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
Image cleanup or removal of unwanted objects, dragging an object and moving to a different location, generative fill, placing a subject in an entirely different background, altering the weather condition or time entirely different from when the actual condition or time the actual shot was taken, these are some scenarios that I think should be disclosed whether done with the help of artificial intelligence or any software.

In some cases where minor adjustments like brightness or saturation or vibrance, or any usual/traditional edits, should be fine/acceptable, and be left to the poster whether to disclose or not, but it would be helpful aside from the camera settings as to what minor adjustments were made to be disclosed.
You make some good insights that should be considered. I guess when you look at apps like Photoshop historically, whilst sometimes used to correct images, the term 'photoshopping', became ubiquitous to manipulating/correcting flaws in an image. In many ways, a precursor to what we now call AI.
 
  • Like
Reactions: uacd
I was having a lively debate about this at work. My opinion is that AI editing is no different to Lightroom. As soon as you mess with a raw image in any way it is no longer the truth that was captured by the camera but a digital lie. Most 'great shots' on the internet were the result of a lucky spray-and-pray, heavy editing and some skill in composition.

A truly great photographer doesn't need to edit images; they made sure that the conditions were all perfect before they even pressed the shutter button and visualised it before they even touched the camera. I should caveat the hypocracy in this statement because I edit as much as the next person and far from a truly great photographer. But all of my best images are unedited, taken at the right time of day but these are few and far between.

EwtaQv9WUAIjLsA.jpeg.jpg

This image I took of a bubble just as it landed on some grass was both sheer luck and required some editing to bring out the sharpness.

2023-11-09.jpg


At the same time this image of the December snowfall was taken just after dawn and has been unedited. I took my camera out to purposefully capture the morning after a heavy blizzard in the local woodland. I got dozens of great shots that day and have never tweaked a single one.
 
I was having a lively debate about this at work. My opinion is that AI editing is no different to Lightroom. As soon as you mess with a raw image in any way it is no longer the truth that was captured by the camera but a digital lie. Most 'great shots' on the internet were the result of a lucky spray-and-pray, heavy editing and some skill in composition.

A truly great photographer doesn't need to edit images; they made sure that the conditions were all perfect before they even pressed the shutter button and visualised it before they even touched the camera. I should caveat the hypocracy in this statement because I edit as much as the next person and far from a truly great photographer. But all of my best images are unedited, taken at the right time of day but these are few and far between.

View attachment 2538883
This image I took of a bubble just as it landed on some grass was both sheer luck and required some editing to bring out the sharpness.

View attachment 2538885

At the same time this image of the December snowfall was taken just after dawn and has been unedited. I took my camera out to purposefully capture the morning after a heavy blizzard in the local woodland. I got dozens of great shots that day and have never tweaked a single one.
Two lovely shots. However these type of photos are unlikely to need heavy AI intervention. Now taking a shot at a very popular tourist attraction that isn’t open early or late, meaning you will have people in your shot or you have to clone them out.
Neither is right or wrong. Just depends on how it is presented.

But I agree as much as you can, the skill is getting it right in camera as far as possible.
 
I'll just leave one of my favourite quotes here:

"Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships."
Ansel Adams


As soon as you mess with a raw image in any way it is no longer the truth that was captured by the camera but a digital lie.
I will beg to differ, in the nicest possible way. Please do not take this personally, it is just a response to your emotive use of the terms Truth vs Lie and then I go rambling along as I tend to do. 🙂

What "truth" is being captured by the camera?
The F-stop influences our depth of field. Who has eyes that see like that?
So stright off the bat that little camera ain't telling s any kind of "truth".
Denoising and adjusting the colour grading is not being dishonest. If a particularly egregious power cable snakes across the sky, I will take it out.
But, I have to ask you, where is this big lie?
Was I not there? Does the scene not exist?

Black and White photos get a pass? My world is full colour… so B&W is one heck of a LIE.

Are all iPhone photos lies? You do realise the incredible amount of manipulation iOS puts our images through before we get "The Truth".

I do not think we should be confusing traditional post processing whether analogue or digital with AI image creation.

Some time ago on these forums there was a poster who was very happy with his shots straight out of the camera (I think the long time posters here will be reminded.) No amount of critique or debate could convince him that his genuinely dull snaps were well… just dull. Not just poorly composed, but utterly washed out and dreary.
But he was happy as Larry. So good for him and you would probably applaud his Truthfulness.

