By now, we’re used to certain artificial intelligence applications in our music-listening routine. We can hold up our phones when an unknown song is playing to learn its name, and we receive algorithm-based recommendations from streaming services for new songs based on our listening habits.
Letting the machines take over our music composition has always been dicier. There is a distinct human element to music that most people cherish; ceding the moment of creation to a computer would render it cold and hollow, according to conventional thought. That’s part of what’s exhilarating about Holly Herndon’s new AI-assisted album, PROTO: once again, Herndon has proved conventional thinking wrong.
For the past two years, Herndon has built, taught, and refined an artificial neural network housed inside a gaming PC nicknamed “Spawn.” The AI has learned enough from Herndon’s own voice and music to recreate a version of it by itself, and on PROTO, its generative sounds and Herndon’s input are combined with a large folk vocal ensemble. The result is a thrilling, very human album, combining new strange mutations with choir pieces that sound centuries old.
Having recently completed her Ph.D. at Stanford’s Center For Computer Research In Music And Acoustics, Herndon is now taking PROTO, and Spawn, on the road. How the AI will perform in real time remains to be seen. But make no mistake: this isn’t music strictly for the shifty, awkward, coding art-gallery crowd. For a rare upcoming show in San Francisco, when the beat to “Eternal” drops, expect to see a wave of dancing as the humans claim their space.