As the late spring sun set over Lake Léman, beatbox rhythms pumped out of the United Nations Building on its southwestern shore. Inside, hundreds of artificial intelligence researchers, UN staff and curious locals listened, watched and tapped their feet as London-born composer and human beatboxer Reeps One “battled” against an AI-powered real-time music generator trained on his own riffs.
Reeps and others working at the intersection of art and AI performed at the May 29 cultural event AI Pushing the Limits of Artistic Intelligence, part of the AI for Good Global Summit 2019 in Geneva, Switzerland. The evening showcased talented artists and cutting edge AI technologies as part of the growing trend of human and machine collaboration on creative expression.
An award-winning visual and vocal artist, Reeps One uses his voice as an instrument to create a variety of sounds, beats, and melodies, mimicking the effect of digital beatboxes. He is an artist in residence at at Nokia Bell Labs, where he has been experimenting with voice and machine learning technology.
Last year, Reeps launched the art project Second Self with artist and programmer CJ Carr, developing a deep learning music system trained on his various beats and voice sound inputs to replicate and predict his musical compositions.
Reeps One records inside the Nokia Bell Labs Anechoic Chamber, which is one of the quietest rooms in the world and where he generated 140 variations on a kick drum sound in pursuit of beatbox excellence.
In his Pushing the Limits performance Reeps faced off against his AI-powered student beatboxer, joined by dancer Robozee, whose robot-like moves thrilled the audience.
A Self-aware art installation?
Also at the event, Creative Director at the Berlin-based Waltz Binaire art studio Christian “Mio” Loclair introduced his installation Narciss, telling the audience“We are very interested in who we are… So we thought let’s initiate this moment from a digital perspective. What if machines start to wonder about who they are just like us.”
Inspired by Alan Turing’s imitation game, Loclair designed Narciss as an imitation character. The installation comprises various electronic components and processors arranged before a circular mirror. Locair says that “by constantly panning and zooming, Narciss receives a feed of different perspectives and sub-regions of its hardware.” The installation uses a natural language processing system to deliver interpretations of its reflection on a display screen — a self-narration symbolizing the device’s budding self-awareness.
Narciss has made some 2000 assumptions about its physical existence as seen in the mirror, some of which might indeed suggest a search for an identity. It has variously described itself as “a laptop computer sitting on top of a desk; a man looking at a laptop; a person standing in front of a tv holding a game controller; a close up of a person holding an open umbrella; a shadow of a person in the mirror; a person’s reflection in the mirror; a person is taking a picture of their reflection in a mirror,” and so on.
Video of the Narciss project is on Vimeo.
Journalist: Fangyu Cai | Editor: Michael Sarazen