May 20, 2023 at 11:19 am #4589
I have spent the morning reading articles about AI. One excellent one was in the PD about young people at Credo School and their exploration of the subject. Another was in Petta Pixel. There are many ethical questions concerning AI that as a people we need to address. Some of my concerns are in the field of the arts. Let’s get real though, we are already using AI in our post-processing and have been for a while now. Don’t you just LOVE how you can bring out faces that are in shadow because of a light b?! Of course, you do, and how about grabbing the bg and darkening it or blurring it a bit? Maybe you change out a sky or maybe that is going too far for you. If you do or if you don’t depend on where your own boundaries of what you consider as too much in manipulating a photo. We each have opinions on what we think is and should be considered “photography.”
The definition of photography has been constantly evolving since its inception. As I said, I have grave concerns about where this latest evolution is headed. I also recognize though that I have been using, happily, AI for a few years now. I think the real question is how do we choose to define AI. We will each determine for ourselves where we place our borders on how we use it, but as a society, I do think we need to define what is created using AI so it is separated from what is created without its use. The question is how to do that and where to draw those lines.
One suggestion I heard was to refer to AI-manipulated art as a graphic. I don’t think that is a complete answer but to me, it is a start at least separating it from the label of photography which is dependent on light, which AI is not. However, we decide to deal with the issue is worthy of conversation and in a club such as ours there seems to me to be no time like the present. I suspect a great many of us already use AI in one form or another. I have used it here several times in the form of spell check. 😉May 20, 2023 at 9:04 pm #4590
I have been using Leonardo.AI for the past few weeks, and I’m trying hard not to be seduced by AI.
I’ve gotten deep enough into it to see trends and possibilities, and as Competition Chair, I see a clear context where AI can be used and still fit within our competition rules even though it amplifies one’s abilities greatly.
A feature of Leonardo.AI is the ability to create what is known as a custom AI model by inputting one’s own images to train it (I made a custom model using images from my portfolio entered in 2021 – see below). Let’s say that I import 500 of my own images to train the model. The textual prompt that I might use to instruct the processing would go something like this:
“Using only images within my model and no other image source, extract and isolate interesting objects from my images and create a fictitious scenario that layers these objects into a realistic arrangement and landscape where these objects appear in a realistic context as well as make an intriguing scene. Use cinematic lighting with a 50mm perspective and output at a 4×1 aspect ratio.” …you get the idea. The prompt can be much more specific such as specifying the location: at Ocean Beach in San Francisco just past sunset with large waves crashing on the shore. This assumes that I have one or more images of this scene in the model trained from my own images and that there is sufficient imagery contained therein for the AI to identify SF’s Ocean Beach.
It can generate up to eight possibilities at a time.
It will then do in seconds what would take ages: go through 500 images and manually remove backgrounds from objects and save the objects in separate files or layers to be composited together and then play with dozens and dozens of clipped image objects to arrange into a realistic imagined scene.
One can already import one’s own images and have it creatively fix or change part of the image, but one can’t limit it 100% to one’s own images (SRPS / N4C rules), but I’m sure that capability is not far in the future.
Our rules rely on the honesty and integrity of the member. AI can’t change that. 😉
Here’s one of the input images:
And here is one example of many it generated:May 21, 2023 at 12:39 am #4591
Here’s an example of a civil war portrait generated. Note that unless you enter a “negative prompt” to exclude various things, anomalies appear. Can you spot one in this image?May 21, 2023 at 12:58 am #4592
Another:May 21, 2023 at 7:21 am #4596Michael FunkModerator
Too many fingers LOL.May 22, 2023 at 8:42 am #4605
As I am understanding what you have said, all the photos you are talking about were yours. The program basically takes your photos and given the directions you have put into the program arranges those photos in several different renderings.
We have available to us many programs that take a photo and apply a filter or rearrange the pixels. One such would be Topaz Impressions. In this I find similarity. The difference and I think it is a big difference is that with such programs as Topaz, YOU are making the artistic choices that create a final product. The YOU does not seem to be present in the AI programs. Yes, you are telling it what to do, but IT is making the choice on what the final product is.
I think the creativity comes with the human factor, the decisions that a person makes to get to the end they envision. With AI it seems to me that has been handed over to a machine to dictate. Now for something like spell check, medical advancements, and many other areas AI is being explored, I see no problems. I think though when we let any other identity human or otherwise make those artistic decisions for us we have given up the very thing that makes art valuable. It is the experiences of the human factor that has gone into any work of artistic creation. I may let a machine spell for me, but I am not ready to let it speak for me or create my artistic vision.May 22, 2023 at 10:13 am #4606philliphannParticipant
I see a vast difference between using AI to create a mask in my image (e.g., to adjust the background exposure, focus, etc.), vs using AI to create a whole new image using my own image library as a basis.
In the former case, I am using AI to perform a task that I could perform by myself. AI is saving me time, but I remain in full control of the actual editing process. In this case, AI is a tool that is automatic/faster, but otherwise provides the same result as the manual selection tools already provided by Photoshop and other apps. I feel perfectly comfortable with this scenario.
In the latter case, however, AI is creating the image, not me. Even though the source pixels may have come from my own image stock, and AI is following a set of criteria that I provide, the resulting image is a surprise….and so it is not really my creation. I don’t believe these should currently be eligible for any SRPS competition.
