LA-based digital media artist Refik Anadol is fascinated by libraries – maybe because both his parents were teachers. Growing up in Istanbul, after learning to code at an early age, his skills were drawn to the spacing and ordering of the repositories of learning.
This comes as no surprise: Anadol’s radical visualisations of generative neural networks are based on mining the data held within them. As a digital artist, he’s been given access to the likes of NASA’s space archives for Machine Hallucinations, or 200 million images of Earth for his Quantum Memories show; he’s digitised every note in Mozart’s repertoire and absorbed each poetic verse of Rumi’s epic Masnavi, consigning it all to the virtual realm, for imaging.
Vast growth in processing power does not make his job easier, but it does make it faster. He shows me two images created using the same data from Istanbul’s SALT archive. One, square, like white, gold and black paint dragged across a canvas, took six years to generate a decade ago. The other, current one took just 90 minutes and shows a nuanced blend of the same shades: small clouds gathered in a circle, as if in a petri dish.
The former, he says, “uses algorithms in X and Y directions, correlating to a human need for narrative, while the latter, with its many axes, shows the non-linearity of how an AI mind now looks at the data”; there’s a reason it looks like bacteria – or a brain scan.
“The craftsmanship is in the algorithms that interrogate the data, like a chisel to marble,” he adds. “I will spend a year sorting, creating meaning in data, and only then does the real work start.”
What he’ll do at Oxford University, with its Bodleian Library, is anyone’s guess, but its multi-dimensionality is hinted at above. “I love Borges’ imagining in The Library of Babel, of all the world’s information held in one place – and what it might look like.” Could Anadol’s work be a path to a new way of seeing? We are no longer in Borges’ hallucination.
