The Pixel Genesis: How a Painter in the 1970s Created the World’s First AI Image
We live in an era where typing “a cat wearing a spacesuit riding a skateboard” into a prompt box yields a hyper-realistic masterpiece in under five seconds. Generative artificial intelligence has flooded our feeds, transforming text-to-image tools into everyday utilities. Yet, looking back at the long evolutionary chain of digital creation reveals a surprising truth: the first-ever AI image wasn’t born from a massive neural network or trained on billions of scraped internet photos.
Instead, it was conceived in the early 1970s by a celebrated human artist who decided to teach a computer how to see.
Meet Harold Cohen and AARON: The Pioneer of AI Art
Long before GPUs, diffusion models, or deep learning, there was Harold Cohen. In the late 1960s, Cohen was a prominent British painter who had represented Great Britain at elite global exhibitions like the Venice Biennale. He stood at the pinnacle of the contemporary art world.
Yet, Cohen found himself consumed by a radical, almost heretical question: What fundamentally makes an image evocative? How do human brains look at a collection of arbitrary lines and dots on a page and instantly recognize a face, a figure, or a landscape?
To answer that, Cohen took a sharp pivot that shocked his peers. In 1968, he left London for the University of California, San Diego, taught himself computer programming, and began writing a software program that would change art history forever. He named it AARON.
Inside the Mind of the First AI Artist
Created around 1973, AARON was vastly different from modern AI. Today’s models rely on statistical probability and massive data crunching. AARON, on the other hand, was built on symbolic reasoning and cognitive rules.
Cohen didn’t give AARON a dataset of thousands of pictures to copy. Instead, he painstakingly coded his own deep understanding of art into the program line by line:
- Fundamental rules of perspective and geometry.
- The mechanics of how objects overlap and cast shadows.
- Principles of composition, balance, and human/plant anatomy.
By feeding these structural rules into the machine, Cohen enabled AARON to make autonomous compositional choices. When the computer drew a line, it wasn’t just blindly following a pre-drawn template; it was calculating its own placement based on the logic Cohen provided.
The First Images: Abstract, Intentional, and Revolutionary
The earliest iterations of AARON produced minimalist, abstract black-and-white line drawings that reminiscent of primitive art or petroglyphs.
To physicalize these creations, Cohen built specialized robotic drawing machines attached to the computer. These mechanical arms gripped a pen and meticulously dragged it across paper, bringing the computer’s internal calculations into the physical world. For the first time in human history, an intelligent system was autonomously executing creative works.
As the decades progressed, Cohen continuously upgraded AARON. The program evolved from drawing simple shapes to generating complex human figures, lush jungle foliage, and eventually, vibrant color paintings that were exhibited in world-class institutions like the Tate Modern and the Whitney Museum of American Art.
Why the First AI Image Still Matters Today
Harold Cohen passed away in 2016, having spent nearly five decades in a collaborative partnership with his digital creation.
Looking back, AARON serves as a humbling reminder of what generative art is truly about. While contemporary tools focus on speed, hyper-realism, and scale, AARON was born out of a philosophical quest to understand human cognition and artistic intent.
Leave a Reply