'mixed, black & white' Stable Diffusion dancers
As an adjacent thread in my interest in creative technology I’ve also been speculating what AI could possibly mean in terms of dance improvisation.
Improvisation and instant composition in dance entails generating and performing the dance in relation to the number of given parameters in real-time. The ‘dance’ in itself is not written beforehand but the language, score or parameters that performers use to generate the dance are pre-established. In this way the dance unfolds spontaneously in relation to a range of evolving circumstances while remaining truthful or authentic to itself (the performers, the space, the music, the audience, etc). The performance is ‘different’ each time.
What could this possibly imply when AI is involved. Could an AI partner generate coherent improvisation?
Using an AI dataset at a future date could it be possible to generate my improvisation in a manner that would be authentic to me and truthful to the occasion? In what time is the creative moment of such an improvisation situated - Past? Present? Future or Continuous time? What are the notions of now, live or real-time in that case? Are those even definitive conditions? Can improvisation be pre-emptive or prospective? Is that still valid as improvisation ? Isn't it a bit paradoxical ?
The next stage of the process is a new collaboration with Niklas Fricke, a reseacher in AI and neurophysics at Champalimuad Centre, Lisbon. The Champalimuad Foundation intersects scientific research and healthcare.
This first stage of our collaboration at Bridges to the Unknown will focus on body movement and machine learning. At the end of our residency we will present work process at the third edition of Metamersion - Healing Algorithms, 16 - 18 may 2024.
