Pixels Before Prompts: The Fascinating Story of the First AI-Generated Image

Today, generating a breathtaking digital painting, a hyper-realistic photo, or a wild fantasy scene takes mere seconds. You simply type a few words into a text box, hit enter, and watch an artificial intelligence bring your imagination to life.

But long before text-to-image generators like DALL-E or Midjourney existed, the concept of a machine creating art was purely science fiction. The journey toward machine-made art didn’t begin with neural networks or vast training datasets—it started in a laboratory in the late 1960s with a British painter named Harold Cohen.

Meet AARON: The World’s First AI Artist

In 1968, Harold Cohen was a successful, internationally recognized painter representing Great Britain at prestigious global art exhibitions. However, he became deeply fascinated by a radical question: What is the core essence of drawing? What are the cognitive rules that allow humans to look at a blank canvas and make sense of lines, shapes, and proportions?

To answer this, Cohen stopped painting himself and began writing code. He moved to the United States to work at the University of California, San Diego, and developed a computer program that would eventually change art history. He named it AARON.

By 1973, AARON was officially producing its earliest autonomous art pieces.

How Did AARON Create Art Without Machine Learning?

If you compared AARON to modern AI, you would notice a massive difference. Modern AI tools learn by analyzing billions of images scraped from the internet. AARON, on the other hand, didn’t “learn” from existing pictures at all.

Instead, AARON operated entirely on manually coded rules:

• The Logic of Drawing: Cohen systematically translated his own understanding of human anatomy, perspective, overlapping objects, and composition into programming logic.
• Autonomous Decision-Making: While Cohen wrote the rules, AARON made independent choices. Once the program started running, it decided how to compose a scene, where to place figures, and how lines interacted on its own—meaning no two drawings were ever identical.
• Physical Output: Early versions of AARON didn’t even output to a digital screen. The computer drove a custom mechanical drawing machine—a robotic turtle-like arm holding a pen—to physically ink the lines onto paper.

“AARON didn’t copy human art; it was given an understanding of visual rules and left to figure out the expression for itself.”

From Black-and-White Lines to Complex Landscapes

For its first decade, AARON generated simple, striking black-and-white line drawings featuring abstract shapes, rocks, and rudimentary human figures.

As computer technology evolved, so did AARON:

• The Addition of Color: In the 1980s and 1990s, Cohen updated the software so that AARON could pick colors and paint its own canvases, following complex rules for color harmony and texture.
• Thematic Shifts: Later versions of the software moved away from isolated human figures to complex, vibrant jungle-like scenes filled with stylized plants and silhouettes.
• Global Recognition: AARON’s work became so respected that it was eventually exhibited in world-class institutions, including the Tate Modern in London and the Whitney Museum of American Art in New York.

The Legacy of the First AI Images

Harold Cohen passed away in 2016, but his life’s work laid the philosophical and technical groundwork for everything that followed.

While today’s AI relies on deep learning architectures like GANs (Generative Adversarial Networks) and diffusion models to instantly whip up photorealistic masterpieces, AARON proved a profound concept decades earlier: a machine could be creative. It showed that computers were not just calculators for math problems, but potential partners in human creativity.

The next time you prompt an AI to design a futuristic cityscape, remember Harold Cohen and his robotic pen—the pioneers who drew the very first lines of the AI art revolution.