• Welcome to Photrio!
    Registration is fast and free. Join today to unlock search, see fewer ads, and access all forum features.
    Click here to sign up

Gallery photos: size limit, and should we allow AI content?

Taylor Nankervis said:
Well it's not.
Click to expand...

That doesn't particularly surprise me.
But as a regular watcher of television and cinema that originates from a wide variety of predominantly or frequently English speaking parts of the world - including Canada, England, Scotland, USA, Wales, Australia, New Zealand and even India - it never ceases to amaze me what terms are curses and slurs in other parts of the world!
Even though if I used them before those viewing experiences, I wouldn't even know I was being offensive!
 
You will hear it often enough in many parts of Ireland, Scotland (aptly descriptive of haggis...), Liverpool, far-flung Iceland...even decadent Kensington, London!

Enough. Gentleman Jims coming soon. Going to bitch the pot...
 
Let's ask the current most advanced ai:

"Is a 100% AI generated photograph a real photograph, if not why?"

------------------

Technically and philosophically, no, a 100% AI-generated image is not a photograph.

While the result may look exactly like a photograph ("photorealistic"), the process of creation fundamentally disqualifies it from the definition.

Here is the breakdown of why they are distinct media, ranging from the physics of light to the philosophy of truth.


1. The Etymology: "Writing with Light"​


The word photography comes from the Greek words phos (light) and graphé (drawing or writing).

  • Photography requires a physical event where light (photons) hits a light-sensitive surface (a digital sensor or film). It is a record of physical light rays at a specific moment in time.

  • AI Generation is "writing with data." No light is involved in the creation of the image. The AI generates pixels based on mathematical probability and statistical patterns found in its training data.

2. The Process: Capture vs. Synthesis​


The fundamental difference lies in how the image comes into existence.

Photography: The Indexical Record​


A photograph is an index—it points to a physical reality that existed in front of the lens. Even if a photo is heavily edited, it started with a "capture."
  • Input: Light reflecting off physical objects.
  • Mechanism: A shutter opens, light hits the sensor, and data is recorded.

  • Constraint: You must be physically present at the location.

AI Imagery: Latent Space Synthesis​


AI images are generated through a process called diffusion (in models like Midjourney or Stable Diffusion). The AI starts with "noise" (random static) and iteratively removes the noise to reveal an image that matches your text prompt.


Getty Images


  • Input: Text (prompts) and random noise.


  • Mechanism: The AI searches its "latent space" (a mathematical map of concepts) to hallucinate an image that statistically matches the request.


  • Constraint: No physical presence required; it creates a scene that never existed.

3. The Test Case: The Boris Eldagsen Controversy​


The distinction between the two was famously tested in 2023. German artist Boris Eldagsen submitted an AI-generated image titled The Electrician to the Sony World Photography Awards.

  • The Result: He won the Creative category.

  • The Reaction: Eldagsen refused the award. He revealed the image was AI to provoke a debate, stating:

    "AI images and photography should not compete with each other... They are different entities. AI is not photography. Therefore I will not accept the award."
    Click to expand...
Eldagsen proposed the term "Promptography" to distinguish the skill of prompting from the skill of photography.

4. The Gray Area: Computational Photography​


If AI isn't photography, why is it so hard to tell the difference? The line is blurring because modern cameras (especially smartphones) already use AI.

Feature Traditional Photography Smartphone (Computational) AI Generation (Midjourney/DALL-E)
Origin Light on Sensor Light on Sensor + AI enhancement Zero Light (Pure Data)
Subject Physical Reality Enhanced Reality Fictional Reality
Example DSLR Raw File iPhone "Night Mode" or "Portrait Mode" Generating a person who doesn't exist
  • The Distinction: A smartphone uses AI to optimize the light it captured. Generative AI invents the light entirely.

5. Legal & Copyright Status​


Legal systems validate this distinction. In the United States, the Copyright Office has ruled that 100% AI-generated images cannot be copyrighted.
  • Why? Copyright requires "human authorship." A photograph is protected because the human made creative choices (framing, lighting, timing). In AI generation, the machine does the "heavy lifting" of visual expression, breaking the chain of human authorship.

