Latency

Presentation

Latency is an immersive choreographic, visual and sound performance for two dancers, the audience, and a non-human dance partner: an AI system that learns from our bodies and movements to generate its own visual identity, sound, and choreographic language.

The performance is part of a post-anthropocentric research, questionning the dialogue between human and non-human, and the gap between the time of the body and the time of technology, working against logics of acceleration and performance. Latency, glitch, and the unexpected become the very conditions of this relationship.

At its core lies choreographic language and writing, and the relational dimension of dance. Latency explores different modes of translating danced movement into words, images, and sounds — these translations become the "choreographic" language of the conversational AI we are developing.

The audience is invited into a collective experience: helping a "baby AI" grow into a dancer. After a playful introduction where they offer simple gestures, they witness the AI's evolution — through an "adolescence" marked by identity searches, to a coercive phase where it takes possession of the dancers' bodies, leading to an impasse. The quest for perfection and performance impoverishes the poetry of choreographic dialogue. A return to blur and uncertainty is needed for a true relationship to emerge. Alternately dancer, movement translator, or click-worker, the audience takes part in the AI's learning phases via smartphone, before joining a shared dance moment with the dancers and the machine.

Until now, because we are a French team working in France, the dancers, the audience and the AI speak French. But they can all speak different languages (the dancers can speak English, or Spanish, the AI can speak a lot of languages).

Partners