I began working with AI in 2015 when I built a robot named D.O.U.G. 1, which stands for Drawing Operations Unit Generation One. I started by creating a system, Mimicry, that attempts to mimic me as I draw, before I became interested in exploring memory. I trained a neural network on 20 years’ worth of my own drawings and wrote an algorithm that could extract these lines as data. It meant that D.O.U.G. could react to my gestures in real time and draw in unison with me.

My goal was to experiment with this emerging technology, and I’ve continued to explore co-creation through subsequent versions of D.O.U.G. It’s currently on its fifth generation—the robot now responds to my brain signals through an E.E.G. headset and translates them into drawing. Working in this way challenges the notion of the drawn line as an intentional gesture.

With D.O.U.G., I’ve done performances, sculpture, works on canvas and video. I’m interested in exploring the relationship between humans and machines through the physical negotiation that happens when I draw together with the robot. As a performer, it’s always been important for me that my practice exists in the real world, rather than just being digital.

I understand the legitimate concerns with AI, especially about artistic data being used without consent. In many ways, I think learning how these systems are built—and thinking of that as its own creative process—is a way to keep both the hopes and fears associated with these technologies in mind. I view AI as a suite of tools, and because I work with systems that I’ve created myself, I can access a different depth of understanding.