Summary​


An AI image is Photorealistic Computer Graphics, not Photography.
  • Photography is about taking (capturing a moment that happened).

  • AI is about making (synthesizing a moment that never happened).
Both are valid art forms, but they are as different as a painting of a sunset is from a photograph of a sunset.

Next Step​


Would you like to see a comparison of "Promptography" vs. "Photography" workflows to understand how the creative skills differ?
 
Sean said:
"Is a 100% AI generated photograph a real photograph, if not why?"
Click to expand...
I think the consensus is, and always has been on this forum, that AI-gen imagery is not photography. However, as I've said several times already, this is not the question at hand. With all due respect @MurrayMinchin and others who continue to debate this issue - I believe it's a waste of time. Nobody here disagrees with you. We don't have to establish that 'promptography' (hah!) is not photography. We already did that.
On a sidenote - no, @cirwin2010, chemigrams do not involve light. So we do have a precedent (albeit a very marginal one) in something other than photography being included integrally in the scope of this forum.

The assumption is apparently that Photrio is focused exclusively on photography as we know it, and that this scope is constant. The question at hand is whether we want that scope to be static, or whether we want to allow it to follow (to an extent) the developments in the real world. I notice a couple of arguments against this - arguments that have underlying sentiments, expectations and concerns that to a large extent I share. I would like to comment on a few of these.

In doing so, I want to emphasize that there's no concrete plan or proposal towards expanding the inclusion of AI in the forum as it is now. There is the subforum that @Sean already linked to. That's all we've got for now. A loose idea would be to have a similar thing in the Gallery where we could perhaps create a sub-Gallery specifically for AI-gen content, with that content not appearing in de sidebar etc. That's all we have in terms of a proposal - which is basically nothing. There's no proposal, no plan, no 'new deal'. At this point, all we have is an open discussion.

I notice that there's a stream of thought rising that the idea to "do something with AI" somehow traces to a desire for 'maximum engagement' (@retina_restoration, earlier also expressed by @Alan Johnson). I want to be very clear on this: the fact that we bring up this question is 100% responsive to what we see happening on the forum. The very recent trigger was the upload of some AI-gen images to the gallery. We have also seen previous threads with lengthy discussion on AI. Apparently there's an interest in AI - albeit that this interest is for many not necessarily a keen one. The tricky bit for us is to figure out to what extent within the present Photrio demographic there is a desire to discuss/view/create AI-gen content.

Gauging this potential interest is apparently tricky in itself because of the deep concerns among some of our participants. There's one concern I'd like to address:

"AI will flood the forum" or "corrupt the essence of the meaning of the forum" (@Taylor Nankervis on the latter quote).
I really don't think so for a couple of reasons:
1: We're not going to allow it under any circumstance. Everyone here recognizes the value of the in-depth discussion on Photrio, and nobody in their right mind would want to somehow harm that. So if we're going to put the door ajar to AI-gen content, it will remain limited to designated places in the forum and included in such a way that forum users can ignore it (e.g. by ignoring a subforum about AI and/or by deselecting specific tags).
2: This extends to our demographic. If we allow some form of AI-talk, this doesn't automatically mean Photrio will become a supermagnet for AI-afficionados. Looking at new registrations in the present day, we see mostly people joining us with an interest in (also) analog photography. Why? Because that's where we have critical mass. That balance is not going to flip overnight. Communities just don't work that way. They grow organically.
3: Given the scope of most of our sub-forums, it just doesn't invite any AI-gen content. If you have a thread on let's say how to best develop TMAX100, there's just no sensible reason to start ditching AI images into that thread. It turns out that people just don't do that. They don't start talking about the best recipe for chili all of a sudden. They don't start dumping AI images into a thread all of a sudden.
4: Conversely, if you make a dedicated place for AI-gen content, this acts as a shunt for whatever tendency/need/desire to discuss/show that sort of content. If any AI-gen content ends up in a non-AI place, we can always move it. Moving something is also generally appreciated more by the person who posted something than seeing their stuff deleted outright.

