If you have Ray-Ban Meta glasses, you must check your privacy settings

Meta has to update The privacy policy for AI, Ray-Ban Meta, giving the technology giant more strength on the data it can store and use to train artificial intelligence models.

The company sent an email to the Ray-Ban Meta owners on Tuesday with a notification that the features of artificial intelligence will now be enabled on the glasses virtual, according to freedom. This means that Meta’s AI will analyze images and videos that are taken with glasses while playing some of the artificial intelligence features. Meta will also store audio recordings for customers to improve their products, without the option to cancel the subscription.

To be clear, the Ray-Ban Meta glasses do not constantly register and store everything around their wearer. The device stores only the speech that the user says after the word “Hey Meta”.

Dead Privacy notice on audio services For wearable devices, it says that texts and sound recordings can be stored for up to a year to help improve the Meta products. “If the customer does not want Meta to train AI on their voice, they will have to delete all records from the Ray-Ban Meta Companion manually.

Change in terms of conditions is similar to the Amazon Change the last policy It affects echo users. As of last month, Amazon will play all Echo orders across the cloud, which removes the most suitable option for domest audio data processing.

Companies like Meta and Amazon are giving up these waves of audio recordings because they are useful training data for artificial intelligence products. With a wide range of audio recordings, Meta’s AI can do better work in treating different accents, dialects and styles.

But improving artificial intelligence comes at the account of the user’s privacy. The user may not understand that if they use Ray-Ban Meta glasses outside the box to photograph a member of his family, this person may find his way to Meta training data, for example. Artificial intelligence models behind these products require obscene quantities of content, and they benefit companies to train artificial intelligence on the data that their users already produce.

The user data is not new. Indeed, Meta is trained by Llama Ai models on General jobs American users share Facebook and Instagram.