A Complete History of Early AI Art and Pictures

The Weird, Wonderful World of Early AI Pictures: From Plotters to Pixels

When most people think of AI images today, they picture hyper-realistic photographs, surreal fantasy landscapes, or cinematic art generated with a simple text prompt. But long before modern apps could render a photorealistic astronaut riding a dinosaur in seconds, artificial intelligence was taking its very first, mechanical baby steps.

The story of early AI pictures is a fascinating journey of code, computers the size of rooms, and robotic arms that laid the groundwork for today’s digital canvas.

1. The Robotic Artist: AARON (1970s)

Long before deep learning existed, artists and computer scientists were already experimenting with autonomous systems. One of the absolute pioneers was Harold Cohen, a painter and professor who, in the late 1960s and 1970s, began developing a computer program called AARON at the University of California, San Diego.

• How it worked: AARON didn’t use neural networks or internet data. Instead, it ran on a set of symbolic, rule-based programming written by Cohen. It understood basic rules about geometry, composition, and human-drawn shapes.
• The Output: AARON was connected to a custom-built mechanical drawing machine (a plotter) that held a physical pen, drawing lines on large sheets of paper. Cohen would then often color in these computer-generated lines by hand. Over decades, AARON evolved from making abstract geometric grids to rendering complex plants, figures, and room interiors.

2. The Algorithmic Pioneers of the 1960s and 70s

While AARON focused on drawing objects, other computer pioneers were exploring algorithmic art—using mathematical formulas and computers to generate complex visual patterns.