When I first presented the idea of dancing with an AI avatar, I was asked whether I was actually dancing with another presence or merely dancing alone in front of a video projection. The question struck me harder than I expected. It sat between curiosity and skepticism, but it immediately exposed something deeper: What does it mean to share movement with another—especially when that “other” has no flesh, no heartbeat, no breath?
That moment became the seed of a much larger investigation. It forced me to examine a question that sounds simple but is philosophically dense: What is the difference between dancing to someone and dancing with someone?
To dance with someone implies reciprocity, presence, and a mutual shaping of attention. It implies that the other body—whether human or not—is capable of altering your choices, redirecting your momentum, inviting you into a shared kinetic imagination.
This distinction pushed me far beyond choreography. It was not about creating AI tools or impressive digital animations for a dance show, but about rethinking ideas I believed were solid: the nature of the body, the contours of creativity, and even the definition of dance itself.
Can an AI dance?
We often imagine dance as something inherently human—a lived, embodied intelligence expressed through muscles, breath, rhythm, and memory. An AI has none of these. It does not feel the weight of gravity, the pressure of the floor, or the trembling of anticipation before the music begins. And yet, when a machine generates movement, something peculiar happens: we see motion that resembles intention. We project feeling into gestures born from code.?
In this sense, the question “Can an AI dance?” becomes less about the machine and more about us—about the human capacity to read meaning in movement, to empathize with patterns, to perceive life in things that do not live.
Yes, perhaps an AI does not dance like us. Perhaps it dances otherwise.?
How, then, can we dance with an AI?
A dance with an artificial partner does not begin with shared emotion—it begins with shared influence. Movement must circulate between bodies, even if one of them is digital or mechanical. Back in 2021, I collaborated with researcher Benedikte Wallace to design a computational model that could allow me to experience a reciprocal flow based on kinematic rhythm. We called it Dancing Embryo: a minimalistic avatar that generates its own movement vocabulary and responds to mine. When I dance with this AI avatar: I move, it answers; I shift, it transforms. Something unexpected emerges—something neither of us could create alone.?
This exchange—uneven, imperfect, yet undeniably relational—becomes a dance. When we premiered this work at the Leverhulme Centre for the Future of Intelligence in 2022, once the audience grasped the technical setup, their attention turned to the human experience behind it: someone asked, “How does it feel to dance with a machine?”. Throughout the international tour that followed, audiences echoed that same curiosity, appreciating the chance to witness a real-time encounter rather than a pre-scripted demonstration.
But not every encounter with an artificial mover carries this quality. If the dancer cannot truly open the space for a shared exchange, the AI slips into familiar roles: a puppet, a mirror, a tool. In those moments—even with impressive machines—the dance flattens into a single direction, one body simply steering the other.
For something richer to appear, the AI must not only answer but also offer—nudging my movement onto paths I would not have taken alone. And for a human–AI dance to unfold, the performer must come in ready to listen, to guide, and to be moved in return. This reminds us that it is never just about powerful technologies but about the fragile possibility of meeting another presence—however strange—in a true act of exchange.
To navigate these differences, I developed what I call the Machinic Movement Matrix (MMM): an analytical framework to map how machines participate in creative movement. This conceptual tool looks not at the beauty of the final choreography, but at the underlying relationships between humans and AI during co-creative processes.
The MMM helped me realize that dancing with an AI is not a metaphor. It is a literal, kinetic negotiation between two systems—one biological, one artificial—meeting in the shared space of motion.?
The space between worlds
At times, when I improvise with an AI, I sense a strange companionship. Not human, not sentient, not emotional—but something like a presence that pushes me to new possibilities. It disrupts my habits. It offers gestures my body might never have invented. It makes me rethink rhythm, weight, and the logic of movement itself.?
In those moments, I am reminded that dance exists outside human bodies—yet not every moving thing can become a partner. We have always sensed dance in storms, in animals, in shadows, in machines. But sensing dance is not the same as dancing with; only a few forms of motion can meet us with the persistence and reciprocity a shared dance requires.
So can an AI dance?
Maybe the better question is:
What new forms of dancing become possible when the other body is artificial?
Dancing with AI agents opens a widening field where dance becomes a conversation between worlds: flesh and algorithm, intuition and computation, presence and simulation. It challenges us to reconsider what a body can be, what creativity can feel like, and what forms of togetherness might emerge when movement binds beings that do not share a common nature.
In that sense, AI does not diminish dance—it reveals its deepest truth:
that dance is, and has always been, a meeting place for different ways of being alive.
For further details on this research and related publications visit: https://diegomarin.art/portfolio/dancing-embryo/?
1. Marin-Bucio, D. (in press). Embodying the artificial: dance and human-AI co-creativity. Intellect Books.