Originally posted by sbh What I found interesting in that announcement is their new Apple RAW format which includes both the original pixel data and the processing information from their AI (or however you may call it). It basically includes all the processing information how to achieve the end result as Apple thinks to be a good result. I think any other camera with a processor could do that kind of stuff as well, or not?
Usually processing information is related to implemented processing functions, i.e. concrete software. Since there's no standard for such implementations there's a dependency to Apple algorithms.
I have an old Canon PowerShot G10 and in the beginning I used the Canon raw converter Canon Digital Photo Professional to convert the CR2 files. If I remember right, the processing information was written into the CR2 files. I guess the same is possible for Adobe created DNG files. This processing information yields to designated results if you use the original manufacturer Raw converter.
The AI (Artificial Intelligence) thing in my opinion is widely discovered for marketing purpose. There may analytical algorithms be used - that's not really AI specific! One important property of real AI systems is that they learn by themselves. How would this be mapped into raw files? ... I won't go further here but like to say.
So what we have is a very proprietary raw format for a camera with a ~1/1.7" sensor and 3 lenses. The longest equivalent 52/65mm called tele that in reality is more a normal lens, plus a very cluttered description of optical zoom options, even the real focal lengths of the three lenses are not mentioned (have a look into the technical details). Apple is a selling artist ... and I like their computer gear. ;-)
To answer your closing question, yes - every digital camera could do such with the appropriate processor(s) and firmware / software. Digital cameras are computers with an integrated and/or attachable optical system
Sorry for this sober analysis and getting off topic.