The First AI-Generated Image: The Story of AARON in 1973

The Genesis of Synthetic Canvas: How 1973 Gave Birth to the World’s First AI-Generated Image

When we think of artificial intelligence creating art today, our minds instantly drift to the hyper-realistic, photorealistic wonders summoned by text prompts in a matter of seconds. We live in an era where diffusion models and neural networks can conjure up anything from Renaissance oil paintings to futuristic cityscapes with a keystroke.

Yet, the story of machine-made art didn’t begin with modern deep learning. It traces back more than half a century to a quiet laboratory in California, long before GPUs, the internet, or consumer computers existed.

The year was 1973, and the birthplace of machine-born imagery was not a high-tech corporate office, but the mind of a visionary British painter named Harold Cohen.

Meet AARON: The Pioneer Program

In 1973, Harold Cohen was an accomplished, internationally recognized painter representing Great Britain at prestigious global showcases like the Venice Biennale. Seeking to challenge his own understanding of what constituted the creative act, Cohen took up a fellowship at the Stanford Artificial Intelligence Laboratory (SAIL).

He wasn’t interested in using a computer merely as an advanced paintbrush or a photo-editing tool. Instead, he wanted to explore a profound question: Can the cognitive process of drawing be reduced to a set of rules and executed by a machine?

To answer this, Cohen began developing a symbolic AI program that would eventually change art history: AARON.

Unlike modern AI, which relies on vast datasets, billions of parameters, and neural networks trained on scraped internet images, AARON was built entirely from scratch using hand-coded logic.

• The Logic: Cohen didn’t feed AARON pictures to mimic. Instead, he systematically translated his own understanding of human cognition, spatial reasoning, and visual perception into code.
• The Rules: He taught the program how to draw closed shapes, understand the boundary between foreground and background, manage overlapping lines, and balance a composition.
• The Autonomy: Once the rules were set, AARON made its own independent compositional choices. Cohen had no idea what the final layout would look like before the machine went to work; the computer synthesized the arrangement entirely on its own.

Bringing the Image to Life

Because digital screens capable of displaying complex graphics were virtually nonexistent, AARON communicated with the physical world through a mechanical flatbed plotter—a bulky robotic arm holding a pen that hovered over a large sheet of paper.

In 1973, as the machine whirred, clicked, and dragged its pen across the page, it brought forth abstract, black-and-white line drawings. These weren’t traces of pre-existing human drawings; they were pixels (or rather, pen strokes) generated algorithmically by a machine acting on conceptual rules.

Art historians and computer scientists universally point to these stark, early 1970s printouts as the first true AI-generated images in history.

A Lifelong Collaboration

What makes the story of AARON and that initial 1973 milestone so remarkable is what happened next. Most tech experiments are swiftly abandoned for newer models, but Cohen spent the rest of his life—over four decades—co-evolving with his creation.

As computing power advanced, Cohen continuously upgraded AARON:

• The 1980s and 90s: AARON evolved from abstract line drawings into complex organic shapes, eventually learning to render human figures and lush botanical foliage.
• The Introduction of Color: Cohen later taught AARON how to handle color mixing and painting. While the computer chose the palette and filled in the shapes, Cohen would sometimes act as AARON’s physical assistant, manually painting the colors onto the canvas under the program’s precise instructions.

AARON’s works eventually graduated from experimental labs to world-class institutions, being exhibited at the Tate Modern in London and the Whitney Museum of American Art in New York.

The Legacy of 1973

When we marvel at the lightning-fast capabilities of today’s generative AI tools, it is easy to forget the decades of foundational groundwork laid by pioneers like Harold Cohen.

The first AI-generated images of 1973 were quiet, minimalist, and mechanical. Yet, they planted a radical seed: the idea that a machine could possess an autonomous aesthetic “mind.” Cohen proved that art was not strictly locked behind the barrier of human biology, bridging the worlds of code and creativity long before the digital revolution ever dreamed it possible.