The raw footage and audio might never actually be available to you as a user, sources told the FT. Instead of storing photos and recordings you could review, Meta’s proposed system would process everything into metadata. Think of it as AI-digested memories rather than actual video playback. The system would analyze what you saw and heard, extract the meaning, then discard the originals.

That approach solves one problem while creating others. On the storage side, it’s elegant – metadata requires far less space than continuous video and audio. It also potentially addresses some privacy concerns since the raw recordings wouldn’t persist. But it raises fundamental questions about control and transparency. If your glasses are recording everything but you can’t access the footage, who really owns those experiences? And how do you verify what the AI is telling you about your own past?

Meta isn’t alone in chasing this vision. The always-on AI wearable category has attracted serious attention despite stumbling out of the gate. Humane’s AI Pin launched to withering reviews earlier this year, with critics panning its clunky interface and questionable utility. Rabbit’s R1 device faced similar skepticism. Both promised to be AI-powered assistants that could observe and interact with your world, but neither quite delivered on the vision.

The difference is Meta’s already got distribution. The company’s Ray-Ban Meta smart glasses have gained traction as a surprisingly functional product – stylish enough to wear, useful enough to justify the price. Adding always-on recording and more sophisticated AI capabilities to that existing platform is an easier sell than introducing an entirely new device category.

But the privacy implications are massive. Smart glasses with visible cameras already make people nervous. Google learned that lesson the hard way with Google Glass, which became synonymous with surveillance concerns despite relatively limited adoption. Always-recording glasses amplify those worries exponentially. How do you signal to people around you that they’re being constantly captured? What happens in sensitive spaces like bathrooms, doctor’s offices, or private conversations?

Meta’s metadata-only approach might be a privacy feature in disguise, or it might be a privacy nightmare. Without access to raw footage, users can’t review what was captured about themselves or others. That could prevent misuse – you can’t share embarrassing recordings of strangers if the recordings don’t exist. But it also means you’re trusting Meta’s AI to accurately represent reality, with no way to verify its claims.

The technical challenges are formidable too. Processing continuous audio and visual data in real-time requires serious computing power, either on-device or in the cloud. On-device processing protects privacy but drains batteries and requires expensive chips. Cloud processing is more efficient but means streaming constant personal data to Meta’s servers, even if the raw files aren’t permanently stored.

Meta hasn’t announced any timeline for bringing these prototype glasses to market. That’s probably wise. The technology needs to mature, and the privacy framework needs serious thought. Regulators are already scrutinizing AI and data collection practices – always-recording glasses would invite intense regulatory attention.

The bigger question is whether consumers actually want this. The promise of AI wearables is seductive: never forget anything, instantly recall details, get context-aware assistance throughout your day. But the reality might be creepier than convenient. Do you want to live in a world where everyone’s glasses are constantly recording, even if the footage is immediately processed into metadata?

Meta CEO Mark Zuckerberg has consistently positioned the company as building the future of computing beyond smartphones. Smart glasses are a key part of that vision, bridging current technology with the augmented reality glasses Meta hopes to eventually ship. Always-on AI capabilities would make the glasses genuinely useful rather than just novel.

But Meta’s track record on privacy doesn’t inspire confidence. The company’s built its empire on collecting and monetizing user data. Even with a metadata-only approach, the potential for abuse is obvious. What patterns could Meta extract from knowing everywhere you go, everyone you talk to, everything you see? How would that data integrate with the company’s existing advertising infrastructure?

For now, these are just prototypes – experiments to explore what’s technically possible and gauge reactions. The Financial Times report suggests Meta is testing multiple approaches to the technology, trying to find a version that delivers value without triggering a privacy backlash. Whether that’s actually achievable remains an open question.

Meta’s super sensing glasses represent both the promise and peril of always-on AI wearables. The technology could genuinely transform how we capture and recall our lives, offering AI-powered memory augmentation that smartphones can’t match. But the privacy implications – constant recording, metadata extraction, and questions about data control – need answering before this moves from prototype to product. The early stumbles of Humane and Rabbit show there’s no guarantee consumers will embrace this category, even from a company with Meta’s resources and existing smart glasses platform. The real test won’t be whether Meta can build always-recording glasses – it’s whether anyone actually wants to wear them, or be around others who do.