This project is an attempt to talk to artificial intelligence via sounds and about sounds. And the focus of this conversation is something only human can experience – the first “hearing” sensations that a human embryo feels inside the womb. Researchers say that the first thing that human “hears” or rather feels is their own heartbeat and their mother’s heartbeat. In this work we’re trying to convey this to AI and get its response.

The audio part is based on two rhythmic lines: the one that symbolizes baby's heartbeat and is panned right; and the one that symbolized mother’s heartbeat and is panned left.

The melodic part is also divided into two main lines: the one that is played by a human and has more natural piano sound; and the one that is played by AI and has more synthesized sound. These fragments follow each other and repeat several times during the performance. They are also reflected in the visuals. The “human” melodic part is embodied in a metaphor of a birth, a blossom. The “machine” part features pictures created by a neural network. Both parts are accompanied by audio reactive generative graphics, and they combine in the end, where human and machine play together.

I used Ableton Live to produce the audio track. It features field recordings and sounds synthesized by analog synth Korg MS-20. For AI music, I used Google Magenta Studio plug-in for Ableton Live. This performance was played live and I sang and played with audio effects during it.

AI-generated pictures were produced via Artbreeder. Generative graphics were made in Touchdesigner where it was all put together.

Generative graphics by Tatiana Anichkina

Music, idea, live recording, visuals by Yulia Stepanova