The First AI-Generated Image: History and Impact

Before the Prompt: The Surprising History of the First AI-Generated Image

When we think of artificial intelligence and art today, our minds instantly snap to a familiar ritual: typing a descriptive prompt into a modern web browser, waiting a few seconds, and marveling as a hyper-realistic image materializes out of a cloud of pixels.

We live in an era where generative AI models like Midjourney, DALL-E, and Stable Diffusion can render photorealistic cinematic scenes, intricate vector graphics, and convincing digital oil paintings on a whim. Yet, the story of the very first AI-generated image didn’t begin with a text prompt, a neural network, or a massive data scrape of the internet.

In fact, it started more than half a century ago, born from a radical question asked by a celebrated British painter who decided to trade his paintbrush for a computer terminal.

Meet Harold Cohen and AARON

To find the genesis of artificial intelligence art, we have to travel back to the late 1960s and early 1970s. At the time, Harold Cohen was a successful, internationally renowned abstract painter representing Great Britain at prestigious global showcases like the Venice Biennale.

In 1968, while serving as a visiting professor at the University of California, San Diego, Cohen developed a profound curiosity about computing. He didn’t view computers as a threat to creativity, nor as a simple appliance; he viewed them as an intellectual mirror. He wanted to understand what constitutes the core act of drawing. What are the cognitive rules that allow a human being to translate a mental concept into lines on a page?

By 1973, Cohen had created a groundbreaking software program that would permanently alter art history: AARON.

AARON was not an image generator in the modern sense. It did not possess a database of millions of scraped internet pictures, nor did it use deep learning or statistical neural networks. Instead, AARON was a symbolic, rule-based program built during the era of GOFAI (“Good Old-Fashioned AI”).

How the First AI Art Was Born

Cohen didn’t teach AARON what art looked like; he taught it how art was made.

Working painstakingly over years at UCSD and later during a residency at the Stanford Artificial Intelligence Laboratory, Cohen manually translated his own deep understanding of artistic composition into computer code:

• How lines intersect and close to form shapes.
• How to balance positive and negative space across a canvas.
• How objects occlude (overlap) one another in a three-dimensional space projected onto a two-dimensional surface.

When AARON generated an image in the mid-1970s, it operated entirely autonomously. Cohen provided the hardware—often custom-built mechanical drawing machines and flatbed plotters connected to primitive computers—but AARON made the creative decisions. The program decided where to place a line, how to weight a shape, and how the composition would grow. Cohen would frequently watch the robotic arm glide across paper, tracing out primitive geometric patterns, abstract figures, and primitive organic forms, serving more as an assistant to his own creation than its master.

From Lines to Jungles

Over the decades, AARON evolved alongside computing technology.

• The 1970s: AARON produced stark, mesmerizing black-and-white line drawings. In many cases, Cohen would meticulously hand-color these computer-plotted prints, bridging the gap between machine logic and human touch.
• The 1980s and 1990s: As computing power expanded, Cohen successfully programmed AARON to understand representational structures, introducing human figures, rocks, potted plants, and complex botanical elements into its repertoire. Later iterations even selected their own color palettes and painted directly.
• The 2000s: AARON transitioned into rendering lush, dense, jungle-like environments, earning exhibitions at world-class cultural institutions like the Tate Modern in London and the Whitney Museum of American Art in New York.

When Cohen passed away in 2016, he left behind a legacy that fundamentally challenged our definitions of authorship and creativity.

Why the First AI Image Still Matters

Looking at AARON’s primitive black-and-white plots next to today’s sleek, ultra-detailed generative masterpieces highlights a stark technical contrast. Modern generative models harness massive transformer architectures and diffusion mathematics to convert text prompts into cinematic art in seconds.

Yet, AARON’s legacy remains profoundly relevant. Modern AI learns by mimicking human art wholesale through statistical correlation. AARON, by contrast, was built on an attempt to understand the grammar of cognition—asking a machine to simulate the actual mental mechanics of a creator.

The first AI-generated image reminds us that the intersection of technology and art is not a modern Silicon Valley invention. It is a decades-long philosophical journey—one that began when a painter sat in a brightly lit university lab, stared at a glowing terminal, and dared to ask whether a machine could learn to see.