By
Steve Kuss
I will start with a “plagiarized” paragraph for you:
While the rule of thirds is a commonly used guideline in photography, there are times when deviating from it can create a more impactful image. Overusing the rule of thirds can lead to predictable and boring compositions. Certain subjects or situations may benefit from different types of framing or symmetry, and breaking the rule of thirds can help achieve this.
Sound good? I always try to spend more time talking about breaking rules than keeping them, as far as photography is concerned. It would seem that the author of the above paragraph agrees. That author? ChatGPT, an AI bot. Huh?
Okay, I went to my computer, signed into a webpage called ChatGPT, and typed in, “Write a 3-sentence paragraph about why one should sometimes not use the rule of thirds in photography.” That’s one of the options I got back. It was written by what is called a “chatbot” or simply, a text-generating program run on a very high-powered computer. The “bot” part is just a nod to “robot.” There’s really nothing magic here. AI (artificial intelligence) is still nothing more than a very sophisticated computer program. It’s another tool that modern people have at their disposal. It is not the death knell for…well, anything, really. It still requires input—a human has to operate it to make it do anything.
I looked at Midjourney.ai and Leonardo.ai, AI bots that use words to create images. One of the “pro” techniques of a photographer is to “pre-visualize” your image. Decide on the photo you want to create, then, and only then, point the camera and shoot. I tried that on the AI bots. I thought of an image I wanted to create and described it as succinctly as possible. As you might imagine, my pre-vision was not the same as the bot’s. In the end, the blue fractal image that I got out of Midjourney wasn’t bad, but it took many iterations of carefully selected words to get there. Fractals are very complex mathematical structures that can take hours for a computer to produce. What you see here was produced in less than a minute and so is an approximation, not a true fractal—a drawing of a fractal, if you will.
One of the features of Leonardo is that it can be “seeded.” You can give it a starting point from which to build the desired image. I gave Leonardo the Midjourney fractal as its seed. Again, it took a few iterations, but I eventually got something fairly interesting.
I then asked Leonardo to make a logo for the Cookeville Camera Club (CCC) with a fractal background. It pretty much took all of my allotted “tokens” (the tracking means to prevent people from hogging all the computer time for the free service) to get anywhere near close to what I wanted. Most logos looked like college shields, and nothing really was up to my vision. Leonardo seemed to think that my “prompt” (the text phrase used to describe my desired image) for “three C’s” was only a suggestion and frequently only used one or two, sometimes mixed in with other letters.
So, in the end, AI was fun to play with, certainly can be a useful tool, requires developing skills to use it effectively, and, no, it will not take over the world of photography, or any world for that matter, any time soon.