Do I want to take photos like that? Hell no.
Would I go and look at an exhibition of his work? Unlikely.

Photography is — to me — art.
Period.
I earn my daily crust as an artist. I do not use AI and I only turn to my computer to clean up the scans of the originals. Does that make me a dishonest artist?

With photography (from which I do not earn a living) I seek to recreate what I saw and not just what I was looking at.


I think we should distinguish between Post processing, editing and Ai manipulation.

Here is a pic I took — I didn't pay attention so it was underexposed.

Raw — The TRUTH (apparently)
P8010023 (1).jpeg



DXO PhotoLab (Still not a great pic, but… No AI) THE LIE (Apparently)
P8010023 (2).jpeg



ChatGPT (And this is what Adobe and Google and Apple et al are pushing on us.)
Now this image IS a LIE.
P8010023 (3).png


I know which image I am OK with and which image makes me throw up.
 
Last edited:
A truly great photographer doesn't need to edit images; they made sure that the conditions were all perfect before they even pressed the shutter button and visualised it before they even touched the camera.
You know very well that not every scenario can have a perfect shot, kids, animals, plants, events, these can be planned but are challenging scenarios to get that perfect shot without having to edit a single thing.

Kids - they rarely stay in one place
Animals - they hunt for food specially the wild ones
Plants - windy or rainy weather
Events - social gathering or newsworthy photos
 
Simple I consider AI generated images to be “artwork”, not “photography”.

There’s nothing wrong with AI images, clearly we’re on the cusp of them.
They can be lovely artwork.

But they are fake, not reality.

I’ve been taking Astrophotography images since 2009, literally 1,000’s of them. It was very clear to the forums that ”Astrophotography” kept the photons captured from the image sensor in their spacial position / relationship to each other and enhanced them via various techniques; PP, image stacking, using lens filters to block UV light, mounted on a equatorial rotating device, etc.

Moving objects was strictly a no-no, except the man made satellites that took away from the space objects; ie, plane trails, satellite streaks, etc.
 
You know very well that not every scenario can have a perfect shot, kids, animals, plants, events, these can be planned but are challenging scenarios to get that perfect shot without having to edit a single thing.

Kids - they rarely stay in one place
Animals - they hunt for food specially the wild ones
Plants - windy or rainy weather
Events - social gathering or newsworthy photos
And that is one of the reasons that I love film; it makes the notion of perfection - 'the perfect shot' - a lot harder to achieve, for it insists that the photographer actually think before pressing the shutter (precisely because film limits the number of shots you can actually take).

The nightmare of a world where photography is composed of perfectly curated shots, the notion where, what my eye falls upon is a curated visual lie, fills me with utter horror. I don't want perfection - I want an image to be an authentic expression of the eye of the person who created it.

Anyway, excellent thread, and a fascinating and thought-provoking discussion, and thank you, @mollyc for starting it, and @Clix Pix, for having suggested it.

Obviously, as with digital photography, (a format I have yet to graduate to - I still use film), AI will transform (transform utterly) how we see, shoot, (if we even shoot, or simply appropriate as needed), process, and present an image.

I am an historian by background, a training that has given me a respect for facts, sources, and what is (rather, what was), and also a tendency to look to a longer perspective; now, I have no quarrel with fiction or fantasy, but they do not purport to represent reality.

Thus, part of this unfolding discussion and debate of whether AI has a role in photography (alas, I think that this debate may already have been lost) and where we, in this forum seek to draw the line as to its use in image creation - a very necessary and timely discussion and debate - concerns whether it is deemed acceptable (in a world where tweaking images is already possible with digital photography), and to what extent - which is where I think that the real debate will (or is) taking place - it should be used when creating images.

In an earlier life, for a year, I worked with the parliamentary debates office as an editor; one of the things that was interesting about that was how the edited (and published) debates did not always reflect exactly what had taken place, what had been said, (sometimes, atrocious grammar was tidied up, as, if one actually had transcribed exactly what had been said, the contribution would have been completely incomprehensible, closer attention was paid to what the front bench - especially the government - may have said than was paid to the utterances of an anonymous backbencher, and - depending on context - errors may occasionally have been corrected prior to publication) and, internally, much discussion (and debate) occurred about how best to render what had been said into a published form, a form that rendered accurately what had been said in a manner that was comprehensible to a reader. I found it fascinating, and - as an historian - at times, a little troubling.

