Shadowed Spacetime is an experience through which visitors come together to participate in the act of improvisatory, communal sound creation in order to speak to questions of how “feedback” functions in machine learning contexts to “improve” the user experience. Supposedly AI tools can only evolve by “learning” from us – we are sometimes overtly directed to train such models by AI companies themselves (for example ChatGPT “asking” us to evaluate its responses), other times the “training” comes in the form of information stolen from large pools of our collective data and creations, and at other points this training emerges through the ways we use these systems — the prompts we suggest, for example.

Shadowed Spacetime is a station in which users provide “feedback” to the MLTB system to “improve” the overall user experience — but this improved experience is hopefully felt through a collectively devised piece that allows users to consider who else has been in the space through their echoes.

Credits:

  • Concept & Creative Direction: Enongo Lumumba-Kasongo
  • Physical station design: David Mesiha
  • Developed within the Data Fluencies Theatre Project