I have one of my AI pieces in a small gallery in Brooklyn coming up soon. Here is the info:
Web link: https://awomaninthearts.com/
This is in fact the first time one of my new AI pieces will be shown in public.
William T. Ayton is a British artist based in Tempe, Arizona & New York's Hudson Valley. His work deals with the human condition, social issues and myth. He creates drawings, paintings & augmented reality pieces, and has recently started to experiment with AI-generated art.
I have one of my AI pieces in a small gallery in Brooklyn coming up soon. Here is the info:
Web link: https://awomaninthearts.com/
This is in fact the first time one of my new AI pieces will be shown in public.
This was intended to be a throwaway piece...I was showing my nephew, Matt, how Leonardo.Ai works, as he is also in the AI field, though not from an art viewpoint. I just quickly went through some steps to show him various parts of the interface & this came out...which I like. It's an idyllic scene in the mountains with an improbably blue-domed palace, a paradise image of sorts, perhaps. Now I think it's a keeper.
The Blue Palace AI, 2024, made with Leonardo.Ai |
This is a series that I wasn't planning to do, but it happened organically when I started experimenting with Leonardo.Ai's Universal Upscaler, which upscales (embiggens) but also adds in more than just interpolated detail. It really does very well working at a vast scale, although it sometimes gets carried away with itself, adding bizarre details (plus the usual extra limbs and the like), so I have to edit it from time to time. Below is a selection, you can see them all at: https://www.flickr.com/photos/billayton/albums/72177720318233270
The themes covered include the usuals: life, death, heaven, hell, good, bad, the past, the future, mythology, reality, and of course, our place in the universe.
A New Utopia, 2024, AI generated image |
Final War, 2024, AI generated image |
The Age of Enlightenment, 2024, AI generated image |
Exodus, 2024, AI generated image |
In Hell, 2024, AI generated image |
Life, the Universe, and Everything, 2024, AI generated image |
So I created a Bosch-style triptych in AI, with a central panel of a dark figure carrying a pink sphere, & the two outer panels depicting Heaven & Hell...on the left is the fountain of eternal life, & on the right is a tower over a fiery pit...two different gifts.
AI Bosch, The Gift, AI triptych, 2024 |
I was creating so many Bosch-inspired images that I've decided to create a blog on Tumblr specifically for them at: https://ai-bosch.tumblr.com/
BoschScape, AI generated image |
With this one, I used Leonardo.Ai's Alchemy Upscaler to improve image quality & upsize it. I'm not sure it made a huge difference, as all the little figures are quite small, so fixing hands & heads, etc. is not a major issue here.
Bosch AI Apocalypse v.2 |
I feel that this one ought to be in a low-budget horror movie...Hollywood, are you watching?
The Creature from the Toxic Lagoon AI |
I've found repeatedly that Leonardo.Ai does a pretty convincing & effective impersonation of Hieronymus Bosch, and its tendency to mess up anatomy (too many legs, etc.) only helps. Here's the Apocalypse:
AI Bosch Apocalypse |
In this AI-generated image of Leda and swans, how many swans can you count? Partial swans are not to be included...
Leda and Swans, 2024, AI-generated image |
This is another image that wasn't exactly what I was looking for when I generated it. However, upon reflection, I decided it was worth uploading here, as opposed to the Tumblr blog where my more "finished" AI pieces live.
Annunciation, 2024, AI generated image |
Another oddity/experiment in Leonardo.Ai. With this one I extensively repainted and regenerated the original image until I had this. The original had a sea urchin on a man's head...
The Hat with the Eye, 2024 AI generated image |
In case anyone's wondering if I still make non-AI artwork, the answer is a definite "yes". Here's one from today:
Lost Among the Spheres, ink drawing 14" x 11", 2024 |
I typed in "the man who was a nuclear weapon" and got this faux movie poster with questionable spelling. You never know what you're going to get...coming soon to a movie theater near you...
The Man Who Was A Nuclear Weapon / AI generated image |
This is not exactly what I wanted, trying to generate a winged head of the Greek god Hypnos, but I liked the resuts anyway:
Still Life with Hypnos variation AI |
Still Life with Hypnos variation AI |
Still Life with Hypnos variation AI |
Still Life with Hypnos variation AI |
Here's an odd experiment with Leonardo.Ai--I was trying to generate a version of Bosch's Tree Man from the right-hand panel of the Garden of Earthly Delights--but instead I got this:
Breughel's Tree / AI image |
The Man Tree / AI image |
Here is a screenshot of a WIP in the Leonardo.Ai interface (Realtime Canvas section), to demonstrate that AI image generation is slightly more complicated than some people might think. In this case, it's a combination of a text generated image with more text input, plus I've painted extensively on the image to change the shapes and appearance of objects. It's a refining process which can be quite lengthy...AI is really a collection of power tools that enhance creativity.
Leonardo.Ai screenshot |
And here is the finished piece:
Light Among the Ruins AI |
Leonardo.Ai did a rather good job of creating these epic images of the Tower of Babel in flames. As you may know, I've depicted the tower in many pieces, but these came out very cinematically, and, as I mentioned, epic.
Battle for Babel AI #1 |
Battle for Babel AI #2 |
Battle for Babel AI #3 |
Battle for Babel AI #4 |
This is a case where the text prompt turned up quite amazing images with little to no editing required...below are 2 versions of the image.
Airship Battle (AI) |
Airship Battle 2 (AI) |
Here's another image I've been playing with, a Narcissus variation. I used Leonardo Ai's "Canvas" feature to interactively draw parts of the image, & also used text prompts to define it.
Narcissus in the Pool AI |
Sometimes, using Leonardo Ai, it manages to achieve good results very quickly. Other times, I have to give up on it & assume it doesn't understand what I'm trying to evoke. In the case below, it did a pretty good job, all 4 of the original images generated were potentially usable. This one could ideally use a bit of tweaking, the giant foot of the subsidiary sculpture stands out, as does the inconsistency of sailing vessels depicted...but that's part of the fun of AI.
Colossus Variation (AI) |
I've just created a new Tumblr blog to hold my AI artwork. You can find it at aytonai.tumblr.com. I'll be uploading more work as time goes on. Check it out.
Homage to Bosch version 1 (AI generated image) |
Here is another AI experiment from Leonardo.Ai. I find that it takes many edits to get a good image, so while this is in many ways quick & automatic, there are a lot of wasted images. I also find that the AI responds quite well to poetic or "nonsense" prompts...
End of Empire, AI-generated image |
My first successful (in my opinion) experiment with AI image creation--I used the Leonardo AI web interface to create this image of the Empire State Building in NYC attacked by monsters. It's the closest image I've created in AI that looks like I might have made it myself. More coming, I hope...
Empire State Monsters, AI generated image |