On Wednesday, The Information reported that Meta is working on facial recognition for the company’s Ray-Ban glasses. This sort of technology—combining facial recognition with a camera feed—is something that big tech including Meta has been able to technically pull off, but has previously decided to not release. There are serious, inherent risks with the idea of anyone being able to instantly know the real identity of anyone who just happens to walk past their camera feed, be that in a pair of glasses or other sort of camera.
The move is an obvious about-face from Meta. It’s also interesting to me because Meta’s PR chewed my ass off when I dared to report in October that a pair of students took Meta’s Ray-Ban glasses and combined them with off-the-shelf facial recognition technology. That tool, which the students called I-XRAY, captured a person’s face, ran it through an easy to access facial recognition service called Pimeyes, then went a step further and pulled up information about the subject from random the web, including their home address and phone number.
When I contacted Meta for comment for that story, Dave Arnold, a spokesperson for the company, said in an email he had one question for me. “That Pimeyes facial recognition technology could be used with ANY camera, correct? In other words, this isn’t something that only is possible because of Meta Ray-Bans? If so, I think that’s an important point to note in the piece,” he wrote.
This is true. But entirely misses the point of why the students created the tool with Meta’s Ray-Ban glasses. They said themselves in a demonstration video they identified dozens of people without their knowledge. You do that by wearing a pair of glasses that look like any other. Meta’s Ray-Ban’s do have a light that turns on when it’s recording, but according to the new report, Meta is questioning whether new versions of its glasses need this.
I replied to Arnold spelling this out a bit. “Yes, it could theoretically work with other cameras. That being said, they didn’t stick a GoPro onto their body—which would undermine the surreptitious nature of it. They used smart glasses, which Meta makes,” I wrote.
We published the article under the headline Someone Put Facial Recognition Tech onto Meta’s Smart Glasses to Instantly Dox Strangers. Arnold had an issue with that.
“I saw the article, and respectfully, I think the headline as written could be misleading to readers. It says someone ‘put facial recognition tech ONTO Meta’s Smart Glasses’, which sounds like someone’s hacked the device, or installed facial recognition device on the glasses themselves, when in reality, they are live-streaming from the glasses to Instagram and then they use a program to monitor the stream. I think that is an important distinction. Is it possible to update the headline to reflect the nuance here?”
I didn’t update the piece or reply because this is a distinction without any real meaning. Into the glasses, onto the glasses, with, glued to, whatever. The impact is what matters, and the project was done on Meta’s glasses.
Arnold left it at that for the time being. We then recorded a podcast episode about the students’ tool, and Arnold had some more complaints. He wanted to talk about the headline of the podcast, which was “The Smart Glasses That Dox Strangers.”
Arnold wrote:
You say in the podcast that the glasses don’t have facial recognition capabilities, and you’ve previously acknowledged that this could be done with any camera/recording device, but a headline saying ‘smart glasses that dox strangers’ clearly makes it sound like this is an issue that is specific to the glasses, or that the facial recognition was executed on the glasses themselves vs. the reality, which is that this was all run by a program on their laptop. This is despite the fact that the students themselves have said publicly: ‘We do not want this to be a criticism of their product at all, and we just had them on hand—this could have been done on a phone camera.’ I realize we may not agree on everything here, but surely you can appreciate how headlines like this are misleading for readers.
Joseph Cox
No, I don’t appreciate that. I also don’t care if the students didn’t intend their project to be a criticism of Meta’s smartglasses or not. The fact is, they built a doxing tool with Meta’s glasses, tested it on multiple unsuspecting members of the public without consent, and specifically chose Meta’s glasses because they allow a user to scan a face in a stealthier manner than jamming a phone’s camera in someone’s face, for example. The intellectual dishonesty at the expense of protecting the brand at all costs here is embarrassing.
I replied: “We think the headline is fair and accurate.”
I followed up with Arnold on Thursday to ask for comment for this piece too. I asked, does his earlier stance that this is not specifically a Meta problem, but a broader one, still stand? And what was their statement on The Information report?
Arnold wrote back: “404 Media’s previous reporting misleadingly implied that the students’ experiment was uniquely possible with Ray-Ban Meta glasses, despite the fact that the students themselves acknowledged ‘this could have been done on a phone camera.’ Our objection to this still stands.”