A hot potato: If there’s one product that really doesn’t need any more negative stories, it’s Meta’s smart glasses, but here we are. Some users found that the wearables recorded their voices without the wake phrase, an issue the company says has since been fixed.
A Redditor recently wrote that he noticed something concerning about Meta’s Ray-Ban glasses: they recorded conversations while talking to other people, even though the wake-up phrase “Hey Meta” wasn’t used.
The conversations appeared in the Meta AI app as saved audio recordings and text transcripts. It also notes that the “Reply without Hey Meta” setting was disabled and that the white notification light would sometimes activate when the person was speaking normally.
This might have been an issue with a single pair of Meta glasses, but the same behavior occurred with three different pairs, including 2nd and 3rd generation models. Most tellingly, it occurred after a firmware update was rolled out to the devices.
Meta Support was contacted about the strange behavior, but their response was to recommend disabling microphone and camera permissions, disabling wear detection, deleting voice history, and disabling automatic updates.
The user complained that these steps not only broke the core functionality of the product, but also did not explain or fix the problem.
The post adds that the user is consulting with an attorney about the possibility of starting a class action lawsuit if other customers discover the same unwanted voice recording.
Several replies to the post confirm that other Meta Ray-Ban users had the same problem: the glasses were randomly recording conversations without the wake word.
Meta has since responded to the Reddit post, but never provided any information about what exactly happened and why. “This issue has been resolved. People can manage, review, and delete voice logs of their interactions with Meta AI at any time through Settings,” the company told Android Authority.
Wearing smart glasses in public is now becoming as socially acceptable as walking around with a swastika badge. Norway is considering temporarily banning them in parks, beaches and schools, and they will be banned in courtrooms in the United States and elsewhere.
This is not the first data protection nightmare related to Metas smart glasses. In March, contractors in Nairobi said they were paid to study hours of raw footage shot using Meta’s “live AI” feature, analyzing everything from simple interactions to some very intimate moments. This data labeling allows Meta to train its computer vision systems, although the people recorded going to the bathroom, getting dressed, and having sex may not care too much about the reasons why.
