Meta Acknowledges Ray-Ban Glasses Recorded Without Wake Word After Firmware Update

Meta has confirmed that its Ray-Ban smart glasses were inadvertently recording users’ conversations without the required “Hey Meta” wake phrase, a malfunction the company says is now fixed but refuses to explain in technical detail.

The issue surfaced after a firmware update pushed to both second- and third-generation models, according to multiple users who reported the behavior on Reddit and to Meta support. Affected owners found full audio recordings and transcripts of private conversations appearing in the Meta AI app despite having the “Respond without Hey Meta” setting disabled. In several cases, the glasses’ white privacy indicator light — designed to illuminate only during active recording — activated during normal speech with no wake word spoken.

Meta’s initial support response instructed users to disable microphone and camera permissions, turn off wear detection, delete their voice history, and disable automatic updates — steps that effectively cripple the device’s core functionality. One user, who documented the issue across three separate pairs of glasses, said they are consulting an attorney about a potential class-action suit.

“This issue has been resolved,” a Meta spokesperson told Android Authority. “People can manage, review and delete voice logs of their interactions with Meta AI via settings at any time.” The company declined to specify what triggered the recordings, whether the firmware update introduced a regression, or how many devices were affected.

The incident compounds mounting regulatory and social scrutiny of camera-equipped wearables. Norway’s data protection authority is evaluating a temporary ban on smart glasses in parks, beaches, and schools. Courthouses across the United States have begun prohibiting them outright. In March, a report revealed that contractors in Nairobi were paid to review hours of raw footage captured through Meta’s “live AI” feature, including intimate moments in bathrooms and bedrooms, to train computer vision models.

Privacy advocates argue the latest failure undermines the consent model Meta relies on: a visible light and a wake phrase. Neither safeguard functioned as designed. The company has not said whether it will issue a post-mortem, notify affected users directly, or change its firmware testing protocols.

As of Friday, Meta has not released a new firmware version number associated with the fix, nor updated its public changelog.