In the late 1960s, Harold Cohen’s career took a radically different trajectory when the British painter moved to the U.S. to take a position at the University of California, San Diego. It is there where Cohen began toying with computer-generated artwork, creating the software he would later dedicate his life to.
Before the rise of artificial intelligence (A.I.) tools like DALL-E, Midjourney and Stable Diffusion, there was “Aaron.” Developed by Cohen over decades of exploration, the program was the world’s first A.I. artmaking software. Now, works generated by Aaron will star in an upcoming exhibition at the Whitney Museum of American Art, one that details the beginnings of A.I. and the history of its relation to the art world.
Opening in February, “Harold Cohen: AARON” will feature artwork created by the software and held by the Whitney, which is the only museum holding different versions of the program from varying periods. The show will also display operating recreations of Cohen’s early drawing machines. “Watching Aaron’s creations drawn live as they were half a century ago will be a unique experience for viewers,” said Christiane Paul, curator of digital art at the Whitney, in a statement.
How did Harold Cohen become a digital art pioneer?
Before changing mediums, Cohen was a successful painter back in Britain. He represented the U.K. at the 1966 Venice Biennale and exhibited at shows like Documenta III, the Paris Biennale and the Carnegie International.
He first conceived of Aaron while teaching at UC San Diego in 1968, where he would work for three decades as a professor and eventually the director of its Center for Research in Computing and the Arts. The program was developed further when he went to Stanford University’s Artificial Intelligence Lab in the 1970s. Cohen named the software Aaron in 1973, drawing inspiration from the biblical figure and the assumption that it would eventually be followed by a “B” version.
Cohen trained his software with a basic understanding of artistic strategies, including the practice of drawing in the foreground before moving to the background. While Aaron’s initial output consisted of abstractions, Cohen developed the program to focus on figures and plant life—by the early 2000s, Aaron was able to generate complex foliage that followed rules regarding plant size, branching levels and even leaf formation patterns.
In an attempt to produce accessible works, Cohen partnered up with computer scientist Raymond Kurzweil to create a screensaver version of Aaron with Kurzweil’s CyberArt Technologies (KCAT), a software released in 2001 that remains one of the program’s most well-known versions. Several outputs from the KCAT software, including images depicting jagged and colorful figures posing alongside plants, will be included in the Whitney’s upcoming show.
Working with Aaron for much of his artistic career, Cohen considered the software a collaborator. The program wasn’t so much a tool as a “sort of assistant,” said the artist in a 1982 paper to the National Bureau of Standards. “We are living on the crest of a cultural shock-wave of unprecedented proportions,” added Cohen while describing his awe for the potential of A.I. “We are in the process of coming to terms with the fact that ‘intelligence’ no longer means, uniquely, ‘human intelligence.’”
Aaron operated in a fundamentally different manner than today’s A.I. artmaking tools—instead of relying on algorithms to generate output from datasets, Aaron worked from rules and knowledge stored in its memory. The evolution of the technology will be a key element of the Whitney’s exhibition, which is scheduled to run through May 2024.
“By bringing AARON back to life in our galleries, we hope to show how our exhibitions can be as innovative as the artists we present,” said Scott Rothkopf, the museum’s director, in a statement. “This exhibition draws from the museum’s collection to present an important timely exploration of the early A.I. tools for artmaking, decades before they entered mainstream conversations.”