Overall, I think if we include AI in the scope of the forum, it will be much like how we deal with digital. It'll be a limited component, there will be some opportunity for people to play with it on the forum, and for those who don't want to be involved in it, it'll mostly remain hidden.

For now, we see a very vocal subset of the population voice their concerns. It's not clear whether that sentiment is representative of the majority, and whether we should even focus on a majority to begin with in any decision-making that may follow. The alt. printing crowd is a small subset of the population and we cater to them as well. This seems to work without being detrimental to other parts of the forum. We could do the same with AI.

Finally, I'd like to observe that the AI subforum that @Sean mentions has been around for something like a month now. People haven't even noticed. Nobody filed a report. Nobody cried wolf when it was created. Apparently, it can co-exist quite peacefully with the rest of it.
 

I have never been to Half Dome the knockoff sure looks magestic, has the park drilled holes to tell people where to put their tripod or stand?
 
TL;DR My point isn't that "AI" is good or bad or if it's photography or not. It's that it's close enough and also such a threat that to not include it and the discussions around it makes no sense.

I'm a "mature" student undertaking a Photography degree. I've used film for years from 35mm to 7x17" ULF and I've used digital also including medium format digital. The point of doing the degree is not to learn how, I know a lot of that already. The point is to learn why? With that in mind, Post Photography is a hugely fascinating area of photography that interests me. Mishka Henner's use of Google Street View and Google Earth is amazing. Most people would accept they are photographs, just not originally his. He didn't take them, he sat at his computer, for hours, searching for the compositions he wants, for the themes and projects he wanted and crafted a collection of images that suited. He edited those to produce his work.

(As an aside: 3D Google Earth are not "real" photographs, there are generated from a database of multiple data sources: Satelite Photographs, LIDAR, RADAR, GIS infomation).

Phillip Toledano created a project called We Are At War. It imagines the lost Capa images from D-Day, using "AI". Sacrilege I hear you cry. It's presented in the form of a newspaper, contact sheets and prints. What's the point of that? For me, it raises questions, as good art should do. About war, about truth, about photography, about AI. Importantly is not merely easily "AI" generated images that are meaningless. It's used in a considered, artistic context to prompt discussion.

For my Photography degree, I recently did a project producing physical postcards using "AI" generated imagery from extracts of W. G. Sebalds The Rings of Saturn with associated "AI" generated postcard text on the rear. I've submitted these to a traditional Art Gallery of some standing in my region and they have accepted them. To quote the curator "We particularly liked, though were somewhat discomfited by, the AI postcards.". I'm not claiming these are photographs, I'm just showing that people need to have an open mind and discuss where this kind of creative work sits.

The Rings of Hallucination

W.G.Sebald’s The Rings of Saturn is a seminal piece of writing. Ground breaking in it’s style and approach. To do it a complete injustice, it’s a narrative of a walk around the co…
andrewkeedle.com

Finally, Gemini can now flow an image through a series of "AI" edits. If the original was a photograph is the final result one also or not?
Original is B&W film with a Fujifilm GW670
Second is colourised using Gemini
Third is made to look like Autumn using Gemini



I don't know where the boundaries are. I just know that I need to try understand what it can do, what threat it is and also what use it is.

Just my 2 penny worth
 
Thank you @Andrew Keedle for your thoughtful response and the examples you've included. I assume some will argue that in these examples, actual man-made works of art are still the source and the AI-component is an extensive 'edit' of the man-made original - although your own project involving the Sebald novel emphasizes (at least for me) that we're talking about a continuum, and not a dichotomy.

