Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Is the use of AI okay in the MR photography threads?

  • No AI at all in any thread on the MR photography forums

    Votes: 50 41.7%
  • AI for enhancement only (canvas extension, generative fill for small areas)

    Votes: 50 41.7%
  • AI for larger areas of the canvas, but you started out with a base photo

    Votes: 4 3.3%
  • If AI is used, it must be disclosed at the time of posting

    Votes: 61 50.8%
  • AI on a case by case basis (please explain)

    Votes: 6 5.0%
  • AI okay for a sample photo if the photographer does not have something relevant in their archives

    Votes: 5 4.2%

  • Total voters
    120
I don't have a strong option on if AI is okay in the MR photography threads. But it does remind me of similar discussions on the use of photoshop many years ago. Some may remember the National Geographic cover were the pyramid was moved. I personally find AI in post processing of my own photos to be unacceptable. It's unacceptable because I do enjoy spending hours working on an image in photoshop. When I started my career in photography I worked in a darkroom printing photographs. What made a good photograph and a good photo printer was the ability to manipulate the light with dodging and burning. For me photoshop is just that next step. So for others AI will be the next step for them. But for me I find no joy in making an image with AI. One area when I am open to AI is computational photography where AI is helping the photographer create an image.
 
Least we forget the elephant in the room, Apple AI usage in their iPhone photos workflow

b0354e53e3b04e05868398abe315cdb0.png

33fa1ad0178f0435c07f5da95283fda5.png
I think it would literally be impossible for Apple to get great images out of iPhones without some sort of AI capabilities given sensor size, optics (which are honestly much better these days), and the physics of those things.

Honestly, we throw around "AI" a lot without really taking the time to define what it really means. I take it as a machine making some decisions for me based on "recognition training" - for example <fill-in-the-blank> auto-focus (eyes, people, cats (cats!)) and I'd wager that we all rely on this every time we pick up a camera. There's also generative artificial intelligence where it "fills in or removes" content. You can do that by hand or you can have your assistant (machine) do it for you.

A completely machine generated image based on a prompt - "create an image of a dock in a lake in Switzerland" - isn't photography. Literally.

Or is it :D ? I don't think it is, personally.

I define photography (historical precedent and literal definition notwithstanding) as capturing photons through optics on recording media such as film or a digital sensor. A generative AI image is made up of content that happens to have been captured as photons through optics on media, for the most part by lots of other people. I'm not even getting into the concept of ethics. For me, the image needs to be sourced by one human with a photon recording device, with a photon recording piece of media and where that human has had to make decisions - including composition, media (film, digital), lens, aperture, shutter speed, ISO (and digital equivalent) and so forth. That one human should be the one posting the image in these forums :cool: . Post processing is cool. It's the game of photography. The human may decide to swap a sky - it's been done forever - or dodge, burn, compose multiple images, and so forth, but the key is "human" and who's making the decisions.
 
I think this is an excellent discussion to have. I feel like a photo should at least start with something captured by a camera; otherwise, it leans more into digital art than photography. That said, I don’t see an issue with using editing tools (whether traditional or AI-based) as long as the photographer took the original image. Perhaps it comes down to transparency: if something is heavily AI-generated, it should probably be disclosed so everyone’s on the same page.
 
So from what I can see in this thread no one seems to advocate that 100% AI images with no camera involved is photography. That is my view as well. Have a separate sub forum for digital art (or whatever you want to call it), but it’s not photography.

Editing is more subjective. I don’t think you can really find a happy level where we will all agree (or how you would police it!).

For me the more I can get right in the camera the better. Less time in front of a screen is after all one of the reasons I took up photography!
 
I define photography (historical precedent and literal definition notwithstanding) as capturing photons through optics on recording media such as film or a digital sensor. A generative AI image is made up of content that happens to have been captured as photons through optics on media, for the most part by lots of other people. I'm not even getting into the concept of ethics. For me, the image needs to be sourced by one human with a photon recording device, with a photon recording piece of media and where that human has had to make decisions - including composition, media (film, digital), lens, aperture, shutter speed, ISO (and digital equivalent) and so forth. That one human should be the one posting the image in these forums :cool: . Post processing is cool. It's the game of photography. The human may decide to swap a sky - it's been done forever - or dodge, burn, compose multiple images, and so forth, but the key is "human" and who's making the decisions.
Quoted for emphasis.

100% Agree.
 
Kinda ironic
“Hey Siri, are AI images considered photography?”
Reply with 3 different messages

d8895c08b16b2fbe1cee28661d1bbe01.png


1 There’s this:
“No, AI-Generated Images Aren’t Photographs
But they’re still art. Here’s why”




2 Then here’s a different take:

“AI is, Quite Seriously, no Different from Photography in Practice
As we know, a lot of the anti argument is the following:

AI has no soul
AI steals
AI is bad for the environment
AI is lazy
AI is slop
AI is taking jobs
However, let's compare AI to photography.

Both involve quite a lot of setting changing, parameter-tweaking, and post-processing (such as photoshop).
Both involve some level of skill or work to get a good image.
Both are the result of a machine.
Both niches are filled with the causal and the professional.
Now, the differences:

AI models require what is known as training, whereas cameras don't.
A camera takes a picture of a typically physically present item, while AI generates an entirely new one.
AI needs large amounts of energy to train, and cameras require nowhere near as much.
Cameras are and were intended to "capture reality"; AI is intended to make something new from human imagination.
Now, in practice, AI and photography are essentially one and the same, as we can see.

However, AI requires much more energy for training, much less for generating (about the same energy used in 1 google search now), and work similarly to the human brain.”


3 Finally there’s this:


“The topic of whether AI-generated photography should be recognized as authentic photography has sparked heated discussions among scholars and practitioners in recent years^[1]. There are contrasting viewpoints on this matter. On one side, proponents assert that AI-generated images differ fundamentally from traditional photographs, thus disqualifying them from being considered genuine photography. Conversely, others contend that AI-generated images embody a novel form of photography that expands upon and complements traditional techniques^[2]. In this discourse, we will delve into these divergent perspectives and examine their implications for the essence of photography.”
 
Threads should include levels of adjustments allowed in the titles (at best) or in the body of the OG post (as well or at worst). If everybody knows the rules going in, they can participate freely and without deception whether wanton or accidental.

For example, “Night Sky pics! Filters allowed: physical, balance correction in post only, please” or “No Man’s Land nature shots; add a whimsical element (of any kind, including AI) but with otherwise minimal processing” etc
 
Threads should include levels of adjustments allowed in the titles (at best) or in the body of the OG post (as well or at worst). If everybody knows the rules going in, they can participate freely and without deception whether wanton or accidental.

For example, “Night Sky pics! Filters allowed: physical, balance correction in post only, please” or “No Man’s Land nature shots; add a whimsical element (of any kind, including AI) but with otherwise minimal processing” etc
I really hope we don't get to the point of having to disclose the type of editing we do.

Editing is not a dirty word. ALL images are edited - if you aren't doing it in post, the camera is doing it for you. My film photos don't convert themselves and they require more editing than digital photos in some ways.

And AI isn't really a dirty word either, but is different than holding a camera in one's hand.
 
  • Love
Reactions: arkitect
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.