"Cartooning is for people who can't quite draw and can't quite write. You combine the two half-talents and come up with a career." — Matt Groening
Using AI for art is an exploration, not an act of creation.
You do not create with AI, you search --- and zero-in on an idea. Hopefully, you do this, while leaving yourself open to the discovery of new ideas along the way. To use AI for art is to explore infinite possibility. To find your way, you must learn to map the latent space* you are searching.*
Imagine a vast landscape where every possible idea exists somewhere as a point on a map. This is "latent space" – the linear algebraic territory that AI navigates when creating images, text, or other work. It's where new mixtures of established concepts live before someone brings them to life.
Today's AI tools force us to navigate this landscape using only text commands. This writing of what we wish to have AI generate, termed “prompting,” is the clumsiest the UX for using AI will ever be. It's like trying to find a location by shouting directions out the window. The evolution of these tools are inevitable.
What if artists could explore this latent space visually?
What if we could triangulate where we want to go, and calculate ever finer approaches to the ideas we wish to explore?
Latent space might be billions of multi-vector possibilities, but in essence, it is just a massive space to explore. We could simplify the concept to three vectors and a small triangular area.
Bear_Rabbit_Robot is a demonstration of this approach for artists. Instead of relying on words alone, it uses a simple triangle to map creative possibilities – giving artists a visual way to explore the spaces between concepts.
The project places three character types at the corners of a triangle:
Robot at the top
Bear at the bottom left
Rabbit at the bottom right
Each point inside this triangle represents a unique blend of these three concepts. By selecting 21 specific points in this triangle, we created a collection of characters that shows how these elements can combine in different ways.
What makes this approach powerful isn't just the mathematics of blending. It's that each point represents an artistic judgment about what combination "feels right." When we place a character halfway between Robot and Bear with minimal Rabbit influence, we're making a statement about what we consider a harmonious blend of those elements.
Each position in the triangle represents a creative decision.
It’s mapping the algorithmic blend of the AI model’s latent space, to the artistic instinct that dictates what we, as humans, would determine to be the blend between the elements. This is merely 21 images with three vectors of concepts. Imagine the control of this thinking blending 100’s of thousands of ideas together.
In navigating latent space, we can draw inspiration from the sextant—the tool sailors used to find their way by measuring the angles between stars. Our triangle method works the same way: by establishing three clear reference points (Robot, Bear, and Rabbit), we create a map that helps us navigate creative possibilities. We don't need complex math to understand that moving between these points creates new combinations, just as sailors didn't need to understand astronomy to find their position at sea.
While data science typically focuses on solving problems, artists can borrow these approaches for a different purpose: discovery. The tools mathematicians use to find patterns—like mapping distributions or finding relationships between variables—can help us explore creative territories beyond what we could imagine on our own.
You're a DJ with three turntables.
One plays pure "Surreal" beats, another drops "Minimalist" tracks, and the third spins "Baroque" melodies. Your magic happens when you blend these sounds – not by random mixing, but through your unique sense of what sounds amazing together.
That's the concept. Latent space can become your creative mixing board.
Let’s use Rabbit_Robot, with the slightest hint of Bear. This isn't some soulless formula – I chose the instance to use in that spot.
Is this a precision-engineered bunny with just enough bear-like sturdiness to give it presence? Or perhaps a lightning-quick mechanical being with a touch of wilderness in its circuitry? Does it feel right to you or a certain audience? Who makes the judgement about what a Robot-Rabbit looks like and how much is just a “dash of Bear?”
As AI becomes more integrated into creative workflows, your greatest advantage will be your unique artistic sensibility – your personal judgment about what combinations feel right. The triangle method doesn't replace this judgment; it gives it a map where it can be expressed and taught. I am sure there will be more of these kinds of frameworks nearing infinite complexity.
As AI tools evolve, the artists who thrive won't be those with the most technical knowledge – they'll be the ones who can most clearly map their creative instincts. Think of it like teaching a new assistant: You wouldn't just give random orders; you'd help them understand your taste and preferences.
User experience design has already given us powerful tools for creative exploration – scroll bars, toggles, bezier-curves, and eye droppers became the iconic controls of the previous generation of digital art.
But mapping latent space will usher in an entirely new breed of tools that let artists blend and morph through their AI-powered creations with the same intuitive ease. The triangle is just the beginning – a simple demonstration of what's possible.
Your artistic sensibility – the judgment that says "this combination works but that one doesn't" – is your superpower in the AI age. The mathematics doesn't replace this judgment; it amplifies it by giving you a playground where your creative instincts can run wild in organized ways. No matter how complex the system for searching might become, it will always be driven by your instinct to find the next great idea.
So grab your metaphorical sextant and start mapping. The technology industry is waiting for the artists to show them what’s possible.
Place your three favorite concepts at the corners of a triangle and just start. What undiscovered ideas live at the uncharted coordinates? And maybe ask, what happens when you blend concepts that nobody thought to combine before?
I am looking forward to seeing what you discover.
Nye Warburton is a creative technologist and educator from Savannah, Georgia. The Bear_Rabbit_Robot experiment was imagined through improvisational concepting with Deep-Seek R1 . Imagery was created using Leonardo.ai and Stable Diffusion. Writing was mainly human labor, with editing assistance from Claude Sonnet 4.5. For more information visit https://nyewarburton.com
