I have been thinking about this a lot lately: what does it mean to build technology that only works if everyone around you assumes you are a good person?
That question sits at the heart of the AI wearables debate, and it is not a comfortable one. The Verge’s recent deep dive into the cultural backlash around Meta’s smart glasses puts it plainly. These devices are discreet by design. That is the feature. And that is exactly the problem.
The Intent Gap Is a Design Flaw
Let us be honest about what is happening here. Meta glasses, AI recording rings like Vocci, and a growing wave of ambient wearables are being built around a core assumption: that the person wearing them has benign intent. The LED privacy lights are easy to miss. Battery limits prevent true 24/7 surveillance. Most users genuinely just want hands-free video of their dog or automatic meeting transcripts.
But design cannot legislate intent. And as someone who follows the AI space closely, I think this is a fundamental product failure that is being papered over with good vibes and celebrity endorsements.
The AirTag comparison is instructive. Apple got hammered by domestic abuse advocates for building a tracking device without robust enough anti-stalking safeguards. They iterated. They added alerts, refined detection windows, and responded to real-world harm. It is not perfect, but there is a feedback loop between misuse and design response.
With AI wearables, that loop does not exist yet. LED lights can be taped over. Recording can be triggered silently. And the moment you hand someone a piece of hardware, you have lost control of how it gets used. Telling people to just trust the wearer is not a privacy strategy. It is wishful thinking.
What This Means for Developers
If you are building in this space, or adjacent to it, pay attention. The backlash to Meta’s glasses is not just Twitter noise. It signals a real and growing cultural resistance to ambient recording devices. And that resistance will eventually translate into legislation, venue bans, and platform restrictions.
We are already seeing it. Private venues are turning away Meta glasses wearers. Zenni Optical is selling anti-facial recognition lenses. France and other jurisdictions are actively looking at wearable surveillance regulations. This is the trajectory.
For developers building on wearable platforms, a few things worth internalizing:
- Consent flows need to be hardware-level, not just software-level. An app asking for recording permission is not enough when the device itself is invisible to bystanders. Think about physical, unmissable indicators.
- Modular design deserves a serious second look. Yes, a removable camera module is less elegant. But elegance that erodes public trust is a short-term trade. Xreal’s modular camera attachment is an early example worth studying.
- Audit trails matter. If your wearable app records anything, build in transparent logs that users and, where legally required, subjects can access. This is table stakes in enterprise; it should be in consumer too.
The deeper issue is that the privacy implications of these devices are being treated as a PR problem rather than an engineering one. That is backwards. No amount of good branding survives a viral clip of someone getting secretly recorded in a gym locker room.
Good Intentions Do Not Scale
Here is the part that I keep coming back to. The reviewer in The Verge piece describes testing a recording ring on her spouse and colleagues without their knowledge, purely for evaluation purposes, always disclosing afterward. She is a professional with an ethics policy. She thought carefully about every step.
Most people will not do that. Not because they are bad people, but because friction is the enemy of habit. If a device makes it easy to record and hard to remember to disclose, the average user will take the path of least resistance. Every time. That is not a moral failing. That is just how humans interact with convenient tools.
Scaling good intentions across millions of users is an unsolved problem. And it is one that the industry is not treating with nearly enough seriousness. Building ambient AI hardware without robust, user-proof consent mechanisms and then pointing at the minority of bad actors as if they are the only concern misses the larger structural issue entirely.
The wearable surveillance state is not coming because of a few creeps. It is coming because well-intentioned people with genuinely useful tools are normalizing a world where recording everything, all the time, is just… what you do.
The more interesting question is not whether regulators will eventually step in, but whether any company in this space will choose to lead on this before they are forced to.