The True Story Behind the First AI-Generated Image in 1973

The Genesis of Machine Creativity: Inside AARON and the First AI-Generated Images of 1973

Long before modern generative models could conjure hyper-realistic photos from a simple text prompt, the journey of artificial intelligence art began in a quiet laboratory environment with a computer program named AARON. Created in 1973 by British painter and programmer Harold Cohen, AARON represents the historic milestone where a machine first took autonomous control of the artistic process.

Who Was Harold Cohen and Why 1973 Matters

In the late 1960s and early 1970s, Harold Cohen was a successful, internationally acclaimed painter who represented Great Britain at prestigious global art exhibitions. However, his curiosity was captured by an emerging technology: computing.

Rather than using a computer merely as a fancy calculator, Cohen wanted to understand the cognitive mechanics of drawing. He asked a radical question: What is the minimum set of cognitive rules required for someone to recognize a drawing as a representation of something?

By 1973, Cohen’s coding efforts culminated in the first breakthrough iterations of AARON. This software was not just executing random code; it was explicitly designed to make autonomous creative choices.

How the First AI Art Was Made

Unlike contemporary AI tools that rely on massive neural networks trained on billions of scraped internet images, AARON operated on manually coded symbolic logic and cognitive rules:

• The Code as a Brain: Cohen didn’t feed AARON pictures to copy. Instead, he programmed rules about geometry, spatial awareness, occlusion (how objects overlap), and composition.
• The Robotic Arm: Once AARON calculated its composition via early computing hardware, the program outputted data to a physical drawing machine—a flatbed plotter—which held an ink pen and physically drew the lines onto paper.
• Total Autonomy: Once Cohen pressed start, he did not intervene. The machine decided where lines went, how shapes interacted, and what the final composition looked like entirely on its own.

The resulting 1973 outputs were fascinating abstract line drawings and primitive closed forms. They looked starkly different from modern pixel-perfect graphics, consisting primarily of clean, calculated black-and-white lines.

A Four-Decade Evolution

What makes the 1973 milestone remarkable is that it wasn’t a one-off science experiment. Cohen spent the next 40 years continuously updating AARON:

1. The 1980s–1990s: AARON evolved from drawing abstract lines to generating complex organic shapes, human figures, and eventually learning how to paint with color using custom color-mixing routines.
2. Exhibitions: AARON’s works graduated from lab printouts to major international institutions, including the Tate Modern and the Whitney Museum of American Art.

Legacy in the Era of Modern Generative AI

Today, when anyone can type a few descriptive words into an app and instantly get a masterpiece, it is easy to forget how much friction used to exist between human imagination and machine output.