AI Art Challenge #4, 2026 "From One, Many"
Thanks to Danni @danni-gurrl for the nostalgia. I haven't used this approach in a while. It's interesting to look at it from a new perspective.
I found my style a long time ago. I move within it. It's a paradox, but whether in AI or "analog, real" drawing, I always return to where I started. No matter what I try or how I experiment. Because I like it from the start. I feel it.
Why is the work "Out-of-competition"?
I do what I enjoy. I'm not trying to please others.
I'm simply sharing what I know and can do.
If I want a competitive spirit, I always have places for that.
My criterion for evaluating AI work is recognition. If I recognize the artist's style, then this picture is here.
Promt was the same everywhere.
Only the artist/style were different
I don't know what to comment on here. Only Aivazovsky himself could convey Aivazovsky's style better than Bing.
This is my favorite artist. I "feel" his paintings. I've seen them live many times. They don't just convey a real atmosphere. They are more than real. They are tangible.
The Aivazovsky-style painting was the first image I generated using AI.
While my main criterion when working with Aivazovsky was the landscape, with Goya I focused on character. I think the girl captured the spirit of the era depicted in Francis's paintings.
I'm also very pleased with Bing's work here.
As well as with Salvador Dali's works.
It's understandable that AI has difficulties working with such styles. But nevertheless…
I found it odd that Bing had such difficulty recognizing the style of the founder of abstract art. On the other hand, Kandinsky's paintings are extremely diverse, and it's truly difficult to formulate a "template."
I tried to choose what best matches his style.
The DreamUP AI worked perfectly with Kandinsky's style. When this AI worked.
I cosplayed one of his works for a challenge on DeviantArt.(
Since Bing learns from images it's just generated, including similar prompts, at a certain point the styles started to blend. It was just an abstraction, with no author or style, no matter what I specified in the prompt.
AIs recognize many artists. But not all. And different AIs recognize different artists. For example, Monet's style didn't work. And Rembrandt didn't work because Bing learn from previously generated images. It wasn't Rembrandt anymore. Kandinsky's style was once perfect and instantly recognizable in DreamUp. That's no longer the case.
As for "non-classical, non-realistic styles" (abstraction, surrealism, etc.), things are a bit more complicated.
The AI usually draws the main character in a standard style, or as close to realism as possible.
But the background and landscape are quite a decent match to the specified style.
So, in this case, to fully replicate the style of such an artist, it's necessary to completely rework (simplify, remove details) the prompt. Or create a new one.
(You won't get a painting in the style of Malevich if you add "Malevich" to the current prompt)
Moreover, this was absolutely not part of the initial assignment.
I haven't used any other AI for this task except Bing on Dall-E3 and DreamUp. But I don't think anything will change radically without changing the prompt.
Thanks to everyone who scrolled to the end :)
I wish you creative success!
@alyssa-ai @ai-satin-chic @anderiix @gigiprinceton2 @gigiprinceton @danni-gurrl @mohairmaster @astrogurlnikkipinkai @fluffyfaza @sarahwellshunter @mentorcreation @aifineart @patrickdreessensworld @jamiejason2-0 @softsmooth69 @miphisticated @synthialove