However, while the forum has already made its stance quite clear on AI generated posts (posts should be written by the person themselves in their own words), in the photography section, it makes sense to discuss the use of AI generally, and, more specifically, whether - or, to what extent - it should be allowed when participating in the many threads (weekly contests, POTD, and others), which is what this very welcome thread is doing.

Personally, I have no wish to live in a world where the visual fraud of AI generated images smothers any expression of creativity, and extinguishes authenticity of expression, and the need for the photographer's subjective eye.

My own personal preference would be for no AI whatsoever, - at least, in these fora - although I accept that this is, perhaps, excessively optimistic.
 
Last edited:
I don't like it, but since day one of photography, man has enhanced and altered photos. It's more a question of method than principle. What is permissive, then? Well, in my view anything that nears actual realism, since photos often are imperfect representations of reality in different regards. Also, alterations that are self evident, i. e. B/W or sepia photos. Cloning and AI removal of objects are not, and will never be. AI pictures that resembles reality aren't photos at all, so they can't be discussed, in my view. A few thoughts. But, I realise my thoughts are irrelevant for the future; I'm mainly the past and only a tiny bit of the present.
 
I'm against the use of generative AI in photography, including things like canvas extension and generative fill.

I barely even use "classic" tools like cloning for such tasks. Except for literally two or three images, all photos I posted here or elsewhere, or most personal photos otherwise are only straightened, cropped, and corrected for luminance or colors.
 
Is it really photgraphy when the computer generates the image?

I think its a slippery slope where AI is used for usual post production touch ups, to then adding a bit more definition, then to altering the images to convey a feeling, meaning, to eventually being almost completely generated.

What comes to my mind is the Ship of Theseus sort of thing - whether an object that has had all of its original components replaced remains the same object.
 
I'll just leave one of my favourite quotes here:

"Dodging and burning are steps to take care of mistakes God made in establishing tonal relationships."
Ansel Adams



I will beg to differ, in the nicest possible way. Please do not take this personally, it is just a response to your emotive use of the terms Truth vs Lie and then I go rambling along as I tend to do. 🙂

What "truth" is being captured by the camera?
The F-stop influences our depth of field. Who has eyes that see like that?
So stright off the bat that little camera ain't telling s any kind of "truth".
Denoising and adjusting the colour grading is not being dishonest. If a particularly egregious power cable snakes across the sky, I will take it out.
But, I have to ask you, where is this big lie?
Was I not there? Does the scene not exist?

Black and White photos get a pass? My world is full colour… so B&W is one heck of a LIE.

Are all iPhone photos lies? You do realise the incredible amount of manipulation iOS puts our images through before we get "The Truth".

I do not think we should be confusing traditional post processing whether analogue or digital with AI image creation.

Some time ago on these forums there was a poster who was very happy with his shots straight out of the camera (I think the long time posters here will be reminded.) No amount of critique or debate could convince him that his genuinely dull snaps were well… just dull. Not just poorly composed, but utterly washed out and dreary.
But he was happy as Larry. So good for him and you would probably applaud his Truthfulness.

Do I want to take photos like that? Hell no.
Would I go and look at an exhibition of his work? Unlikely.

Photography is — to me — art.
Period.
I earn my daily crust as an artist. I do not use AI and I only turn to my computer to clean up the scans of the originals. Does that make me a dishonest artist?

With photography (from which I do not earn a living) I seek to recreate what I saw and not just what I was looking at.


I think we should distinguish between Post processing, editing and Ai manipulation.

Here is a pic I took — I didn't pay attention so it was underexposed.

Raw — The TRUTH (apparently)
View attachment 2538910


DXO PhotoLab (Still not a great pic, but… No AI) THE LIE (Apparently)
View attachment 2538911


ChatGPT (And this is what Adobe and Google and Apple et al are pushing on us.)
Now this image IS a LIE.
View attachment 2538912

I know which image I am OK with and which image makes me throw up.
There is indeed an argument that every photograph is a lie, but no more than a painting.
 
  • Like
Reactions: uacd and Janichsan
You know very well that not every scenario can have a perfect shot, kids, animals, plants, events, these can be planned but are challenging scenarios to get that perfect shot without having to edit a single thing.

Kids - they rarely stay in one place
Animals - they hunt for food specially the wild ones
Plants - windy or rainy weather
Events - social gathering or newsworthy photos
When it comes to casual snaps I find the imperfections make the image. ‘Remember when she’d never sit still? Look at all the blurry images!’ Is a common comment in my house! It’s the same with all those tourists. If want to remember how busy it was. I’d come back at 4am if I wanted it deserted.

