Meta is facing fresh privacy concerns after several Ray-Ban Meta smart glasses owners reported discovering audio recordings and transcripts of conversations they never knowingly asked their devices to capture.

The issue first surfaced in a Reddit discussion, spotted by Android Authority, where one user claimed their smart glasses had been recording conversations without the usual “Hey Meta” voice command or manual activation of the built-in AI assistant.

According to the user, the unexpected recordings appeared across three different pairs of Ray-Ban Meta glasses, including second- and third-generation models. Several other owners subsequently reported experiencing similar behaviour, raising questions about whether the problem was more widespread.

Meta has since acknowledged the issue in a statement to Android Authority, saying it has been resolved. However, the company provided little additional information about what caused the recordings, how many users were affected or what steps were taken to address the problem.

Under normal circumstances, Ray-Ban Meta glasses should only capture voice interactions when users activate the AI assistant using the “Hey Meta” wake phrase or a physical button on the device.

However, Meta’s own privacy documentation acknowledges that voice services can sometimes be activated inadvertently. It also states that certain voice recordings may be reviewed by company personnel to help improve its voice services.

For users concerned about whether their glasses have captured conversations without permission, the Meta AI smartphone app provides a way to review stored voice activity.

Owners can navigate to Glasses > Device settings, select their connected glasses and then open Glasses privacy > Voice activity log to inspect and delete saved recordings.

While recordings can be removed through the app, users cannot completely disable audio capture while retaining all voice assistant functionality. Some Reddit users have responded by switching off voice activation altogether, although doing so removes one of the glasses’ main hands-free features.

The reports add to existing scrutiny surrounding the privacy implications of camera-equipped smart glasses, particularly as Meta continues expanding the capabilities of its wearable AI devices.

Although Meta maintains that the problem has been fixed, the incident highlights the importance of ensuring that users have clear control over when their smart glasses are listening and recording, especially when conversations involving other people could be captured without their knowledge.