As I have mentioned before, I study jazz piano. Jazz, more than most musical forms, relies on improvisation. It creates fitting rhythms and responses in the moment, avoiding mere formula.

No one knows how accomplished jazz musicians do this. But when they do it well, it’s amazing—and it’s nothing like the so-called “creativity” of recent AI attempts.

An artificial intelligence reporter reviews one such attempt, MuseNet, which

… shows how effectively such a model can capture and reproduce statistical patterns that reflect the character of something like a piece of music

Will Knight, “This AI-generated musak shows us the limit of artificial creativity” at Technology Review

But it’s not, as Knight notes, “real creativity.” He quotes Zach Lipton, an assistant professor at Carnegie Mellon University and an accomplished Jazz musician:

“It is uninteresting in precisely the same way as every generic ‘we trained an LSTM to generate” … “I don’t think there is anything here that a musician should find interesting.”

Will Knight, “This AI-generated musak shows us the limit of artificial creativity” at Technology Review

It’s regurgitation, not creation.

AI systems do not “create,” they extract patterns from the training data and then regurgitate (Knight’s term) “some statistical variation.” They can only copy, and then not very well. Artists have nothing to fear from AI if, as philosopher of mind Sean Dorrance Kelly says, creativity does not follow computational rules.

See also: AI creates kitsch, not art. And it’s the perfect tool for the job (Brendan Dixon)

Creativity does not follow computational rules

and

Does AI spell the end of the artist’s way of life?