The floodgates for artificial intelligence’s interference (or improvements, depending on your point of view) in digital photography have opened up since Apple and Google made it mainstream, with advancements such as Portrait Mode and Night Sight. Adobe took it another direction when it added an A.I.-powered “enhance” feature that enabled the enlarging of photos beyond what’s naturally possible. But is this making photography less real than it should be?
Coming at you from Drone Film Guide is a hot take on the subject, with an emphasis on the “should be” part of that last line. While I’ve long been a skeptic of computational imaging, I can’t help but enjoy some of the modern benefits afforded by the tech. For instance, the only copy of this photo of my grandparents that exists, as far as I know, is a measly 0.8 megapixels. But through the magic of Adobe’s AI Enhance, I now have a family photo that is 3.1 megapixels. Not huge, but certainly enough to make a decent size enlargement.
Artificial Intelligence helped me take this photo of my grandparents from the 0.8 megapixels it was scanned in at to 3.1 megapixels. Not bad.
So, I can see the argument for AI imaging technology, especially when it comes to bringing family memories kicking and screaming into 2021 resolutions (or maybe 2000s resolution in this case). That said, as AI’s introducing new pixels that didn’t exist before into the mix to produce the enlargement, is AI something that really has a place when accurate archiving is the goal?
But what about those other replacements, such as the popular AI sky replacements that can be done with Luminar? Admittedly, those same modifications have always been possible with the likes of Adobe Photoshop, Stewart argues. It’s just that AI makes the whole endeavor a bit easier and less time-consuming. If the goal is a compelling image and not necessarily reality, then why not make the process of getting to that image a bit faster?
The whole umbrella of computational imaging tech and after-the-shot AI manipulation includes a lot of tech that just isn’t understood by photographers, thus complicating the “legitimacy” issue around photography, but does that matter outside the realm of photojournalism?
What are your thoughts on AI in photography? Leave your thoughts in the comments below.
The discussion is similar to how much to do in Photoshop, my idea is "Don't lose the skill to get as much right in your camera as you can before resorting to it"
Listen to this guy. I know nobody here likes Pentax, but listen to the photographer: https://youtu.be/8Mq1MPMyw5U