Every day humanity inches closer to the possibility of an AI-driven robot apocalypse. Now artificially intelligent machines have made one more step towards superiority over nature by creating new sounds that no one has ever heard before.
As Wired reports, members of Google Magenta, the team of researchers using Google Brain to try to create AI systems that will make art, have trained their system to create new sounds based on the mathematical makeup of preexisting sounds. Called NSynth, the program breaks down the sound of any one of a thousand traditional instruments, and melds them to create sounds that no human has ever heard.
According to the researchers, this is much more than just layering two tones on top of one another. NSynth actually turns the sounds into algorithms and adjusts them to make them fit together into one aggregate sound.
Whether it’s a hybrid of a flute and a bass guitar, or a flugelhorn and a glockenspiel, the resulting noise would be difficult or impossible for a human to reproduce. Although whether that qualifies as art is in ears of the behearer.