Most AI-based artists work exclusively in front of their computers. Sougwen Chung is different: they train robots to physically paint in tandem with them on massive canvases. Before they used AI, Chung, who identifies as nonbinary, painted expansive abstract artworks filled with bold, flowing lines. They then trained a neural net on decades of those paintings—and built robots trained on those neural nets to paint with them in real time. When they paint a line, the robots mimic Chung’s line and then extend it outward with new ideas and patterns. “What I’m chasing is that surprise and wonder in that machine translation,” Chung says.

Chung, 38, travels the world, painting with their robots for live audiences. Chung compares their relationship with their robots to that of a musician with their violin. “In some ways, the robotic system is a kinetic instrument that I’m navigating with,” they say.

Illustration by TIME; reference image courtesy of Sougwen Chung

Chung is now on the fifth version of the D.O.U.G. (Drawing Operations Unit Generation_X) robot. The most advanced robot absorbs not only their past artwork but also their present state of mind: it can be linked to their EEG data and alpha brain waves, and will paint more actively when they hit a flow state of meditation.

Chung splits their time between London and New York these days, and is continually pushing their experiments into new territory. They have been working with a 3D motion-capture system to create sculptures, and researching ways to power their systems with alternative energy sources like microbial cell batteries. They also head up a studio, Scilicet, in which a growing group of artists are exploring AI art alongside them. “The technology we’re building helps reshape how I paint, meditate, perform—and that changes the nature of the drawing entirely,” they say. “Allowing that feedback loop really catalyzes technical development, but also creative growth as well.”

More Must-Reads From TIME

Contact us at letters@time.com.

Yann LeCun
Abeba Birhane
Audrey Tang
Grimes
Demis Hassabis
EDIT POST