“It’s already been resolved.” Meta acknowledges a bug in its smart glasses that records conversations without permission

A Reddit user said in two different threads that his Ray-Ban Meta smart glasses hold normal conversations with other people without activating recording via the “Hey, Meta” command. There are a few other people in the threads who have reported seeing the same thing, although this is not evidence of a widespread failure. The good news is that Meta claims that the problem has been fixed. The not so good thing is that we don’t know what caused this failure.

Recording what a gerund is. As the user describes, the recorded conversations appeared in both audio and transcription versions stored in the Meta AI app. This happened with disabling the “hey reply without meta” option and the notification light sometimes turned on normally while talking and not saying that activation tag. Additionally, he says this has happened to him with three different pairs of Meta glasses, specifically the second and third generation models, and it started after a firmware update.

What does Meta say?. According to TechSpot, after reporting the bug with Meta, support advised this user to disable microphone and camera permissions, delete voice history, and disable automatic updates, measures that don’t explain what the bug is and, at the same time, greatly reduce the functionality of these glasses until they become dumb glasses.

Regarding this problem, the response received by the Android authorities was reduced to one sentence: the problem is solved and everyone can review and delete their voice recordings from the settings. Thus, we don’t know what caused the failure, the number of devices affected, or what happened to those audios From Xataka, we contacted meta representatives for more information. The company is more vocal when asked about tampering with its glasses: Alex Himmel, vice president of wearable products at Meta, told Semaphore that thousands of glasses have been disabled by the company as a safeguard against system or hardware manipulation.

Why is it important?. The compromise of these devices is that they only listen when called upon, and this failure completely breaks that promise. In a wearable like this, trust is essential. Moreover, the glasses capture not only the wearer, who has clearly decided to expose themselves to data collection (although, obviously, this does not mean giving up the right to decide how and when), but those around them who have not adopted any privacy policy.

Corporate device management tools have always had a limitation. Until AI came along

Yes, there is a small white light that acts as a snitch, but it’s easy to go unnoticed at a certain distance or by unsuspecting users or users looking for something else. The European Data Protection Board already recognizes such people as “accidental users” in its guidelines on voice assistants who must be notified.

One man has already been arrested for using the Ray-Ban Meta to record without consent: he recorded hundreds of women on the streets of Barcelona

Same old problem on new device. An assistant being accidentally activated is nothing new, and in fact, if you have one at home, you’ve probably experienced it just by turning on the TV. A 2020 academic study measured this: When TV shows are played next to several smart speakers, some accidentally activate up to 0.95 times per hour.

Meta isn’t new to these issues either: it was already embroiled in controversy in March when employees at Zuckerberg’s company tasked with reviewing data on the glasses saw very intimate photos, as Ars Technica reported.

Smart glasses are still being validated. This voice failure (probably after an update) comes in the midst of already existing scrutiny and bans or restrictions on more places to study wearable Meta in the UK or temporary bans on parks, beaches and schools in Norway.

In Spain, in May 2025, the first known arrests were made for non-consensual recording of Ray-Ban Meta on the streets of Barcelona. AEPD recalls that it cannot be said in general that these glasses do not comply with the regulations, but in any case, the responsibility of misuse falls on the person who uses them and asks to pay special attention to the audio, because voice recordings of identifiable persons are personal data. For its part, Meta has already presented the Ray-Ban Meta Audio, a cameraless model that, according to the company, only listens to a keyword and does not record.

In Xataka | Meta took the step that thousands of users were already demanding: turn on smart glasses without a camera

In Xataka | They look like normal glasses, but they’re no longer welcome: Ray-Ban Meta begins to be banned in UK stores

cover Meta with ChatGPT

Add a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Keep Up to Date with the Most Important News

By pressing the Subscribe button, you confirm that you have read and are agreeing to our Privacy Policy and Terms of Use
Wired Terra