- Parallel proposal to disable the recording indicator light has reignited a fierce debate over privacy, consent, and the unsettling prospect of always-on covert surveillance.
- Under the “super sensing” concept, the glasses would continuously sample the wearer’s environment: capturing snippets of audio and snapping periodic images across hours of daily activity.
Meta is developing a new generation of AI-powered smart glasses that would passively capture audio and images throughout the day — building a comprehensive memory of the wearer’s life. But a parallel proposal to disable the recording indicator light has reignited a fierce debate over privacy, consent, and the unsettling prospect of always-on covert surveillance.
The plans were first reported by the Financial Times, which cited multiple people familiar with the project. At the centre of the effort is an internal feature known as “super sensing,” a mode that would mark a decisive shift in the philosophy of wearable AI — from deliberate, user-initiated capture toward ambient, background recording that simply never stops.
A memory that never forgets
Under the “super sensing” concept, the glasses would continuously sample the wearer’s environment: capturing snippets of audio and snapping periodic images across hours of daily activity.
The AI would then build a searchable index of a person’s day, letting users ask questions like where they left their keys, which colleague they spoke to before lunch, or what was discussed in a meeting they barely remember.
Raw audio and images, according to one proposal being discussed internally, would not be permanently stored or made directly accessible to users. Instead, the system would extract metadata — structured information about what was seen or heard — so the AI can answer questions while sidestepping the need to retain the original recordings.
Meta is also reportedly weighing whether such data could serve as training fuel for its broader AI model development.
The concept pushes well beyond the current Live AI mode, which already allows users to ask real-time questions about their surroundings. Super sensing would keep the glasses “aware” over much longer stretches, operating in the background rather than requiring the user to manually trigger each interaction. It turns the glasses from a tool you ask into a companion that already knows.
The privacy light problem
If the super sensing concept is ambitious, it is the accompanying discussion around the recording light that has provoked the sharpest reaction.
Meta’s current Ray-Ban and Oakley smart glasses feature a white capture LED that illuminates whenever photos or videos are being taken — a hard-wired signal to anyone nearby that they are being recorded.
The company has repeatedly framed this light as a cornerstone of its privacy architecture, and as recently as this week, Meta announced an update that disables the camera entirely if the glasses detect that the LED has been blocked, covered, or physically tampered with.
“Beginning with our second generation of glasses, the camera is automatically disabled if we detect that the capture LED has been blocked. No photos or videos can be taken until we detect that the light is unblocked,” Meta stated in a FAQ published Tuesday.
The company added that it is “continuously improving its ability to detect tampering,” and that the glasses will now also shut down the camera if the LED has been “physically tampered with or destroyed”.
And yet, according to the FT report, Meta has internally considered whether the super sensing feature should operate without triggering that same light at all. The tension is hard to miss: Meta is publicly hardening the privacy LED as a non-negotiable safeguard while privately entertaining a future where it may not apply.
Critics have been quick to pounce on the contradiction. If the glasses are passively capturing images and audio across an entire day — in offices, on public transit, in homes — and no light alerts the people being captured, the distinction between a smart assistant and a surveillance device becomes uncomfortably thin.
The “glasshole” problem returns
Meta’s privacy challenge is not hypothetical. The current generation of glasses has already been branded “pervert glasses” in some quarters, and their wearers “glassholes,” following reports of users filming women in public without consent.
Wired documented influencers using the glasses to record themselves approaching strangers, raising alarms that wearable AI was normalising surreptitious recording as a cultural default.
The backlash has not stopped celebrity endorsements — Kylie Jenner has been among those promoting the devices — but it has hardened the battle lines. For every user who sees the glasses as a hands-free convenience, there is a privacy advocate who sees a tool purpose-built for exploitation.
The stakes are amplified by the fact that some of the new features, including super sensing, could reportedly arrive through a software update rather than requiring customers to buy new hardware. That means millions of glasses already in circulation could theoretically gain always-on recording capabilities overnight — a prospect that would radically change the social contract under which those devices were originally sold.
This is not the first time in recent months that Meta’s glasses ambitions have drawn scrutiny. In June, two security researchers discovered that Meta had built a complete facial recognition system into its smart glasses companion app — code that could theoretically be activated through a software update, turning the glasses into a real-time identification tool.
Once the story was widely covered, the company reportedly deleted the code.
The pattern is difficult to ignore: Meta builds the capability, tests the boundaries of public tolerance, and retreats — or doesn’t — based on the reaction.
Mark Zuckerberg has been unambiguous about his ambition to make AI glasses the next major computing platform, and Meta has invested billions in pursuit of that vision. But the super sensing leak crystallises a fundamental question that the company has yet to answer convincingly: how much privacy is the public expected to surrender so that someone can find their car keys?
Meta declined to comment on internal prototypes, telling the FT that it does not discuss projects under development. The company has not issued a public statement addressing the super sensing reports.
