The True Story Behind the First Ever AI Generated Image

The Genesis of Digital Dreams: The Story of the First AI Image

Long before the era of typing a text prompt into a modern tool to render hyper-realistic photos or whimsical digital paintings, the concept of machine-made art was born in a quiet laboratory.

Today, artificial intelligence image generation is a cultural phenomenon. Yet, its roots stretch back decades, long before neural networks, massive internet datasets, or advanced GPUs existed. The journey of the world’s very first AI-generated image is a tale of visionary art, painstaking computer programming, and a machine named AARON.

Meet Harold Cohen and AARON

In the late 1960s, British-born artist Harold Cohen was at the peak of his traditional art career. He had represented Great Britain at the prestigious Venice Biennale and his abstract paintings hung in major galleries.

However, Cohen grew deeply curious about a fundamental question: What actually makes art? More specifically, could a computer be taught to make independent creative decisions rather than just blindly following human instructions?

Leaving behind his traditional brushes for a moment, Cohen moved to the United States to teach at the University of California, San Diego, and began teaching himself computer programming. By the early 1970s—crystallizing around 1973—he had developed a pioneering software program named AARON (partly inspired by biblical history).

How the First AI “Drew”

Unlike modern AI models that learn by scraping billions of images from the internet, AARON had zero data to look at. It didn’t use deep learning or neural networks. Instead, AARON was a rule-based system.

• Hand-Coded Logic: Cohen manually wrote out the cognitive rules of drawing. He taught the program how to understand basic geometry, how lines relate to shapes, how to balance proportions, and how objects overlap.
• Autonomous Choices: Once the code was running, AARON made its own choices. It decided composition, where to place lines, and how to construct figures without Cohen guiding its hand in real-time.
• The Physical Output: Early versions of AARON didn’t output to a computer screen; instead, the software drove a specialized mechanical drawing machine (a plotter) equipped with a pen to physically ink black-and-white lines onto paper.

When AARON first began producing its rudimentary abstract shapes and primitive figures, it achieved something profound: The machine was creating original variations that even Cohen himself could not entirely predict.

Evolution and Legacy

Over the next four decades, Cohen continuously updated AARON. What started as simple, abstract black-and-white line drawings gradually evolved. By the 1990s and 2000s, AARON was given rules for color, eventually generating complex, vibrant scenes featuring plants, rocks, and human-like figures that were exhibited in world-class institutions like the Tate Modern and the Whitney Museum of American Art.

When Cohen passed away in 2016, he left behind an extraordinary legacy. AARON proved that the boundary between human creativity and machine execution was porous long before the 21st century.

From Lines to Diffusion Models

Looking back, AARON’s stark, primitive line drawings are a far cry from the instantaneous, hyper-detailed creations generated by modern text-to-image models. Yet, the DNA is the same.