But I do see your point. Lots of great images are off the cuff, I’m not trying to detract from that at all. If you’re in a stressful situation who has time to compose at all?
 
  • Like
Reactions: uacd and Janichsan
Thank you everyone for your thoughtful opinions! This has been interesting to watch unfold. I am not sure we will be able to make an official determination for the overall forum, but it's really great to see such considerate responses.

I would like to ask that this conversation lean to the use of AI within photography vs editing. Although they are related, as has been noted several times here, editing has been around as long as photography and even compositing goes back to the 1800s in the darkroom. I'm not here to debate the merits of editing in general but what role AI now plays and may play in the future.

My personal position is that I don't use AI, although I am not against any manual adjustments that, while time consuming, can be done in the darkroom/LR/PS. I know many talented photographers who make extensive use of compositing multiple parts into a single image, but it's all done by hand, and not with the use of AI. If you are that capable in a photo editing program then have at it. But once you start having AI do the compositing for you, I think you've stepped off the rails a bit.

That said, if you are using AI as merely a time saver to clone out power lines, for example, or add a little bit of canvas because your crop was slanted and you need to straighten it - like if we are talking about less than 3% of the image being AI manipulated, and only because it's faster to AI it than manually do those edits, I don't really think it's that big of a deal and wouldn't need to be disclosed. I have one image in my archive where I essentially recreated an entire side of an image, cloning a hedge from one side to the other....I'm happy to share the SOOC, and I'll never hide the fact that I did that much cloning, but I did it all by hand and it's because I had a vision of the final image that I wasn't able to fulfill in that location (and this image is from roughly 15 years ago, and I don't think we even had content aware back then). I don't usually disclose the cloning when I share the photo, but if I did the same image now and AIed it, I think I'd have to. Cloning manually has a human element to it, where a person is in charge of the creative decisions....AI does not have that human aspect with generating from a prompt.
 
My thoughts are that images posted in the MR forums that have been created in whole by generative AI tools should not be called photographs and should not be used in photography competitions or threads. Images that have had parts of them created by AI (extending a canvas, creating a sky, generative fill, etc.) should likewise not be entered into any competition with other photographs, though they certainly can be displayed to other MR forum users as part of a discussion or to show work one is proud of. Those must be labeled - very obviously - as being created with AI tools.

As far as using digital tools to change images (cloning, etc.), I only use those tools to correct a problem that the camera itself has created. For example, cloning out a speck of dust on the sensor. If there's an awkward tree branch, etc., so be it. That's the world. It's not perfect. I don't use AI tools in such instances.

My background is in photojournalism, and newsgathering organizations are dealing with these issues as well, all with the goal of making sure people can trust their images as being factual. Getty, for example, routinely offers guidance in their policies for photographers who shoot for them. Currently that includes warnings that Getty expects photographers will not use any AI tools, period, in the creation or post processing of images. And that the usage of any such tools would result in the photographer's contract being canceled and their work removed from the service.

I know people coming to the MR forums may not have the same expectations for the work seen here as that from major news organizations. And I don't expect the forum's rules should reflect some other organization's values. I can only speak for myself when saying I try to take the approach of my work reflecting reality. Others feel it is a place for their artistic expression, and the rules for altering images are different. I'm totally good with that (though I would appreciate disclosure when images are changed during post processing).

Like @mtbdudex mentioned, I too occasionally shoot and post astrophotography and often use layers of multiple exposures and similar processes to create an image. Also I have stitched together images to make a landscape panorama on occasion. In those cases I describe in the post that I have used multiple exposures, etc. I haven't used AI tools for my astrophotography, including AI noise reduction tools.

Generally I post camera and lens and exposure information with my photos. That's one way of letting others know about the authenticity of my images, and to give other photographers some insight to my work, if they are interested.

Thanks @mollyc for starting this discussion!
 
I do think it’s a lot harder to avoid AI in phone photos. It’s hard to even get a pure raw file sometimes.
Rarely do I use the memory creator in Apple Intelligence. I only used it one time to create a slideshow of photos from my business trip months ago.
 
I am old school and for me a photograph is something that is taken with a camera. Sample definition of "Photography"

Since this is the "Photography" sub-forum everything posted should have been originated from a camera (that thing that you hold in your hands or is mounted on a tripod for example).

