Pixels and Punch Cards: The Fascinating History of Early AI Pictures

Today, creating a stunning piece of digital art is as easy as typing a few descriptive words into a prompt bar. In seconds, algorithms spit out hyper-realistic portraits, dreamy landscapes, and complex abstract designs.

However, the journey of making art with computers didn’t begin with powerful modern graphics cards or deep learning neural networks. It started decades ago, born out of the curiosity of mathematicians, engineers, and scientists who wondered what machines might be capable of imagining.

Let’s take a journey back in time to explore how early AI and computer-generated pictures came to be.

1. The Pioneers: Art from Mainframes (The 1960s)

Long before personal computers sat on desks—let alone smartphones in pockets—massive, room-sized mainframe computers were crunching data in university labs and corporate research centers.

In the early 1960s, visionaries like A. Michael Noll at Bell Labs and Georg Nees in Germany began writing code not to solve equations, but to draw shapes. Using devices called plotters—machines equipped with a physical pen attached to a mechanical arm—computers traced lines onto giant sheets of paper based on mathematical instructions.

• Algorithmic Art: These early pieces weren’t “AI” in the modern neural-network sense, but they were pioneering examples of generative art—art created by a set of rules or algorithms written by a human.
• The Gaussian Quadratic: One of Noll’s most famous early computer-generated patterns, Gaussian Quadratic (1962), used mathematical formulas of probability to scatter lines across paper, creating a surprisingly aesthetic composition that challenged what people thought a machine could do.

2. Entering the Matrix: The Rise of ASCII Art

As computer screens evolved from simple oscilloscope displays to text-based terminals, a new subculture of digital imagery emerged: ASCII art.

Instead of drawing lines with a mechanical pen, programmers and computer enthusiasts used the standard keyboard characters (letters, numbers, and punctuation marks like @, #, $, and .) to shade and shape images.

• How it worked: Darker characters like M or W took up more visual weight on the screen, while lighter characters like . or spaces created highlights.
• The AI Connection: While much of early ASCII art was meticulously typed out by hand, programmers eventually wrote rudimentary scripts and algorithms that could scan a photograph, analyze the brightness of each pixel, and automatically translate it into a grid of keyboard characters. It was an early form of image translation and data processing!

3. The AARON Revolution (The 1970s–1980s)

One of the most important milestones in the history of machine-created art was AARON, a robotic painting program created by artist and professor Harold Cohen in 1973 at the University of California, San Diego.