Do You Own Ray-Ban Meta Glasses? You Should Check Your Privacy Settings.

Meta recently updated the privacy policy for its Ray-Ban Meta smart glasses, enabling AI features by default. This means that images and videos captured will be analyzed, and voice recordings stored to help improve the company’s AI models. Contrary to popular belief, the device does not record constantly, but only after being triggered by saying “Hey Meta.” Although voice recordings can be retained for up to a year and are used to train Meta’s AI without a direct opt-out option, users can manually delete their recordings via the companion app. This change—similar to a recent policy update by Amazon for Echo—raises privacy concerns, as the data collected feeds into Meta’s generative AI products.

Here are the key details regarding the Ray-Ban Meta glasses privacy policy changes, based on previously provided information:

• Meta has updated its privacy policy for Ray-Ban Meta smart glasses, expanding its ability to store and use user data to train AI models.
• AI features are now enabled by default on the glasses.
• When certain AI features are active, Meta’s AI will analyze photos and videos taken with the glasses.
• Voice recordings will also be stored by Meta with the purpose of improving its products. There is no option to disable this recording.
• It’s important to note that the glasses do not constantly record everything around the user. The device only stores speech after the wake phrase “Hey Meta” is spoken.
• Voice recordings and transcripts may be stored for up to one year to help improve Meta’s AI products. If a user does not want Meta to use their voice recordings for AI training, they must manually delete each recording through the Ray-Ban Meta companion app.
• The purpose of storing this data is to supply useful training material for generative AI models. A wider range of audio recordings allows Meta’s AI to better understand various accents, dialects, and speech patterns.

However, this AI improvement comes at the cost of user privacy. A user may not realize that by using the glasses to photograph someone, that person’s face could potentially be included in Meta’s AI training data. AI models require massive amounts of content, and companies benefit from training on the data users already generate.

This shift is similar to Amazon’s recent change to Echo’s processing, which now sends all commands to the cloud by default, removing the option for local processing.

It’s worth noting that Meta’s use of user data is not new. The company is already training its Llama AI models using public posts shared by U.S. users on Facebook and Instagram.

In summary:

• AI is now enabled by default.
• Media captured by the glasses is analyzed when AI is active.
• Voice recordings triggered by “Hey Meta” are stored and used to train Meta’s AI.
• The only option for users concerned about this is to manually delete recordings via the app.

These changes aim to improve AI performance but raise valid privacy concerns for users.