What Was the First AI-Generated Image in History?

Before the Prompts: The Surprising Origin Story of the First AI-Generated Image

When we think of AI-generated art today, our minds instantly flash to hyper-realistic masterpieces summoned in seconds via a sleek text prompt on platforms like Midjourney, DALL-E, or Stable Diffusion. We type a few descriptive words, press enter, and marvel as a neural network hallucinates a photorealistic world.

Yet, the lineage of artificial intelligence creating imagery stretches back far longer than most realize. Long before GPUs, neural networks, or massive internet datasets entered the picture, a lone artist and programmer quietly birthed the world’s first true AI image generator.

His name was Harold Cohen, and his creation was called AARON.

Meet AARON: The World’s First AI Artist

In the late 1960s, Harold Cohen was a successful British painter representing his country at prestigious international art exhibitions like the Venice Biennale. However, he grew fascinated by a radical question: What makes a drawing a drawing? Could the cognitive rules behind human artistic creation be codified and executed by a machine?

Leaving behind a traditional art career, Cohen moved to the United States, taught himself computer programming, and began developing AARON around 1968. By 1973, AARON successfully produced its first autonomous, computer-generated artistic images.

Unlike modern AI, which learns by scanning billions of images scraped from the web, AARON didn’t use deep learning, neural weights, or training data. Instead, AARON was built entirely on hand-crafted symbolic logic and rules.

Cohen meticulously wrote code that embodied human understanding of:

• Basic geometry and structural proportions
• How objects overlap and obscure one another in space
• The rules of composition and perspective

Operating a custom drawing machine connected to a mainframe computer, AARON would autonomously plot out black-and-white line drawings of abstract shapes, rocks, plants, and eventually human figures. Cohen didn’t design a specific end-image and tell the computer to draw it; rather, he gave the software the rules of art, allowing AARON to make independent compositional choices on the fly. Seeing a machine physically holding a pen and drafting a composition that no human had pre-designed was nothing short of science fiction for the era.

Evolution Beyond the Lines

As computing power evolved over the decades, so did AARON.

• The 1980s and 90s: Cohen updated the software to understand color theory, allowing AARON to paint its own generated lines using specialized brushes and liquid paint mixtures.
• The 2000s: AARON shifted from drawing individual figures to rendering rich, jungle-like environments.

AARON’s work was so revolutionary that it transcended computer science labs, eventually being exhibited at world-class institutions like the Tate Modern and the Whitney Museum of American Art. Cohen continued developing AARON until his passing in 2016, leaving behind a legacy as the undisputed pioneer of autonomous visual generation.

From Rule-Books to Neural Networks

To understand how far we’ve traveled, it helps to look at the generational leaps in AI imagery:

| Era | Primary Technology | How It Worked | What It Produced |
| :— | :— | :— | :— |
| 1970s–2000s | Rule-Based Systems (AARON) | Hard-coded human logic and geometry rules. | Black-and-white line drawings, later filled with color. |
| 2014 | GANs (Generative Adversarial Networks) | Two neural networks (a generator and a discriminator) competing against each other. | Strikingly realistic human faces and objects (