I don't mean to go too much against the narrative in this thread, but aren't cameras on glasses necessary to achieve AR? Are people concerned about these glasses just because it's Facebook, or is there broader opposition to AR technology?
Almost nobody will feel violated if a blind person has camera-based prosthetic eyesight... unless it's also recording for others to gawk at later.
I think the key privacy distinction isn't whether there's a device using a camera to augment vision. It's whether the glasses are recording for publication via Facebook at one extreme, versus just processing locally to perform AR and discarding the stream at the other extreme.
Where personal information may be involved for functionality, ideally it should use sophisticated privacy-preserving data queries, for example looking up faces and gaits without revealing to the cloud who is being looked up (technically challenging but undoubtedly possible to build).
Even if the video is not published, but it's streamed to Facebook for, say, demographic manipulation analysis, race estimation, genetic healthcare analytics (affecting other individual people's insurance premiums based on what it sees), juicy gaffes that are "recommended" by ML to be posted as probable high-engagement rage-tweets, customer service profiling, that sort thing, it's going to make being around a person wearing these things high-anxiety for a lot of people. The list of possible dark applications is endless. For good reason: It has the power to ruin someone's life, and certainly end or prevent a career, should someone else wish to do that.
Just recording the video (or even audio) without publishing it, purely for the owner's use, could be a problem if it's pervasive. It might be used later against you when you thought you were ok being relaxed and informal without paying heightened attention to your every action. Perhaps someone you thought was a friend decided later to use their private recording as evidence for a character assassination campaign or whatever. Imagine if everything you said, every gesture, every eyebrow movement might be plucked out of context and turned into an outrage tweet at any time after it happened. Panopticon is a problem. People need informal contexts where it's ok to not be on heightened alert at least some of the time.
Its worse. They have a consistent track record for asking for trust and then breaking it in most abusive way possible, then issuing a non-apology only after getting caught and then repeating same process all-over again. Oh, I forgot mentioning FB getting rich in the process, so it somehow becomes justified and has zero chance for change.
I think it's fair to say that Facebook grew too quickly to mitigate the issues inherent to social networks at scale. I don't think that makes them abusive or evil, just naive.
https://www.theguardian.com/technology/2014/jun/30/facebook-...
Overall, I would guess that there is no track record of trust because expectations remain unclear. Concepts of privacy, data usage, and such didn't exist before Facebook and others. I suppose I am not eager to assume malintent outside what is standard for any business operating for the sake of profit.
They did exist before Facebook.
But not before "others", which include every company and person who ever existed. Not sure why you added "and others" to that sentence.