So maybe the rule for this sub-forum needs to be: post photographs taken by you with a camera (film, digital, phone etc etc)

Now, the next question is, how much "manipulation" is ok for that photograph?
in my mind, the "traditional" exposure/color etc correction is totally fine, and that includes "minor corrections" like eg removing lens flare or something like that.
Where it gets complicated is when an entire object is removed from a photograph, eg a trash can or a person that just was on the scene but was not the actual object of the photograph. In those cases I think it should be revealed in the posting like eg "removed object xyz or something like that.

And in general, for me, I want a photograph to represent what I saw with my eyes at the moment I took the photograph, so in @Apple fanboy 's example with the storm troopers - that does not fit my description.

And having said all of the genAI-generated images have their place and purpose and maybe there needs to be a new sub-forum when those images and techniques become more "common".
When someone says "I want a photograph to represent what I saw with my eyes at the moment I took the photograph" they eliminate all of the print work of photogs like Ansel Adams, because his ultimate prints were substantially adjusted from what he saw prior to shutter release. Even though he was an advocate of "pure" photography.
 
When someone says "I want a photograph to represent what I saw with my eyes at the moment I took the photograph" they eliminate all of the print work of photogs like Ansel Adams, because his ultimate prints were substantially adjusted from what he saw prior to shutter release. Even though he was an advocate of "pure" photography.
why? say I replace a cloudy sky with a sunny one - is that what I remember when I took the photo? no.
I have no idea what others see when they take a photograph. I only know about what I see.

I can have my own standards for my photographs and still like what folks like Ansel Adams have done, not mutually exclusive. The world is not black & white ...
 
  • Like
Reactions: Scepticalscribe
I’m not digitally talented enough to create photos myself so my Photos Library consists of images I took myself with my equipment or equipment I borrowed or are scans of print images I (mostly) took and had printed. There are a few family photos others took that I scanned.

I use Luminar Neo to enhance many photos. It does use AI, but not to substantially alter the image I took. Meaning, there is no way to insert myself into an image I was never in. (If there is a way, I’m unaware of it.) I admit I’ve changed out a sky or two, likely 3 or 4 times. A few I’ve posted in the photo of the day threads. I didn’t disclose the sky changes but will in the future if it’s decided that’s important.
 
One major complication in this discussion — and determining what is on or off limits — is the rampant branding of all sorts of long-time computing algorithms or long-used tools or processes as “AI”.

It's a bandwagon effect — seeking publicity, funding, hype, buyers, etc. The underlying code or actions may be virtually or even exactly the same, but now the photos are “AI enhanced”!

Here's an example of how that could work. We have long had the “Magic Wand” tool in the Photos app — it auto-magically makes a variety of changes in the photo's exposure, contrast, saturation, etc. Now, if Apple were to brand that as AI, does that change whether the resulting photo is acceptable or not?

No doubt, Photoshop and other editing apps have various processing tools that might be acceptable, but if they’re branded or repackaged as ”AI”, does that change the status of the photo they shaped?

Thinking about this, I wonder if any photograph that has an automatic transformation applied — without the photographer's intervention — is problematic…

That introduces a criterion or notion of human control or action — and another criterion of whether it was random, experimental, or thoughtful.

On a different plane, I do like the generative vs. editing distinction, although that has challenges, too, in the borders between them.

P.S. The editing of parliamentary debates is a fascinating counterpart — in the realm of written language as opposed to visual representation. Such editing strikes me as beyond the limits for official proceedings in a way that it doesn't for tweaking certain aspects of photographs. If someone makes an incoherent or semi-coherent statement, or with grammatical gaffes, the record should reflect that. It's an official record of what was said and argued.

The U.S. Congress has an interesting variation on this — not in the benign hands of a nonpartisan editing staff trying to make things more coherent, but in the hands of the elected official speaking and his staff. They typically request permission of the chair “to revise and extend their remarks”, which occurs at a later time. Thus, what live C-SPAN shows legislators actually stating and arguing may not be reflected in the printed, bound, official Congressional Record. They can literally delete, revamp, tweak, change facts, and alter what they said.

That has parallels with generative AI producing photos or transforming them well beyond the original! As always such analogies break down on close examination, but I think you get my drift…

For example, if it were a photo of a daytime scene with intrusive shadows or background dark clouds, and it's changed to shadow-free and sunny, that would be troubling. Yet, it would not be as bad as adding an object or putting a different person in it — or changing what they were wearing. So, there are definitely multiple dimensions to all this — now we’re into whether the changes are applied to or generated in the focal point vs. the background.
 
  • Like
Reactions: Scepticalscribe
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.