In late April Meta made changes, effective immediately, to the Ray-Ban Meta “smart glasses” privacy policy, which appears designed to turn the device into a surveillance machine for training AI models.
In a message sent to users, Meta said that its “AI on the glasses” – that is, some of the settings – are changing.
The giant’s explanation is that this is allegedly necessary to use Meta AI “more easily” – but also, “help us make product improvements.”
The policy update rests on “opt-outs”: from now on, the Meta AI with camera use is always enabled on the glasses, unless the user goes to the trouble of turning off the “Hey Meta” in the settings.
This is the activation phrase for Meta’s AI assistant. The second change regards the way Meta stores voice recordings from Meta AI users – now, those are kept by default in the cloud.
The reason the company gives for this is “improving” either Meta AI or, “other Meta products.”
The option to disable this behavior is now gone. Once again users are made to jump through additional hoops, and that’s the tried and tested Big Tech way of steering their behavior and interaction with apps and services in the desired, by Big Tech, way.
In this case, Meta AI users will have to go to the settings and manually delete their voice recordings.
When making these decisions, companies like Meta effectively “dumb down” their “smart” products (by removing voice interaction with the assistant, reducing automated usability to manual deletion…).
And that’s on top of rubbing the wrong way those who are uncomfortable with increasingly privacy-intrusive mechanics behind the said products and services.
In addition to selling what’s obviously not an “improved privacy experience,” Meta and its ilk still insist that obscuring what’s happening under the hood means a better (“easier”) user experience.
The most grim and negative scenarios as to why any of this is being done, or how it could be (ab)used in the future, aside – the obvious intent is to take the exploitation of user data to another level, at this junction, to ensure massive datasets are available for AI model training.
The notification users received about the latest policy changes adds a little insult to injury when it concludes by reiterating, “You’re still in control.”
“In control” to turn off “Hey Meta” and manually delete Meta AI interactions, that is.
