Eye-to-eye with Jason Salavon’s algorithm-produced art

This article was taken from the November 2014 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.

These abstract artworks weren't created by an artist -- they were painted by an algorithm. "I tasked myself with building a virtual automaton that would iterate relentlessly though a hyper-accelerated painting career," explains Chicago-based artist Jason Salavon, 44. For Golem, Salavon wrote a script to generate 100,000 such artworks, with the ultimate goal that the work might pass "a kind of Turing Test for abstract painting". The works were generated sequentially -- up to 700 per day, over six months -- with Salavon altering the algorithm throughout the computer artist's "lifetime". Salavon created the project in 2002 but revisited it this year to combine each piece into a web-based map using OpenGIS.

The result: an almost haunting representation of a computer artist's career, complete with clearly defined periods of shifting work -- albeit changes made not through experience but code.

The sheer volume of paintings is typical of Salavon, who has crafted a career through amalgamating huge volumes of data into high-impact pieces. His "Figure 1" combines ten years of Playboy centrefolds into a haunting spectre of the magazine pin-up; "The Top Grossing Film of All Time" combines every frame of Titanic into a huge canvas of pixelated colours. "I'd say much of my work occupies a space between cultural sampling and data mining," says Salavon.

Next up, Salavon will spend the next six months as the first artist-in-residence at Microsoft Research in Seattle. Among the ideas he'll be exploring: "What would [Rennaisance painter] Pieter Bruegel the Elder or Hieronymus Bosch do with contemporary computation?" To Salavon, the interaction between art and big data is still in its infancy. "If you imagine what kind of things artists will make in 20 years," he says, "it's hard not to see how algorithmically informed work won't be central." Painters: better get coding, fast.

This article was originally published by WIRED UK