What Is the First AI Generated Image? History and Origin

The Genesis of Synthetic Sight: What Was the First AI-Generated Image?

When we think of artificial intelligence and images today, our minds instantly snap to hyper-realistic, text-prompted wonders generated in seconds. But the story of AI art did not begin with modern diffusion models or deep neural networks. It began decades earlier, born from an artist’s obsession with code, cognition, and the fundamental mechanics of human drawing.

To answer the question of what the first AI-generated image actually was, we have to travel back to the late 1960s and early 1970s to meet a program named AARON.

Meet AARON: The World’s First AI Artist

The creation of AARON is credited to Harold Cohen, a renowned British-born painter and professor who represented Great Britain at the prestigious Venice Biennale in 1966. At the peak of his traditional fine-art career, Cohen experienced a radical crisis of curiosity: What is it that makes a drawing a drawing?

Rather than continuing to paint with a traditional brush, Cohen moved to the University of California, San Diego, in 1968, taught himself computer programming, and began writing the code for what would become history’s first autonomous art-generating software.

Unlike modern AI systems that ingest billions of images from the internet to recognize patterns, AARON operated entirely on symbolic, rule-based logic (popularly known in early computer science as “GOFAI” or Good Old-Fashioned AI).

Cohen did not train AARON with data sets; instead, he manually coded human rules of art into the program:

• The rules of perspective and occlusion (how objects overlap).
• How to balance compositions and arrange human figures or botanical shapes.
• The cognitive mechanics of physical drawing and spatial reasoning.

What Did the First AI Images Look Like?

The earliest iterations of AARON outputted black-and-white line drawings.

In the early 1970s—with the foundational output solidified around 1973—AARON wasn’t spitting out digital files onto a screen. Instead, Cohen connected the program to a custom-built, turtle-like drawing machine that moved across a flatbed plotting table, clutching a pen to physically ink lines onto paper.

The resulting images featured abstract human figures, primitive geometric shapes, and stark foliage. While they look remarkably simple by today’s standards, they were revolutionary: AARON was making independent composition decisions. The computer was choosing where to place lines and how to construct figures without Cohen actively steering the pen or intervening in real time.

An Evolution Across Decades

AARON did not stay frozen in the 1970s. As computing power advanced, Cohen continuously upgraded his digital creation:

• The 1980s and 1990s: AARON learned to handle color, with Cohen later programming algorithms that allowed the software to paint its own creations or instruct robotic painting arms.
• The 2000s: Later versions like Aaron KCAT generated massive, vibrant projections of human-plant hybrids and jungle scenes, which were eventually exhibited in renowned institutions like the Tate Modern and the Whitney Museum of American Art.

Cohen famously wrestled with the philosophical implications of his own creation, asking: Who is the true artist—the human programmer, or the machine? He viewed AARON not as a tool, but as an independent collaborator.

From AARON to Modern AI Art

Looking at the primitive, calculated black-and-white lines of AARON provides a humbling contrast to the hyper-detailed landscapes and photorealistic portraits generated by contemporary text-to-image systems.

Yet, the core spirit remains identical. Long before neural networks, Generative Adversarial Networks (GANs), and diffusion pipelines dominated technology headlines, Harold Cohen and AARON proved that a machine could conceptualize visual form—officially cementing AARON’s hand-drawn lines as the very first AI-generated images in human history.