February 26, 2023
Is An AI Generated Photography…A Real Photograph?
Midjourney by Jason Allen (and AI)
It was recently revealed that an award winning photograph was actually AI generated. I was discussing this with my son Caleb, who is majoring in Computer Science, and the question arose: is an AI generated photograph, a real photograph?
A day later, Caleb emailed me and said that he had posed that question to ChatGPT, which is an AI chatbot (think of the computer HAL in 2001).
Here is AI’s response to the question:
The answer to whether a photograph created by AI is a “real” photograph can be somewhat subjective and depends on how we define “real.” From a technical standpoint, a photograph created by AI is a digital image file that can be viewed and shared just like any other digital photo. However, there are some important differences between photos created by humans and those created by AI.
One key difference is that photographs created by humans are typically the result of a deliberate and intentional act of capturing an image of a real-world subject through a camera lens, whereas AI-generated images are created through an algorithmic process that may not involve an actual subject or a physical camera. This means that AI-generated images may not necessarily represent something that exists in the physical world.
Another difference is that photographs created by humans are often valued for their ability to capture a moment in time and convey a sense of reality or authenticity, whereas AI-generated images may be valued more for their esthetic or creative qualities. In this sense, the concept of a “real” photo may be less relevant for AI-generated images.
Ultimately, whether a photograph created by AI is considered a “real” photo depends on the context in which it is used and the purpose it serves. In some cases, an AI-generated image may be just as valid and valuable as a photo taken by a human photographer, while in other cases, the distinction between a “real” photo and an AI-generated image may be more meaningful.
Wow, not a bad answer, but it’s a little creepy being referred to as a “human” by a computer.
This week I had friend write an article for his Club’s newsletter, in which he talks about how he described an image to DALL-E (an AI system that can create images) and he then shows the result:
Find Your Vision
Thanks to the amazing job Serge does as the Programs Director for WPS we have one spectacular speaker after another at our meetings. And one of the positive side effects of the COVID-19 pandemic has been our Zoom meetings where we have been able to invite speakers from geographically distant places.
Last night’s program knocked it out of the park again: Cole Thompson came to us from his home in Colorado and gave an awesome inspirational talk illustrated with many of his projects.
He talked about finding your vision and not listening to what other people think your vision should be, or what “rules” you should follow, and that certainly one should not copy someone else’s vision.
He shared an anecdote from his early days when his photography imitated Ansel Adams. A critic told him “Ansel already did Ansel. What can you do that exhibits your unique vision?” That led to a sudden epiphany, and as he put it, he didn’t want to be known as “the world’s greatest Ansel Adams imitator”.
Lately generative AI has been in the news everywhere. ChatGPT has been all over the mainstream press, most recently on the cover of the current (February 27th) issue of Time Magazine. Before that, everyone was talking about OpenAI’s DALL-E 2.
While I certainly don’t want to be the world’s greatest Cole Thompson imitator, inspired by him, I thought I’d have a little fun with DALL-E and see what I could do. So with the prompt,
“A moody high contrast black and white portrait of a cat in the style of Cole Thompson with long exposure motion blur clouds and ocean in the background”
Meow, Sitting for Portrait
Scary, huh?
So, what’s your vision? I hope you share some of your images in our Member Showcase meetings, or even here in our newsletter. Till next time, may you always see beauty in your viewfinder.
–Fuat Baran, President (Westchester Photographic Society)
AI clearly doesn’t do cats very well (unless it was portraying the cat in long exposure?) Nonetheless, this is a fairly impressive and yet scary image, that gives a hint of what’s to come.
AI is a hot topic, and I doubt it’s going away. I suspect my views on it will evolve with time, but for right now, I would not feel good about creating an AI generated image and then calling it a photograph or calling it “mine.”
Hello Cole! In its answer, the machine conveniently left out an important part: that the generated artwork is ENTIRELY based on work created BY humans, and which had been used to train the AI without the original creator’s explicit consent, in most cases…
I had my own thoughts about AI-generated, derivative art, a while ago. I’m adding the link via the comment form. 🙂
Hello again Cole, thanks for telling me that the link to my article is rather obscure. 🙂 Here is it: https://www.alex-kunz.com/exploring-different-worlds/
Hi Cole, I have played around with AI a bit (MidJourney & DALL-E) just to get a feel for capabilities surrounding all the hype and excitement.
In my opinion AI in its current form is really just a new medium, and cannot be passed off as photography. As I understand it AI algorithms pull information from databases to satisfy the prompts they receive. They use existing images and tags to create results (images taken by humans I dare to add), and are unable to create anything unique without that knowledge. If that data does not exist (eg my pet dog, Winston) then AI does its best with what it has (in this case any pet dog).
I found it easy to create astounding images, but I could not say that they were anything I had imagined in my own mind, or successful in recreating scenes I referenced as a starting point.
At some point I see the applications growing and faithfully able to capture the essence of referenced images, but AI is not at that stage now.
Like the introduction of photography, just as photos can’t be labelled paintings I see AI as becoming a distinct artform in itself, and adding more value to the more personal photography that went before.
BTW – coincidentally I am re-publishing a blog entry on your work later today after migrating my website. I believe you OK’d the blog on my previous site but let me know if you’d like me to take it down.
Well, you are famous enough for AI to recognize your style and create something alike. AI-generated images are here to stay and as with many other things we do not have a saying on it. They still don’t do cats or hands well, but will get there. Cheers.
You cut right to the chase Alessandra: we don’t have any say in the matter!
Yes, AI is here to stay, but we still have a choice as whether or not to use it. Contests, publications and galleries also have a choice. I came across a contest which has a separate category for AI images. We’ll soon see!
Hi Cole,
I don’t agree that AI images are photographs.
They are digital art.
If you look up the definition of photography it states that a camera is used in the process. No camera in sight with AI – only algorithms.
Also the word photography comes from a Greek word and the process involves capturing light.
No light is being captured here.
Cheers, Vic.
For me, an AI image doesn’t come from your heart and soul. It’s just not “me”.