What Is the First AI-Generated Image? History & Origin

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

When we think of AI-generated images today, our minds instantly snap to hyper-realistic, split-second wizardry: typing a complex text prompt into a modern diffusion model and watching a pristine digital painting materialize in seconds.

Yet, long before neural networks could dream up photorealistic portraits or surrealist landscapes, artificial intelligence was already putting pen to paper.

The story of the very first AI-generated image takes us back not to a sleek Silicon Valley tech lab of the 2020s, but to an academic computer science lab in the late 1960s and early 1970s. Meet AARON, the world’s first AI art system, and the birth of machine-made imagery.

Meet Harold Cohen and AARON

In the late 1960s, Harold Cohen, a successful British-born painter representing Britain at the prestigious Venice Biennale, made a radical pivot. Intrigued by the nascent world of computing, Cohen moved to the United States to study computer science. He didn’t want to use the computer merely as a digital paintbrush (like Photoshop decades later); instead, he wanted to see if a machine could independently understand the cognitive acts of drawing and painting.

Beginning in 1968 at the University of California, San Diego, Cohen began writing a computer program that would eventually be named AARON.

Unlike modern generative AI, which learns by devouring billions of images from the internet using machine learning and statistical probabilities, AARON operated on symbolic, rule-based logic (popularly known in early AI as “Good Old-Fashioned AI” or GOFAI).

Cohen essentially sought to deconstruct his own brain. He manually codified the rules of art:

• How do lines intersect?
• What rules govern bodily proportions and perspective?
• How do objects overlap in a two-dimensional space?
• How are figures shaded or textured?

The Breakthrough: The Early 1970s

By the early 1970s—with significant refinement during a residency at the Stanford Artificial Intelligence Laboratory between 1973 and 1975—AARON began producing its first autonomous line drawings.

What did the first images look like?
Far from the vibrant, complex renders of today, early AARON produced stark, abstract, black-and-white line drawings. Operating custom robotic drawing machines connected to mainframe computers, AARON would physically drive pens across large sheets of paper on the floor, plotting out original compositions of geometric shapes, abstract figures, and primitive botanical forms.

Crucially, AARON made its own creative choices. While Cohen provided the foundational rules and grammar of drawing, he did not pre-draw or pre-determine the final layout. Every time the program ran, it made autonomous decisions about composition, layout, and placement. In essence, the machine was improvising within the boundaries of its human-coded upbringing.

From Mainframes to Museums

As decades passed, AARON evolved dramatically.

• The 1980s & 90s: Cohen taught AARON how to handle color, introducing rich palettes and complex fills. AARON moved from drawing simple lines to painting intricate flora, fauna, and human figures.
• The 2000s and Beyond: Later iterations of AARON projected massive, lush, jungle-like environments onto gallery walls, continuously generating unique variations in real-time.

AARON’s work transcended the computer lab, earning exhibitions at world-renowned institutions like the Tate Modern, the Los Angeles County Museum of Art (LACMA), and the Whitney Museum of American Art. Cohen continued collaborating with his digital creation until his passing in 2016, leaving behind a legacy as a true pioneer who bridged fine art and computer science decades before the rest of the world caught up.

Why the First AI Image Matters Today

Looking at AARON’s primitive 1970s line drawings next to modern AI art might feel like comparing a bicycle to a supersonic jet. Yet, AARON established a profound philosophical precedent that shapes our modern debate: can a machine be creative?

Cohen didn’t view AARON as a tool to replace human effort, but rather as an independent entity—an electronic collaborator. When you marvel at the rapid evolution of AI-generated imagery today, remember that the journey didn’t start with a neural network scraping the web. It started with a painter who dared to teach a computer how to see.