Hot off demoing AI voice transfer in a performance context at Barcelona’s Sónar Festival, American artist and musician Holly Herndon muses on vocalization through a machine learning process she calls ‘Spawning.’ “Unlike sampling, which is a reproduction of sounds sampled from a recording, Spawning is the ability to create works in the likeness of others by interacting with a [voice] model trained on them,” Herndon writes on Twitter. It’s “21st century sampling, with big implications.”