What Is the First AI-Generated Image in History?

Long before anyone typed “cyberpunk cat drinking coffee” into a prompt box, artificial intelligence was already learning how to draw.

When people ask about the first AI-generated image, they aren’t looking at a product of modern deep learning or neural networks like Midjourney or DALL-E. Instead, they are stepping back to the late 1960s and early 1970s to meet a pioneering computer program named AARON.

Meet Harold Cohen and AARON

In 1968, British-born artist Harold Cohen was a successful, internationally exhibited painter representing Great Britain at prestigious global events like the Venice Biennale. Yet, he became fascinated by a radical question: What makes an image evocative, and can a computer be taught the cognitive rules of drawing?

Leaving behind his traditional studio practice, Cohen moved to the University of California, San Diego, taught himself computer programming, and began building AARON.

Rather than randomly spitting out geometric patterns (which was common in early computer art), Cohen programmed AARON to understand fundamental visual rules: how shapes overlap, how lines imply depth, and how figures relate to ground.

How the First “AI” Made Art

AARON didn’t use neural networks or web-scraped data—technology like that didn’t exist yet. Instead, it relied on a symbolic artificial intelligence framework.

• The Code: Cohen hard-coded cognitive rules of human drawing logic into the program.
• The Execution: AARON generated its images through custom-built robotic drawing machines—often called “turtles”—attached to physical flatbed plotters. The computer would calculate the coordinates, and the robotic arm would physically drag a pen across paper.
• The Collaboration: In the early decades, AARON produced black-and-white vector line drawings. Cohen would then step in to manually paint or color them in, creating a unique human-machine collaboration.

Over the next four decades, Cohen continually updated AARON. By the 1980s and 2000s, the program evolved to independently compose complex scenes, plants, and human figures, eventually mixing its own paints and driving large-scale inkjet print systems.

Why AARON Matters Today

AARON is universally recognized as the grandfather of AI art. It proved that computers could do more than just crunch numbers—they could participate in creative acts of decision-making.

While today’s generative models are fueled by billions of parameters and vast internet datasets, they share a direct conceptual lineage with Cohen’s work: exploring the boundary between human intent and machine execution.