Obviously, this is a hot topic. We photographers tend to disregard it. I don't think we should be so snotty. After all, we all do post-processing and often to a point where the image is more a work of art than a realistic scene. The various AI enhanced editors on the market and the new generation of AI generating programs just lift that to another level. In my opinion, all that is requested is to mark heavily enhanced or AI generated images as such.
In any case, I like to request a subforum for AI processing, just to keep ourselves sharp on the subject. We cannot simply close our eyes.
Here is my first try with getimag.ai, and it turned out as a surprise to me. I would have worked hours to get a similar result.
First the original image I took from a shop window, because I liked the empty look.
I then requested the AI to make that an image of King Charles. I do not know which Charles they took and from which source. But the changes were more and better than I expected.
Very good result indeed !
I think that, when it comes to photographic creations, AI only brings another way of producing images. It can also give a tool to imaginative writers that are not graphical designers to elaborate graphical works.
And as any tool, if you don't use it, it does nothing at this right time...
To my sense, as long as you decide or choose, you create and so, I'm more concerned about AI being a tool to influence people or even worse, to lead to restriction of human free will. That being said, human being itself seems to be already able to do so very efficiently without the need of AI...
Back to the subject, I've seen since many years, graphical artists listing the tools used and explaining their work. In the case of your image, adding simply "Global picture is my own photographic work and the face + royal details are generated with text AI tool" may be enough to answer to any question about the way it was created.
In your specific image, the initial choice is very important (subject, shooting settings, framing etc...), and the humorous "written" addition is too. At least, that's what may make it "efficient" to my eyes.
For me, good editing is the kind of editing where the final result basically reflects the original scene. Yes, we do postprocessing, but the postprocessing should not deviate from reality. Of course, it's hard to define exact limits, but I always ask myself: if someone had stood with me while I took the photo, would they be able to say honestly that the end result basically reflects what they saw?
Bringing up shadows and highlights for example is basically like having our eyes adjust with their large dynamic range. On the other hand, your edit with AI added something new and it's not at all like what you saw.
Yes, AI is another way of producing images, but your end result here is not really photography any more. Personally, I don't think it's a net positive gain to art either.
Personally, I don't think it's a net positive gain to art either.
I completely agree. It was not meant to be one.
The more I think about photography, the more it becomes just a means to show others my views to reality, tell stories about this reality, or disclose things that others do not see. AI generated reality does not fit to that idea.
However, I understand that others might see photography as a way to express their imaginations or dreams.
To me it is all about context. If it is for photojournalism like Nat Geo then even something like too much saturation can be unethical because the viewer is led to believe it is intended to be realistic. If it is for the gallery wall then anything goes. The artists imagination and the ai tech can produce results that are as valid as art as 'sooc.'
Personally I think it fits well in the post processing section without needing it's own forum.
@renegr I must say my own view is not that of photographers, as I (still) don't consider myself as one (maybe it's better like that :D). I come from graphical design and photography had always only been a tool for creating images for my usage. To me, there are just different levels of distortion of reality depending on the quantity of changes you make from the "original" OOC image, and some context you can put around this image.
At the opposite way, I did a project with a Japanese fellow at deviantArt some years ago where the purpose was to provide only OOC images - or at least images that were just as the software renders them in auto mode (whether bokeh is nice or not :D).
He called it the "Zen Project". All the process had to be a kind of "out of thought" one from the shooting to the publishing (so we were never giving context or information with pictures). That was very interesting. Mainly because we have seen that, in the end, interpretations were as various as if the images have been worked out hard, but also because, to respect the goal, I had to train a lot before being able to shoot - only nearly - without thinking (my Japanese mate was already ready at the beginning of the project, I wasn't).
From yet another perspective, it's completely a different thing if your purpose is journalism, of course. But still, and even more in this case, I think the image alone is not enough, most of the time, because you need to give dates, context and other elements in order to make the image a valuable testimony. The interpretations can fully change depending on what information is given with the image (sometimes even a one word title can give a total different meaning, I bet you know what I mean as we can see that every day in news medias nowadays).
A long and rich debate... maybe as the one that been opposing the defenders of figurative painting and surrealist artists, for instance.
"I do not know which Charles they took and from which source."
So you are OK with stealing the work from others?
If you want to create a Photograph, use your camera and darkroom or software to do it yourself. Why be a thief - even though you are blaming a machine?