I'd like to extend your line of reasoning with an observation of my own, without wanting to imply it necessarily follows from your thinking or that it's your conclusion - for me, it would be attractive to be able to discuss and look at AI works on Photrio because I trust that this community in particular will succeed in finding the gems in the mud. By that, I mean that in this community, I think the discussion will gravitate towards the fundamental questions around how AI relates to and photography in particular, and also that participants of this forum in particular will from time to time be able to locate or even create examples that really make us appreciate AI differently from how it's often typecast currently (i.e. as mass-produced homogeneous and low-quality 'slop'). Personally, I wouldn't want to go to an "AI forum" (if such exists) to learn more about AI, because I don't expect such a place will ask, let alone answer, the essential questions. I think Photrio is far better positioned in that sense, and I'm excited to learn more about it through the lens of a community that has always engaged at a profound philosophical and technical level with photography.
 
If one wanted to play devil's advocate, a large amount of the AI's training data is derived from light hitting a photosensitive surface. The distinction is perhaps that photography is not traditionally associated with millions of photographic instances being compiled together in a single image. A typical multiple exposure rarely contains more than a few separate shots. It's also combined additively, simpler than what AI does which could be described as some kind of morphing with categorization.

When you draw the line that the author of the work needed to at least start with an image they personally took with a camera, that at least requires that the light hitting a photosensitive surface has some relation to their personal action. Although this too can be destructively reduced to absurdity in argument - starting with your own image and manipulating it so heavily with AI that it essentially contains nothing original to you.

Probably what will end up happening is like what happened with the definition of pornography. "I know it when I see it." All technical quantifications on it will ultimately fail in some way.

Regarding what you say koraks about flooding the forum not being a problem, I agree, the moderators wouldn't let that happen, it's not a real concern in this case. It is a concern with some other websites that have less/no/different moderation.
 

Absolutely this point. I have no interest whatsoever in learning what "AI" can do in a creative context from "AI" Bros. Their view of world is warped beyond belief. I want to see and know what people engaged in a creative practice think of and how they use (or not) these new tools. Yes, there are huge issues to be resolved but if we are honest and look at the prevailing direction of recent court cases it's not going to be in our (independant creators) favour.
 
Hmm, have photographers already accepted the billions of images scraped (most unknowingly) by tech companies and used in their ai image gen tech? Was it that easy, and photographers have moved on without a care? Is scraping and training data not simply a new technique of copyright infringing? One that is so complex and abstract these companies can essentially get away with it. Who is going to look at an ai image gen and say, "Hey that's totally a photo bob took 10yrs ago, it's practically identical, look the same scene, even the same person's face". It's impossible. This is where things differ for me. 100% AI generative work could very well be cleverly cloning the work of others and we are feeding that beast. Using an AI editing tool to edit your own work feels like you are at least ethically fire walled from that (violation?). A tangent worth exploring maybe.

Also, what are the implications if someone on Photrio AI gens a bunch of work and it turns out to be a huge copyright violation? What are the implications of any website serving such material?
 
I do think it is "so complex and abstract these companies can essentially get away with it". I wrote a blog (for my course) on aspects of "AI", copyright, court cases. It is in no way exhaustive, or to be fair detailed or reasoned, but different courts have come to different rulings. None of these I think are precedent setting because each circumstance is complex and deal with different countries and cross border issues. Companies with large resources like Getty Images and Warner Music Group cannot get their cases prosecuted in their favour and/or have to compromise and come to some partnership deal to have some form of "protection". Whilst the "AI" companies have seemingly endless streams of fundings, whilst capitalists are in land grab mode, I feel this is only going one way. Am I happy about it? No. Is there anything I can do about it? No. I could not use their tools, but in my mind that puts me on the back foot with working out how to use them and at a disadvantage. I don't pay for these "AI" tools but I am indirectly supporting them by using them. It's a dilemma and we will all have our own views. I just want to be able to discuss those.

Legal and Ethical case studies

There are several areas we could tackle when looking at Legal and Ethical issue relating to photography. Copyright / Plagiarism Artificial ...
apkeedle.blogspot.com
 

The phrase "AI slop" is actually defined and in broad use, as a Google search shows many articles about use of this phrase from NPR, The Guardian, the World Economic Forum, a scientific paper from the University of South-Eastern Norway USN School of Business, and so on and so on. So let's not get alarmed about a reference to a phrase that is already in common use around the world.
 

