Before Midjourney: The Fascinating Story of AARON and the First AI-Generated Images

When we think of AI-generated images today, our minds instantly jump to hyper-realistic photos, slick digital paintings, or whimsical designs whipped up in seconds by text prompts on platforms like Midjourney, DALL-E, or Stable Diffusion. We live in an era where typing a few words yields breathtaking visual art.

However, the journey of artificial intelligence creating art didn’t begin in a high-tech Silicon Valley startup last decade. It started more than fifty years ago, born from an unlikely marriage between computer science and fine art.

The story of the world’s first AI-generated image is a tale of a pioneering artist, a room-sized mainframe computer, and a software program named AARON.

Meet Harold Cohen and AARON

In the late 1960s, British-born artist Harold Cohen was at a career crossroads. A respected painter who had represented Great Britain at the prestigious Venice Biennale, Cohen took a bold leap: he moved to the United States to teach at the University of California, San Diego.

Fascinated by what computers could do, Cohen did something radical for an artist of his era. He decided to learn computer programming.

Cohen wasn’t interested in using a computer merely as a tool to draw lines digitally. Instead, he wanted to understand the cognitive side of drawing. He wanted to know: What are the fundamental rules that govern how a human mind translates a three-dimensional world onto a flat surface?

By 1973, working extensively at the Stanford Artificial Intelligence Laboratory, Cohen birthed a software program called AARON. AARON would become the world’s first true artificial intelligence art system.

How the First AI “Created” Art

If you compare AARON to modern AI, the differences are staggering. Today’s models use massive neural networks trained on billions of scraped internet images, relying on statistical probabilities to turn text prompts into pictures.

AARON had no internet, no neural networks, and no database of scraped photos. In fact, it didn’t even use a text prompt.

Instead, AARON operated on symbolic, rule-based AI (often referred to as GOFAI—Good Old-Fashioned AI). Cohen manually coded human artistic logic directly into the software. He taught AARON the rules of:

• Geometry and proportion: How shapes relate to one another.
• Occlusion: How one object overlaps and hides part of another.
• Composition and balance: How to arrange elements harmoniously across a canvas.

Once the program booted up, AARON made its own independent creative decisions. The computer decided where to place objects, how to compose the scene, and what lines to draw next.

Because early computers couldn’t display complex graphics on a screen easily, AARON was physically connected to a custom-built, turtle-like drawing machine. This robotic arm held a pen and crawled across large sheets of paper spread out on the floor, meticulously executing the lines that the AI computed in real-time.

Did you know? The earliest versions of AARON produced abstract, black-and-white line drawings. Over the decades, as computer technology advanced, Cohen continuously updated AARON. The AI eventually learned how to handle color selection, shading, and even how to paint complex organic shapes like rocks, plants, and human figures.

From Museum Walls to the Digital Age

AARON’s creations were revolutionary. They weren’t just viewed as quirky computer printouts; they were recognized as legitimate fine art.

Throughout the 1970s, ’80s, and beyond, AARON’s artwork was exhibited at world-renowned institutions, including the Tate Modern in London, the San Francisco Museum of Modern Art, and the Whitney Museum of American Art. Art critics marveled at the fact that a machine was exhibiting an independent artistic style.

Harold Cohen continued developing and collaborating with AARON for over 40 years until his passing in 2016. Because AARON was an extension of Cohen’s own mind and philosophy, the software ceased evolving when its creator passed away, cementing its legacy as a unique, singular experiment in art history.

Why the First AI Image Still Matters

Looking back at AARON’s primitive black-and-white lines from 1973, it is easy to draw a straight line to the hyper-advanced AI imagery we see today.

Yet, AARON represents something deeper than technological progress. Modern AI often raises heated ethical debates about copyright, originality, and whether a machine can truly be “creative.” Harold Cohen faced those exact same questions half a century ago.

By encoding his own understanding of art into a machine, Cohen forced the world to ask a profound question: Is creativity defined by the tool used, or by the intent behind it?

Long before text-to-image generators sparked a global cultural shift, AARON quietly proved that the intersection of code and canvas was fertile ground for human imagination.