"I do not know which Charles they took and from which source."
So you are OK with stealing the work from others?
If you want to create a Photograph, use your camera and darkroom or software to do it yourself. Why be a thief - even though you are blaming a machine?
Strange twist that you bring to a thread aimed to be about the problems of AI and wanting a discussion about it. I did not steal anything.
Let's make sure to keep the discussion friendly even though folks feel strongly about the topic.
"I do not know which Charles they took and from which source."
So you are OK with stealing the work from others?
If you want to create a Photograph, use your camera and darkroom or software to do it yourself. Why be a thief - even though you are blaming a machine?
I have to excuse myself for the snippy reply. Your valid point deserves a better answer.
ChatGPT is indeed taking its content from all the sources it can get its hand on. This generates not only problems with authorship, but also with reliability. But we all do the same when we are chatting, don't we? Our writing does rarely satisfy scientific standards. An extreme example are forum replies 😉.
Currently, AI is in the state of a very gifted toddler. It just repeats things it has read somewhere. With development, it will be more reflective and learn to express varying opinions. It's just the same development humans go through.
Image generation by AI is very similar. It does indeed steal images from the net without citing or asking. In contrast to a human, it has no other way to get images. (That's why I think that photographers will survive, by the way.) It is difficult to judge, how illegal or bad this borrowing really is. After all, the images are not cited or copied directly. Legally, it might still be copyright infringement.
In my book, taking an image and alienate it to fit into another context is a negligible sin. Sill, citing should be done, and I'd do it.
"I do not know which Charles they took and from which source."
So you are OK with stealing the work from others?
If you want to create a Photograph, use your camera and darkroom or software to do it yourself. Why be a thief - even though you are blaming a machine?
You may not have intended it, but this kind of inflammatory language changes a discussion into an attack or argument. I hope that was not your intent because it would ruin these forums.
Plenty of work is in the public domain and free for use. The source material for AI is part of an ethical debate based on whether it is simply scoured from the internet or sourced into databases based on public domain images.
There is a solid history of combining purchased or public domain images as art. PPA has a category for this kind of work in their international competitions. Many photo clubs and museums accept this kind or imaging - and often without any disclosure. For example, sky replacement can enhance an image and is very easy - but not so easy to do well.
On the other hand, I respect your opinion that you prefer other types of art that only have content from the photographer or only have content from a single image. It is interesting to follow the early evolution of AI in photography. I wonder how we will feel about AI being used for subject detection in new AF systems?
Eric Bowles
www.bowlesimages.com
I think it is kind of like defining what is a fair use of an image as a derivative. For example Andy Warhol was sued for using a famous photo of Marilyn Monroe in his printmaking. He won because derivative use is permitted if the resulting art is different enough for the new work to stand as art in its own right.
This is such a gray area for many (including myself)... I think we will need to go through some lawsuits, petitions, calls to action and other things that have happened in the past to answer many lingering questions about the use of AI for things like photography.
Most early AI algorithms were unable to create things from scratch. Those algorithms would go through millions of similar images they crawled through before, then blend different parts of images in some fashion, adding shadows, making similar tones, etc. We have seen a lot of images like that. The next generation of AI imagery is already here, and it is able to do a lot more than that - in some cases, it is even capable of coming out with a fresh new concept that has never been seen before. Some results are downright scary (scary good, but sometimes also scary bad). I can't imagine what AI is going to do in the next iteration, creating realistic details and very high-res images.
At that point, what images are real vs generated? How could anyone even tell them apart? You already have heard of photographers creating images through AI and winning competitions, with judges having no idea that the photo was not real. I think it is only going to get worse from there.
Personally, I am not looking forward to any of this. I have zero desire to generate AI images. Our lives are already filled with fake news, AI-generated content, virtual reality and so much more. I would like to keep at least part of my life...real.
Here is a link to an article about the Sony World Photography Award winner who submitted an AI image.
https://www.bbc.com/news/entertainment-arts-65296763
Luckily this photographer submitted the AI image to prompt discussion, and immediately declined the award. The photographer is against the use of AI in photo contests.
You wonder if the photographer even can copyright the image. It's a computer answer to a query with specific criteria. Was the image created by the photographer, the software, or the AI bot?
Eric Bowles
www.bowlesimages.com