Beyond the Prompt: The Fascinating Story of AARON and the First AI-Generated Images

Today, typing a few descriptive words into an application yields breathtaking, hyper-realistic images in seconds. This seamless fusion of text and pixels feels entirely modern, a hallmark of deep-learning diffusion models. Yet, the quest to teach machines how to create art is far older than most realize.

Long before neural networks, GPUs, or internet-scale datasets, the world witnessed the birth of the first true artificial intelligence image generator. Its name was AARON, and its creation began not in a high-tech computer science lab, but in the mind of a painter.

The Artist Who Coded a Mind

In the late 1960s, Harold Cohen was a successful, internationally acclaimed British painter representing Great Britain at prestigious global art exhibitions. However, a profound curiosity gripped him: What is the core essence of drawing? What rules govern how humans translate a three-dimensional world onto a flat surface?

To answer this, Cohen made a radical pivot. He moved to the United States—eventually working at the University of California, San Diego, and the Stanford Artificial Intelligence Laboratory—to teach himself computer programming.

By the early 1970s (with breakthrough outputs emerging around 1973), Cohen’s digital creation, AARON, produced its first landmark automated drawings. AARON was not just a tool that executed pre-designed templates; it was an autonomous system programmed to make its own creative decisions.

How the First AI Image Generator Worked

Unlike modern generative AI models that rely on statistical pattern recognition from billions of scraped photographs, AARON operated on symbolic, rule-based reasoning (often associated with “Good Old-Fashioned AI” or GOFAI):

• Manual Logic Encoding: Cohen did not feed AARON a dataset of existing images. Instead, he painstakingly translated his own understanding of cognitive art rules into code.
• Autonomous Composition: The program was taught how to construct shapes, manage perspective, handle object occlusion (knowing when one object overlaps another), and balance spatial proportions.
• The Physical Act: Early iterations of AARON didn’t live purely on a screen. Connected to a custom-built mechanical drawing machine (a turtle-like plotter device), the computer would physically hold a pen and ink lines onto paper on the floor.

The resulting black-and-white line drawings—featuring abstract figures, primitive plants, and organic shapes—were revolutionary. They represented the first time an autonomous machine compositionally mapped out and executed an original piece of visual art without human hand-guidance during the drawing process.

Evolution and Cultural Impact

Over the following decades, AARON evolved dramatically. As computer hardware advanced, Cohen updated the software to handle color selection, painting mechanics, and increasingly complex scenes involving human figures and lush, jungle-like environments.