I’m not a huge fan of generative *ai* art. I think it’s something else. I am aware, however, that it is causing a lot of heartache for my fellow graphic designers. Companies are laying off design teams in droves. AI can do it cheaper and better. Wake up world, AI is not coming for your JOB it’s coming for your entire industry.
Art With DALL-e
OpenAI broke the news with DALL-e. But it wasn’t until ChatGPT shattered the public aura of *ai* that the business world began to sit up and listen to the robot chatter and LLM extractions and transformations. I was working for a company building AI-enabled search results with a great tool, Search Unity. Then as ChatGPT came online, I was heading up the creative and content team building a global website for the merger of a two data center owners into the largest owner, leapfrogging Equinix in number of data centers, but only 50% in market capitalization. So, that’s a bit of a brag that doesn’t hold water. But, at the time OpenAI unleashed this tool, I was building a 2,000+ page website, with a lot of repetitive content, a lot of summaries, and a lot of bad writing from authors and executives who were not writers.
Hello ChatGPT. With one other writer, we GPT’d the entire website in six months, to a celebrated launch in May of 2023. Yay all around. I learned a lot about the limits of ai-generated text. I learned how to prompt more clearly and edit with my own human-trained mind. The limits and leaps of ChatGPT were astounding. My writer and I would never have completed the website content on time without AI.
On the generative art side, we did not have as much call for images. I had been playing with DALL-e and Midjourney since they were released. But I was also a graphic designer with Photoshop chops and a rich history of painting, sculpting, and drawing. The ai art seemed interesting at first. But today, ai-generated imagery is boring to me. (I’m Done With *ai* Dreamy Graphic Designs for AI) Today, I’ve I’m deeper into DALL-e than any of the other ai design tools, and I wanted to share a bit of my process.
How Humans Can Create *Images* with AI
I don’t think ai can generate art. I do know of a friend who uses ai generated layers in his art. The human behind the curtain is more engaged than mere prompt-art, but I’m a bit dubious about the “art” of it all. I guess, for the purposes of this article, my (what is art) opinion is less relevant. Let’s begin with a simple idea.
My human brain was prompted by an image on LinkedIn. Here is the real-world prompt that struck a creative chord in my human LLM.
My mind enjoyed the neurochemical bath this image provided. Immediately, my logical side said, “Let’s do something like that with DALL-e. And here we are. Human to human, trying to understand generative ai from an artist’s perspective.
PROMPT NUMBER ONE: “liminal photograph of a drop of water making a splash in a still pool of water, dark, blues greens and few yellow highlights” I’m going to come back to the prompt tips in a moment, but first I just want to take you on my journey.
Okay, those are nice. Let’s do a variation round. I select my favorite drop and hit the “variations” button.
Let’s change up the prompt a bit. (Now, my human mind is jumping to all kinds of new ideas about what I want. I don’t want to recreate the initial image and creative impulse. I want to have a conversation with DALLe and do some human + ai art.
PROMPT NUMBER TWO: “liminal photograph fish eye view of a drop of water making a splash in a still pool of water, dark, blues greens and few yellow highlights“
Exploring a “eye view” filter that I learned from a friend. Here I try something new “fish eye view.” I think the last one (far right) get’s close. Perhaps I needed an apostrophe. “fish eye’s view.” Either way, these did not light up my imagination.
PROMPT NUMBER TWO: “liminal photograph bird’s eye view of a drop of water making a splash in a still pool of water, dark, blues greens and few yellow highlights”
PROMPT NUMBER THREE: “liminal photograph satellite eye view of a drop of water making a splash in a still pool of water, dark, rainbow lighting with fewer highlights“
PROMPT NUMBER FOUR: “liminal photograph distant star’s eye view of a drop of water making a splash in a still pool of water on Earth, dark, rainbow lighting with fewer highlights”
PROMPT NUMBER FIVE: “hyper liminal photograph of a drop of water making a splash in a still pool of water on planet Earth, dark, rainbow lighting with fewer highlights”
Pausing for everyone to catch up.
AI Art Prompt Tips and Ideas
Prompt number five is where I ran out of time for this little conversation with DALLe. But the wonderful thing about *ai* is you can pick up where you left off. But before I go deeper into my drop of water image, I want to give you some helpful prompt tips for generative art.
- In the style of – this is the big filter – “In the style of Dali” would produce quite different results than “in the style of Warhol”
- Bird’s eye view – try other “eye view” filters, go nuts
- Liminal space – this is a nuanced add that I’m exploring in other ways – liminal is roughly the moment between two major events, so the pause between the action
- Colors and highlights – I like to push into palettes, give the colors, or add NEON or PASTEL to the color
Prompts are easy to engineer if you follow a few routines. 1. The perspective of the viewer. 2. Styles of art. 3. Photorealistic or painting. 4. Type of media (crayon drawing, or impressionistic pastel painting) Then I apply “filters” to add, subtract, and dream with the ai of DALLe to get more ideas.
Let’s return to the drop of water image exploration. I have a few more ideas I’d like to try.
PROMPT NUMBER SIX: “hyper-realistic liminal photograph of a drop of water making a splash in a still pool of water on a distant planet Earth in space, dark, rainbow lighting with fewer highlights”
PROMPT NUMBER SEVEN a slight variation on five: “hyper-realistic liminal abstract pastel drawing of a drop of water making a splash in a still pool of water on a distant planet Earth in space, dark, rainbow lighting with fewer highlights”
PROMPT NUMBER EIGHT let’s change it up a bit more: “liminal impressionist pastel painting of a drop of water making a splash in a still pool of water , dark, rainbow lighting with a few highlights”
PROMPT NUMBER NINE I’m going to dumb it down again: “impressionist abstract painting a drop of water making a splash in a still pool of water , light, rainbow lighting with a few highlights”
PROMPT NUMBER TEN: “impressionist pastel a drop of water making a splash in a still pool of water , light, rainbow lighting with a few highlights”
Back to the Beginning with AI Art
Thus, we’ve almost come full-circle back to the original prompt. What you will begin to learn is how your mind jumps to prompts you’ve tried before, how it jumps to new variations of those prompts, and how the images AI creates spawn more ideas in your mind.
When you begin your journey with an ai-generative partner you will be trained to filters that work for you. Obviously, my “planet Earth in space” filter was ignored or not understood. The same thing happens when you stray to far out of the known masters when using “in the style of” prompts. In the end, I went back to basics. Sometimes simpler instructions give satisfying results.
Enjoy your personal journey training and being trained by generative-ai tools.
John McElhenney — let’s connect online
LinkedIn or Iterativ.ai
A few of my creative ai motions:
- waltwhitman.ai < a book of “whitmanesque poetry” with zero AI
- Radio No Head < *ai* parsed music for you to sing and dream with
- poətic.com < my page about the ai poetry conundrum
- hyper-soul < the science fiction exploration and world-building project
- clickthis.ai < my team for *ai* marketing and sales acceleration
Please check out a few of my books on AMAZON.
