I, for one, welcome our new robotic overlords

What a fascinating article Ryan Blitstein has written for Miller-McCune about a computer program that composes original music (with sound clips, so you can judge whether you think the algorithm—named Emily Howell—is good at it).

Emily Howell isn’t stealing creativity from people, he [David Cope, the programmer] says. It’s just expressing itself. Cope claims it produced musical ideas he never would have thought about. He’s now convinced that, in many ways, machines can be more creative than people. They’re able to introduce random notions and reassemble old elements in new ways, without any of the hang-ups or preconceptions of humanity.

At one Santa Cruz concert, the program notes neglected to mention that Emily Howell wasn’t a human being, and a chemistry professor and music aficionado in the audience described the performance of a Howell composition as one of the most moving experiences of his musical life. Six months later, when the same professor attended a lecture of Cope’s on Emily Howell and heard the same concert played from a recording, Cope remembers him saying, “You know, that’s pretty music, but I could tell absolutely, immediately that it was computer-composed. There’s no heart or soul or depth to the piece.”

-”Triumph of the Cyborg Composer”

A few reactions:

1. I wonder if in a decade or so, we humans will be creating compositions that are flawed on purpose—music and sculpture and stories that are just a little unpolished, that show a touch of human frailty in face of robotic perfection, like exaggerated stitches and buttons now in fashion on etsy to stress that something is homemade. In ten years, will writers make type-os, switch tenses, and dangle clauses on purpose because a computer would not?

2. The most fascinating question for me here is the question of authorship. The creative process through which the music is produced relies on David Cope to offer his program an idea—a musical structure, perhaps, or a series of notes. The machine works this idea into a piece according to the rules that Cope had coded, then produces a few thousand of these workups in the time that Cope has a cup of coffee or takes a shower.

But according to the article, Emily doesn’t know which one of these thousands is the best. It (she?) can’t even tell if any of them are good at all because as a machine, at this stage of development, she can’t enjoy them. The discernment and taste are all Cope’s, and if we like a piece of Emily Howell’s, we like it because that piece is the one that Cope has selected. So who’s the author? If the music is in fact art, is it art created by David Cope or Emily Howell? Is Emily Howell just another tool for idea transcription? Is authorship possible without complete autonomy?

Here, of course, we come full circle, to the original assumption with which Cope started—that none of us are fully autonomous when we engage in the act of creation. That we create art through a synthesis of our influences, by plagiarizing and rearranging. Good artists borrow, great artists steal.

With Emily Howell, however, the question is practical. She can’t evaluate her own work and she can produce a metric ton of it. So if we rely on her to create music, we are just trading one problem for another: instead of struggling to create, we’ll be struggling to select the one great piece from among thousands of pages of crap or, worse, mediocrity. And boy, as an occasional lit mag slush pile reader, do I ever not envy Cope that one.

3. The problem of selection might of course be solved by crowdsourcing. Imagine a web site where the work products of an AI get posted and then sifted by thousands of humans. If we let them claim whatever they found, it’d be like the new California gold rush. Would we offer thousands a new, previously inaccessible path towards the act of creation? Would we get more literature, music, and art, better and faster? And can you just imagine the flame wars?