However, if future AI capabilities allow us to specifically select the images it will use to create a new composite image, and we can tell the AI exactly what we want it to do, then this becomes another tool in our post-editing box, and should be eligible for entry in SRPS creative/pictorial competitions.May 22, 2023 at 10:18 am #4607philliphannParticipant
PS…By the time I got around to sending my thoughts, Terry had already entered hers. I completely agree with what she said.May 22, 2023 at 3:10 pm #4608
“In the latter case, however, AI is creating the image, not me. Even though the source pixels may have come from my own image stock, and AI is following a set of criteria that I provide, the resulting image is a surprise….and so it is not really my creation. I don’t believe these should currently be eligible for any SRPS competition.”
The magnitude, scale and speed are vastly greater. However, rules are incredibly difficult to write without ambiguities.
Who (or what) is the source of the suggested result? When someone is mentoring a member (i.e., Tony, hypothetically) and Tony says, “The way you cropped your image focuses on what you believe to be the subject, but if you crop it differently, it emphasizes what I believe to be the much more intriguing subject in your image.”
As a result, the member notices something that they hadn’t seen before (and hopefully learns to notice other ways of interpreting a scene whether composing a shot or examining in post), and proceeds to enter that into competition. Whether an AI scan of 100,000 of your images reveals many suggested composites or just examines each image individually and highlights by cropping and other post-processing techniques refining the 100 images that are most striking, and some of those are entered into competition, where do you draw the line in the rules? Do you throw the baby out with the bath water and ban mentoring? We are about learning and honing our skills, and AI is revealing incredible ideas. The new skill that we all need to hone is how to have a conversation with AI to elicit and amplify what starts as a thought that is uniquely our own.
Using the Renaissance artist Jacques Louis David as a seed, I asked, “In the style of Jacques Louis David, Minutemen and Redcoats are having a bacchanal at the Boston Tea Party.” :
What if you asked, “What images among my 100,000 come closest to the style of Jacques Louis David for the number, positioning and expressions of people in the image?”, and see what results? And, using the resulting 100 images, ask the AI to composite elements from them into a result that comes as close as possible to that artists’s style? Realistically, an expert in Photoshop who is also knowledgeably credentialed regarding Renaissance artists could do this for 100 of your images, but not 100,000. Where do you unambiguously draw the line?May 23, 2023 at 10:23 am #4609
PetaPixel put this out today. I think it very clearly shows the process at least in Ps Beta. I do not see any skills presented in this other than the basic skill to write and spell ( hm there we are with spell check again, eh?) Other than the initial photo chosen by the artist everything else has been decided by a machine.
To the argument that AI is just another mentor in our lives, I would suggest that going to a museum or studying work online is more beneficial. When I am studying another artist my eyes, and my mind are taking in what I see and transforming them based on my own experiences, tastes, and skills to form a work of my own making. Looking at renderings from AI I am only making a choice on something someone else, a machine has created. It, not I am the creator. Where is the thought process? I can go into a store and tell the person behind the counter I would like a tube of toothpaste that doesn’t mean I created that toothpaste.
I agreed with everything but the last paragraph you wrote Phill. Even if AI allowed us in the future to specifically select the photos that would go into what we are working on, It, not us would still be creating the toothpaste. I could see asking for a car say and AI going through our library to show us all the cars there. ( It can do that now I believe ) If then we select the car and proceed from there to make a work …………. well that is, I think different.May 23, 2023 at 11:18 am #4610May 23, 2023 at 12:20 pm #4611
I think, for me, the question is not do I use AI because let’s face it we already are. The question for me is how far will I go in using AI, where is the boundary? When does a work cease to be an extension of my voice and becomes just another automated process?May 23, 2023 at 1:10 pm #4612Bill TheisParticipant
so playing the devils’ advocate: after AI has had its way with your image as fodder, is it still your image?
it may be “better” in some people’s opinion but isn’t it more fun to go out and take the photos than to futz with software. Imperfection and limitation are challenges that should not be avoided
No I am not a LudditeMay 23, 2023 at 3:22 pm #4613Bill StacyParticipant
Interesting discussion and I think Steve Muench summarizes the issue nicely:
” Even though the source pixels may have come from my own image stock, and AI is following a set of criteria that I provide, the resulting image is a surprise….and so it is not really my creation. I don’t believe these should currently be eligible for any SRPS competition.”
In fact, I think that might be a good way to “draw the line.” If you are surprised by the result, it is probably not your creation.May 23, 2023 at 4:41 pm #4614
Quoted text is Phil Hahn, not me (Steve Muench).
A surprise can come from a variety of sources: a mentor, a judge, or an AI, etc. A mentor or judge could edit your image to show surprising possible transformations and improvements. Ethically, we treat the surprise as a suggestion to try and not a result to enter into competition; nothing’s changed.
Be careful to differentiate refinements from transformations, though I have been ‘surprised’ by refinements done by Luminar’s AI that would not’ve occurred to me otherwise.
Also, even after an AI has gone thru 100,000 of your images and composited original pixels into surprisingly interesting scenes, one can still recognize or find which photos contain the source elements by simply asking the AI (coming very soon) to display each source photo and highlight the elements used. Then, you can composite something similar yourself. Perfectly acceptable to enter regardless of the source of the idea if you do it with your own labor. To me, the real surprise is the scope, magnitude and scale going from a few judges and mentors to zillions of comparisons and analyses distilled into frequently astonishing results.
We welcome the surprises. We are more likely to learn from surprises than minor tweaks (though seemingly minor tweaks can produce winners).
I love surprises. Shared in competition, it thrills and benefits us all.
- You must be logged in to reply to this topic.