The History of the First AI-Generated Image Explained

The Pixel Pioneers: The Fascinating Story Behind the World’s First AI-Generated Art

When we think of AI-generated images today, we usually picture typing a quick prompt into a tool like Midjourney or DALL-E and watching a hyper-realistic masterpiece materialize in seconds. It feels like magic, a modern phenomenon born in the 2020s.

However, the history of artificial intelligence dreaming up visual art goes back much further than you might think. Long before neural networks, graphics cards, and massive internet datasets, a lone artist and programmer began building a computer program that would change creative history forever.

This is the story of AARON, the world’s first AI artist, and the dawn of computer-generated imagery.

Meet Harold Cohen and AARON

In the late 1960s, British painter Harold Cohen was a successful, internationally exhibited artist representing Britain at major global art events like the Venice Biennale. But working at the University of California, San Diego, Cohen became fascinated by an entirely different kind of canvas: the computer.

Cohen wasn’t a computer scientist by training; he was a painter trying to crack a philosophical code. He wanted to understand what made an image an image. What are the fundamental rules humans use to recognize a shape, a line, or a figure?

Instead of writing code that simply repeated a pre-programmed drawing, Cohen wanted to build a system that could make autonomous creative decisions. In 1973, that vision materialized into a software program named AARON.

How the First AI “Painted”

AARON didn’t work like modern AI. Today’s image generators are built on deep learning, meaning they “learn” by scanning billions of images created by humans across the internet.

AARON had no internet, no database of millions of photos, and no neural network. Instead, AARON operated entirely on manually coded rules written by Cohen himself.

Think of it as a set of instructions written in computer code that translated to artistic logic:

• How do objects overlap in space?
• How do you draw a human figure or a plant without it looking distorted?
• How do you balance the composition of a page?

Cohen taught AARON the rules of visual art, but he allowed the program to execute them on its own. When AARON ran, it would output raw data coordinates. Connected to a custom-built mechanical drawing machine (a flatbed plotter with a pen), the computer would physically guide a pen across a sheet of paper, slowly drawing stark, mesmerizing black-and-white line art.

For the first time in history, a machine was creating original visual compositions—making independent choices about where lines should go and how shapes should interact—without a human hand guiding the pen in real-time.

From Simple Lines to Global Galleries

Over the decades, AARON evolved. As computer technology advanced, Cohen continuously upgraded his digital creation.

• The 1970s and 80s: AARON mastered abstract shapes and simple geometric figures.
• The 1990s and 2000s: Cohen taught AARON how to handle color, introducing liquid paint systems where the computer could pick colors, mix them, and fill in its own outlines. Later versions even generated complex, organic jungle scenes populated by stylized human and plant figures.

AARON’s work grew so respected that it broke out of computer labs and entered the upper echelons of the art world, with exhibitions displayed at prestigious institutions like the Tate Modern in London and the Whitney Museum of American Art in New York.

Why the First AI Image Matters Today

Harold Cohen passed away in 2016, but his life’s work laid the philosophical and technical groundwork for everything we see in generative AI today.

When we marvel at modern text-to-image generators, we are standing on the shoulders of pioneers like Cohen who dared to ask a radical question: Can a machine be an artist? AARON proved that creativity isn’t strictly reserved for flesh and bone—it can be broken down, understood, and mirrored by code.

The next time you prompt an AI to create a futuristic cityscape or a surreal landscape, take a moment to tip your hat to a clunky computer plotter from 1973, slowly dragging a pen across a blank piece of paper to draw the very first lines of a digital revolution.