The First Ever AI Image: History of AARON and Early AI Art

The Genesis of Synthetic Dreams: The Story of AARON, the World’s First AI Artist

Long before text-to-image prompts could summon hyper-realistic portraits in seconds, a quiet revolution began in a laboratory in the early 1970s. While the modern world associates artificial intelligence art with neural networks, GPUs, and internet-scale datasets, the true pioneer of AI-generated imagery was a computer program named AARON, born in 1973.

The story of the first-ever AI image is not one of automated theft or massive data scraping. Instead, it is a fascinating tale of fine art, cognitive science, and a painter who wanted to know if a machine could truly understand the act of creation.

Meet Harold Cohen and AARON

In the late 1960s, Harold Cohen was a successful, internationally renowned British painter representing the UK at major global exhibitions like the Venice Biennale. Yet, at the height of his traditional career, Cohen found himself consumed by a radical question: What is the core essence of drawing?

To answer this, Cohen made a bold pivot. He moved to the United States to teach at the University of California, San Diego, and began teaching himself computer programming. He didn’t want to use the computer merely as a digital paintbrush to replicate his physical hand movements; he wanted to build an autonomous entity that could make its own creative decisions.

By 1973, that entity had a name: AARON (partly named after biblical history).

How the First AI Images Were Made

Unlike today’s models that rely on deep learning, transformer architectures, and billions of parameters, AARON was built on symbolic artificial intelligence and manually coded rule sets.

• The Logic of Art: Cohen didn’t feed AARON pictures to copy. Instead, he meticulously translated human cognitive understandings of the physical world into code. He programmed rules about geometry, perspective, occlusion (how one object blocks another), and human anatomy.
• Autonomous Composition: Once the program was set in motion, AARON made independent choices. It decided how to arrange shapes, construct figures, and balance compositions on a canvas without Cohen intervening.
• The Output: The earliest iterations of AARON produced abstract, elegant black-and-white line drawings. Connected to a specialized drawing machine (a mechanical plotter with a pen), the computer would physically guide the pen across paper, etching out lines entirely on its own.

When Cohen publicly demonstrated AARON at the University of California, Berkeley in 1974, the art and tech worlds caught a glimpse of a completely new paradigm: a machine generating original visual art.

A Four-Decade Evolution

AARON was not a one-off experiment; it became Cohen’s life’s work for over forty years. As computing technology evolved, so did AARON:

1. The 1980s and 90s: Cohen successfully integrated color into the software, teaching AARON how to mix its own paints and fill in its drawn shapes autonomously.
2. The 2000s: The software evolved from drawing isolated human figures to creating complex, organic environments, including rich, jungle-like plantscapes (Aaron KCAT in 2001, and Gijon in 2007).
3. Institutional Legacy: AARON’s evolutionary output transcended laboratory walls, earning exhibitions at prestigious venues like the Tate Modern and the Whitney Museum of American Art.

Why the First AI Image Still Matters

When Harold Cohen passed away in 2016, AARON’s journey ceased, but its footprint on history remained profound.

Today’s text-to-image generators conjure hyper-detailed masterpieces by statistically analyzing billions of human-made images. In contrast, AARON possessed no dataset of internet images—it operated purely on an internal logic of how to structure reality. AARON forces us to rethink the boundary between human intent and machine execution, proving that the quest to create synthetic art is not a 21st-century Silicon Valley phenomenon, but a decades-old artistic dream.