The First AI Generated Image from 1973 Explained

Before the Prompts: How Harold Cohen and AARON Created the First AI-Generated Art in 1973

Long before text-to-image generators could conjure hyper-realistic photographs from a casual sentence, the world of artificial intelligence art had a very different genesis. In 1973, inside a university computer lab, a pioneering software program began autonomously drafting original compositions.

Its name was AARON, and its birth marked a profound milestone in human creativity: the moment a machine stopped merely computing and started to draw.

The Painter Who Turned to Code

To understand the significance of 1973, you have to look at the man behind the machine. Harold Cohen was not a computer scientist who dabbled in hobbies; he was an elite, internationally acclaimed British painter who had represented his country at the prestigious Venice Biennale.

By the late 1960s, Cohen had grown deeply curious about the cognitive mechanics of art. What actually constitutes a drawing? Why do humans recognize a collection of lines as a shape, a boundary, or a figure? Crucially, he wondered if those cognitive rules could be stripped away from human intuition and explicitly written down as code.

In 1973, while working as a visiting scholar at Stanford’s Artificial Intelligence Laboratory (and later moving to the University of California, San Diego), Cohen set out to answer that question. He began writing a program that would eventually become the world’s first sustained AI art generator: AARON.

How AARON Brought Images to Life

In an era when computers filled entire rooms and operated on punch cards and primitive terminals, AARON was a marvel of symbolic AI.

Unlike modern neural networks that are trained by scraping billions of images off the internet, AARON possessed no database of historical paintings to mimic. Instead, Cohen gave the program a set of rigid, rule-based instructions regarding spatial organization, boundary-making, and structural closure.

• The Decision Process: AARON did not create images through random generation. It made deliberate choices based on its code, deciding where a line should begin, how long it should travel, when a region was fully enclosed, and how shapes interacted with one another.
• The Robotic Hand: Software alone couldn’t paint a picture. To bridge the digital and physical worlds, Cohen hooked AARON up to a custom-built, floor-standing mechanical drawing machine (often referred to as a “turtle”). Guided entirely by the program’s real-time calculations, this robotic arm clutched a pen and dragged it across large sheets of paper spread out on the floor.

When Cohen first publicly presented AARON at the International Joint Conference on Artificial Intelligence in 1973, the results stunned observers. The machine produced intricate, abstract compositions entirely on its own. Once Cohen hit “run,” he did not intervene or guide the physical pen’s path; the machine possessed total autonomy over the final mark.

A 40-Year Evolution

What makes the story of AARON truly remarkable compared to other short-lived technological demos of the 1970s is its longevity. Cohen didn’t just write the program once—he spent the next four decades of his life continuously updating and expanding it.

As computer processing power grew, so did AARON’s capabilities. Over the years, the program evolved from creating purely abstract geometric shapes into rendering figurative elements, eventually learning to draft complex arrangements of human figures, plants, and rocks. Cohen even taught AARON how to use color, transforming the mechanical plotter drawings into rich, painted masterpieces where the computer handled the composition and Cohen filled in the pigments according to its layout.

The Legacy of 1973

Today, we live in an era where AI-generated imagery is ubiquitous, sparking fierce global debates about copyright, artistic integrity, and the definition of creativity.

Yet, long before terms like “deep learning,” “diffusion models”, or “prompts” entered the cultural lexicon, Harold Cohen and AARON laid the groundwork. They proved that a machine could exercise a form of aesthetic judgment and generate original visual culture without a human holding the brush.

The first AI-generated image of 1973 wasn’t just a technological novelty—it was the quiet opening salvo of a creative revolution that we are still living through today.