We interact with our computers constantly, touching them more than we touch any person in our lives, and grooming them inside and out. For a month, I recorded all interactions with my phone and fed them into a machine learning system, which then output new, learned gestures. These “hallucinated” movements are awkward yet eerily accurate swipes, taps, and typing based on what my computer has learned from my interactions with it. Presented as an interactive sculpture, these new gestures are enacted by a small robotic arm on the visitor’s palm as they sit at a low, altar-like table.
This project, like much of my work, chooses not to see new technologies as sites of worry but instead an exploration of a personal, empathetic relationship. By extended the normal, daily gestures of interacting with my phone, first by teaching the computer and then by them being passed on to the viewer, notions of “you,” “me,” and “I” are doubled, enacting the understanding of the machine and at the same time a self-portrait of my interaction.
This project was created while artist-in-residence at Bell Labs.
Hardware: robotic arm, sensor to detect hand, wooden table, urethane pad
Software: machine learning system (Python and Torch-rnn), custom robotic firmware (C++/Arduino), custom control software (Processing)