‘[data, set]’ showcases a speculative home automation system in which the smells, lighting conditions, audio and visuals within the environment are controlled by an AI-powered system which monitors visitors to the space in real time.
The piece tracks the emotional responses of guests — as well as how long they stay inside the room — to gage which combinations of media are most effective at capturing and maintaining their attention; as a result, the space is constantly optimising itself to present the most engaging version of itself to those inside. Deceptively alluring from the outside, visitors to the space are bombarded with a random, hollow assortment of media that has been grotesquely over-engineered to create an overstimulating, surface-driven experience.
The media (audio, videos and still images) on display consists of a randomized assortment of content extracted from miscellaneous real-world data sets; imagery such as x-rays, crowd fights, and dashcam footage that are actively being used to train neural networks. This project draws from ongoing research into datasets - who compiles them, how they are compiled, and what they can currently (and potentially) be used for.
Produced to accompany this exhibition was a short fictionalised essay contextualising the piece. This has been embedded at the bottom of the page.