John Supko on Artistic Expression in the Age of AI

One of the philosophies guiding John Supko’s research endeavors is that technology is not a replacement for human creativity but a partner in conversation. 

In his recent tgiFHI talk, “Maybe Modernism Can Save Us,” the Associate Professor of Music at Duke explored how artistic values drawn from the 20th century might guide our use of artificial intelligence today in more complex and serendipitous ways.

For Supko, modernism can effectively be seen as a model for how to think, not just how to make art. “Those works of art didn’t tell you what they meant,” he said. “They invited you to figure it out, to put together the meaning for yourself.” 

He argues that the openness and difficulty of modernist art, including the way it resists instant understanding, stand in contrast to how we often use powerful computational tools. 

“My critique of artificial intelligence is that we’re using these vastly powerful tools to do relatively mediocre things,” he explained. “They’re not really thinking. They’re rearranging prior thought.”

Rather than treating AI as an all-knowing machine, Supko treats it as something closer to a collaborator. He describes his compositions that use interactive computer systems as “provocatively incomplete.” 

In a recent piece for pianist Conrad Tao, One Hundred Thousand Billion Pieces for Piano & Computer, the computer generates partial musical environments that the performer must respond to in real time. The computer also listens to what Tao is doing and responds with its own interventions.

“I’m trying to create a spark of inspiration for the human performer,” Supko said. “The computer proposes something alluring but intentionally incomplete, and the performer completes it with their own creativity.”

Image

This collaborative process depends on what he calls “iterative intervention.” A generative system only becomes meaningful through continual adjustment and human feedback. “You can’t simply ask it for an output and be satisfied,” he said. “In order for it to be molded to your creativity and your ideas, intervention and iteration are really important.”

Supko’s interest in ambiguity and process connects to a broader artistic concern: how our culture engages with art and meaning. Drawing on the literary critic Viktor Shklovsky’s concept of 'defamiliarization,' or making the familiar strange, Supko argues that ambiguity is essential to art’s vitality. 

“We risk losing the willingness to sit with complexity,” he shared. “Ambiguity isn’t effective in politics, but it’s essential in art.” And for him, the pressure to make all meanings explicit can flatten artistic experience. 

“Political speech requires clarity and directness—those are things that make for very bad art.”

That tension between clarity and complexity is something he also observes in online culture. For Supko, the speed of social media discourages the kind of slow, layered engagement that art requires. “That kind of enrichment takes time and slowness. It’s like soil that needs to be turned over through many seasons,” he said.

Even in his own creative practice, Supko’s approach emphasizes slowness and observation. The databases his computer systems draw from are made up almost entirely of his own recordings—fragments of sound collected over years of work. 

“When I’m designing a system, I’m observing what works and what doesn’t and trying to change what doesn’t work,” he said. “It’s a slow process of development, but eventually the system stops bothering me. Then I know it’s ready to go out into the world as a work of art.”

Supko acknowledges the practical concerns around generative technologies, including plagiarism and authorship, but sees their value as catalysts for thought rather than replacements for it. “They can help identify patterns or bring to the surface problems you hadn’t considered,” he said. “But they’re not a replacement for thinking or for work.”

In the end, Supko returns to the human side of technology: the need for reflection, patience, and imagination. 

“These systems can unlock creative ideas, but only if we remember that they depend on us—on our curiosity, our willingness to intervene, and our capacity to live with uncertainty.”