If you’re an artist you’ve probably already played with the new generation of AI art tools: Dall-e, Craiyon, Midjourney, Stable Diffusion. They take your written description and turn it into an image.
Each AI is differently trained, and even the same prompt will yield different results. For example, these three set of images were all from the prompt “green 1895 ballgown” but each AI has a different sense of style, don’t you think? It’s similar to human artists with different styles based on their training and interests.
The art AIs generate can be fantastical, weird, creepy, delightful, and with enough refinement of the prompt, it can be very good. I am not at the “very good” phase yet; I’m just feeling my way around the different systems.
The process is fun and messy, a blend of creative awareness and logical adjustment. It feels a bit like painting, at least the way I paint – starting off with a broad idea and then seeing what happens when I put the paint on the canvas. I adjust from there, refining shapes and adding more detail. Working the AI art tools seems about the same, only there is a greater uncertainty over the images that come. But it is easy enough to give a different prompt to get something more in line with your ideas, and not nearly as time consuming as starting a new painting.
Here are a few of my tries with Midjourney and Stable Diffusion, captioned with the prompts I used:
Will AI art replace human art? In some cases, yes. Just like desktop publishing software and autotune and coloring books opened new fields to beginners, AI tools give a boost to people who aren’t trained in art. Those people might have hired me to make an illustration. Now they’ll make their own with Midjourney. That’s fine. No single artist or single tool can fill all needs. Human artists are still going to have work.
Some people will become professional AI artists who are skilled at crafting prompts that give the results that they imagine. Overall, there will be more artists, not fewer, as a result of these tools.
For me, the generated images are inspiring and eye-opening. I want to make that silver and red dress, or try to paint something like the orange cat/bird. I am pretty excited to play with these tools some more and maybe even get “very good” at them. It won’t hurt to be able to offer my clients a different type of illustration that the AI and I collaborate on.
There are beautiful times ahead.
