Modern versions of Chrome and Edge have features that try to guess what an image contains using AI techniques. Same with iOs and Mac OS. Modern iPhones even try to recognize controls on the screen, but my phone is too old for that to work. These solutions are still in their infancy, though, and therefore far from perfect. They might help us jump through a small accessibility hurdle once in a while, but without serious accessibility work from developers, we're doomed anyway.
I'll one-up this a bit: a blind woman called Kristy Viers demonstrates on YouTube how she uses various things, among them iPhone with the onscreen six-finger Braille keyboard. But also she uses the ‘magnifier’ app to tell her what is there around her, i.e. what the camera sees: https://youtube.com/watch?v=8CAafjodkyE
(The videos are filmed by her boyfriend, in case someone has to wonder.)
I realize that the vid is not too informative to a blind person, other than a vague demo. IDK where to find more info on the setup, since I'm not a user of iPhones myself. Comments on YouTube say that the magnifier app is in the accessibility options, that it adds a home screen app when enabled, and that it works at least on iPhone 10 since that's what Kristy uses.
I would also guess that such usage requires persistent mobile connection and consumes considerable traffic and battery—but again, with modern tech magic I won't bet against it working locally on the phone.
On Android, the feature called ‘Lens’ in the Google's camera app can describe some things, more useful to seeing people IMO: e.g. the model of a car—basically the same as Google's image search does, these days. Also purportedly can read and translate text signs. But IDK how far that goes, never heard of it doing stuff like those descriptions by the iPhone magnifier.
I checked the magnifier app on my iPhone 6s and couldn’t find anything related to narrating what it sees.
It’s very possible that recognition happens locally, Apple has been making efforts in this direction for other functions (photos) but maybe the 6s is just too old and can’t do it.
As far as I understand, the narration is handled by the ‘voiceover’ feature, which is a generic accessibility feature, precisely the one that reads screen contents and controls. In the demo, Kristy taps the screen each time to invoke the voiceover description.
iPhone 6S shipped with iOS 9 but is upgradable to the current 15.6. Though features may indeed depend on the particular phone, i.e. the hardware.
Also, apparently Voiceover changes the way the phone should be controlled, to its own set of gestures—so it's not a feature to use just with the camera. However, it sounds to me like blind users would find it helpful anyway.
