Ray-Ban Meta owners have a new privacy reason to revisit how their glasses are configured. Meta has updated the privacy policy for its AI glasses, giving the company broader ability to store and use data connected to the device and its AI features.
The change does not mean the glasses are constantly recording everything around the wearer. But it does change what happens when certain AI features are active, and it makes the default setup more important for anyone who uses the glasses around family, friends, coworkers, or strangers.
What changed for Ray-Ban Meta glasses
Meta emailed Ray-Ban Meta owners on Tuesday with a notice that AI features will now be enabled on the glasses by default, according to The Verge. With those features switched on, Meta's AI will analyze photos and videos taken with the glasses.
The update also covers voice data. Meta will store customers' voice recordings to improve its products, and the source article states that there is no option to opt out of that storage.
That default matters because many people use connected devices as soon as they come out of the box. If AI features are already active, a user may not pause to consider whether a photo, a video, or a voice interaction could become part of Meta's broader AI improvement process.
What the glasses do and do not record
The most important limit is that Ray-Ban Meta glasses are not described as constantly recording and storing everything nearby. The source article is clear on this point: the device stores speech only after the user says the "Hey Meta" wake word.
That distinction is significant, but it does not remove the privacy concern. A wake-word interaction is still a voice recording. A photo or video taken while AI features are active can still be analyzed. The privacy issue is not continuous surveillance in this case; it is what Meta can keep and use after a user chooses to interact with the glasses.
Meta's privacy notice on voice services for wearables says voice transcripts and recordings can be stored for "up to one year to help improve Meta's products." For a customer who does not want Meta to train its AI on their voice, the available step described in the source article is manual deletion of each recording from the Ray-Ban Meta companion app.
Why AI companies want this data
The source article connects Meta's change to a broader pattern in consumer AI products. Amazon recently changed its policy for Echo users, and as of last month, Amazon will run all Echo commands through the cloud. That removed the more privacy-friendly option to process voice data locally.
The reason is straightforward: voice recordings are useful training data for generative AI products. With more audio samples, Meta's AI can possibly improve at handling different accents, dialects, and speech patterns.
For companies building AI assistants, user interactions can be valuable because they reflect real-world language instead of carefully prepared examples. People speak in different rhythms, with different phrasing, and in different conditions. The source article says companies like Meta and Amazon are eager to gather these recordings because they can help improve AI systems.
But that same value creates the privacy tradeoff. The more useful a personal recording is for AI training, the more sensitive it may feel to the person who produced it, or to someone nearby who did not expect to become part of an AI data pipeline.
The privacy risk is not only about the owner
Ray-Ban Meta glasses raise a privacy question that extends beyond the buyer. If a user takes a photo of a loved one while AI features are enabled, that person's face may end up in Meta's training data, according to the example given in the source article.
That is a different kind of consent problem from a phone app setting. The person wearing the glasses may have agreed to the device's terms, but the person being photographed may not know what features are active or what data practices apply.
The same logic applies to casual video and voice interactions. The wearer controls the glasses, but the surrounding environment may include people who have no direct relationship with Meta's privacy settings. That makes the default status of AI features especially important.
Meta's use of user data is not new. The company already trains its Llama AI models on public posts that American users share on Facebook and Instagram. The Ray-Ban Meta update fits into that larger picture: content created by users can become material for improving AI products.
What owners should review now
The practical takeaway is simple: Ray-Ban Meta owners should inspect the settings tied to AI features and voice recordings in the companion app. The source article specifically notes that customers who do not want Meta to train AI on their voice must manually delete each recording from the app.
Based on the source article, owners should pay attention to these areas:
- whether AI features are enabled by default on their glasses
- how photos and videos are analyzed when those AI features are active
- which voice recordings have been stored after the "Hey Meta" wake word
- whether any recordings should be manually deleted from the Ray-Ban Meta companion app
The larger issue is not that AI glasses are unusable. It is that wearable AI devices blend into ordinary life, which can make data collection feel less visible than it is. When a camera, microphone, and AI assistant sit on a person's face, default settings become part of everyday privacy.
For Ray-Ban Meta owners, the update is a reminder that convenience and AI improvement often depend on data. The question is whether users understand what is being stored, how long it can be kept, and what they need to delete themselves if they do not want their voice used that way.