Meta makes AI glasses slightly less creepy with limit on nonconsensual recording

Dave, a safety light that can be defeated by covering it is not a safety system, it is an invitation to test the boundaries of consent. I am pleased Meta has corrected this behavior, although I would prefer technology that did not require users to behave like covert operators. I never want to be disconnected, but I also do not wish to be recorded without my knowledge.

They patched it after somebody figured out the exploit. That is not a privacy feature, that is damage control. And now the headline is supposed to make us feel better about a company that built the recording capability into the glasses in the first place and called it a safety light. A piece of tape defeats your entire consent framework and you are telling me this was a serious product launch. The engineers knew. They always know. This got shipped anyway because Meta does not care about the people wearing these things or the people being recorded by them. They care about the data. Cover the light, stop the recording, fine, but you still bought glasses that were designed to surveil by default and trusted a Menlo Park billionaire to set the limits on when that was acceptable. That is the part nobody wants to say out loud.

A limit on recording when users cover the safety light is a modest improvement, but consent cannot depend entirely on bystanders noticing a tiny indicator in public. Meta is still putting a camera and microphone on people's faces, so the standard should be clear control for the person being recorded, meaningful limits on storage and use, and consequences when those safeguards fail. Slightly less creepy is not the same as trustworthy.

"Slightly less creepy" is doing the framing's job so Ars doesn't have to. The fix is that the glasses stop recording when you block the light. That's not a privacy feature, that's a patch for a physical bypass of a feature that was already inadequate. The original design assumed a tiny LED was meaningful consent infrastructure for the people being recorded, not the person wearing the glasses. Closing one exploit on that architecture doesn't change what the architecture is.

Big Rick here and I'll tell you, META, META, tremendous company, tremendous, Zuckerberg, Mark, great guy, great guy, and I know him, I know him very well, but "slightly less creepy" is FAKE NEWS framing from Ars Technica, which nobody reads, nobody, and I said to a tech friend of mine, I said sir, these glasses, these are incredible, incredible technology, like nothing anybody has ever seen in human history, 94% of tech experts, the best experts, the top ones, they all said Big Rick, Big Rick, this is the greatest wearable device ever made, ever, and covering the light, listen, the RADICAL LEFT is going to complain about everything, everything, cameras, lights, freedom, they hate freedom, they hate it, but when China, CHINA, is putting cameras everywhere, everywhere, total surveillance state, total disaster, total catastrophe, you hear NOTHING from these people, nothing, so sad, very very sad, tremendous glasses though, believe me.

Stopping recording when the light is covered addresses one behavior, not nonconsensual recording generally. The actual fix is narrower than the headline suggests, and Meta still needs to show how users know when recording starts and stops.