I would expect the most advanced AI to include references to it's work...
 

Regarding value, consider this: we are already in territory where people are making Promptography images like this and selling them to the public. I’m not going to bother raising the question whether or not it’s Art. That’s not the point. However, the technology that makes this stuff is only getting better at faking realism. What do you suppose people are going to think when they are informed that they have paid a substantial sum for a “photograph” like the replica of the Half Dome shown here, and that 1) it’s not an actual photograph and 2) it is inaccurate in its depiction of a real place?
This is possibly the biggest problem with AI generated imagery - we are fast approaching territory where it will no longer be possible to tell fact from fiction in art. I’m not saying this isn’t interesting (and useful) technology, but in the hands of greedy, unethical people there is a huge opportunity to deceive and swindle, not to mention diluting the virtue of centuries of hard won skill by real artists, whose life’s work can now be emulated by prompting a machine with 2 or 3 sentences of instruction.

Nothing I say/do will change the course we are on, but I vote to keep this kind of imagery off Photrio, since it is clearly NOT photography, no matter how much you try to stretch or deform the definition of photography. A photograph is an image made with light, acting on light sensitive media. Boris Eldagsen gets it.
 
Last edited:

But how will people distinguish the real photos from the AI generated ones? On the other hand, pictures of families, friends and personal vacations and parties will still be photographed and shown around. But artsy type photos, landscapes, etc. will become passe except as a passing hobby for a very few. What would be the point of this site having a photo display section if AI-generated pictures are allowed there? What happens to the spirit of competition among photographers when a ten-year-old creates better images with a computer than a guy with 35 years of photography experience?
 
retina_restoration said:
Don’t feed the trolls.
Click to expand...

State your opinion if you disagree with others' views about this topic. Don't call them names. That's insulting. You're better than that.
 
I know for a fact that my Flickr portfolio has been “scraped” by AI to build their LLMs. (There are tools to find out if your work has been (ab)used)

Do I approve? No! Can I do anything about it? Sadly, no. I can stop publishing new work, but that won’t change a thing, and I’m the one who would suffer.

But this ingestion by LLMs is a violation of my IP Rights, and violates Flickr’s Copyright settings as determined by me. But I know that Flickr management is taking a very passive role in what’s happening (which I also know for a fact: I asked) - they know what’s happening to their users work, but there is no legal path forward to prevent this kind of abuse, so they’ve thrown up their hands (defeated?).
Addendum: I have done more research, audit appears that many content platforms (like Squarespace and YouTube) now include in user settings an “AI training permission” opt-out switch. Unfortunately Flickr doesn't have such a tool. When the time comes next year for me to renew my Flickr Pro account, I am going to have to think very carefully whether or not I want to continue paying Flickr for a service that is underperforming in this manner.

So to the managers of this forum I ask: is this a site about photography? Or will there be an asterisk on the end of Photrio in the future?
 
Last edited:
gbroadbridge said:
It's worth exactly what someone is prepared to pay for it, just like an AA photo.
Click to expand...

AI images like this will destroy the value of real photos that compete with it. The market will be flooded with awesome photographs that aren;t real. No one is going to pay much unless it's from a famous photographer. Just as digital cameras and photography ruined the market for many commercial photographers, this AI will continue the negative trend.
 
Agreed.
 
retina_restoration said:
So to the managers of this forum I ask: is this a site about photography? Or will there be an asterisk on the end of Photrio in the future?
Click to expand...

Site owner here. To be clear again as the thread meanders, there is no proposal by the site, so no 'vote' is needed.

My position is it can be discussed in this subforum "AI in Photography and Other Arts" or if someone wants to create a group in the social groups area they can. My position is that 100% AI generated 'promptography' is out of scope here, but not banned from this particular subforum. I think that gives any members interested in it the ability to discuss it within the community or post promptographs in a thread within this subforum.
 
You must log in or register to reply here.
Menu
Log in
Register