"Meet us at FBF 2024"
Blog
Blogs

The musical virtuosity of AI composers

Caption

As someone who is constantly trying to convince people of the wave of AI I believe is coming, I find that a lot of people— both lay persons and the tech savvy — struggle to accept that machines will soon be able to engage in complex creative endeavors. This is a completely understandable resistance, and I too feel the force of this gut feeling that there are certain things no computer will ever be able to replicate. Sure, they might be able to perform rapid calculations and analyze patterns in massive amounts of data, but how can these uncomprehending devices ever perform those feats of creativity and genius that are the pinnacle of human civilization? But reality, as usual, is stranger than we can imagine.

In 1981, when American composer David Cope hit a “composer’s block”, he thought it might be interesting to write a program to analyze his usual composition style and then use it to try to predict what the next portion of a particular composition should be. The trick was to generate new content that was similar to but distinct from what had already been created. The mechanics of this process relied on the technique of “recombinancy”, where existing pieces of music are combined into new logical sequences.

After creating some bland variations on his earlier work, he decided to apply his program to the canon of classical music, and ended up producing stunning compositions in the styles of the greats.

Here’s Vivaldi:

Bach:

And Chopin:

Perhaps a professional musicologist will be able to tell at once that these are fakes, but to my amateur ears they sound just as transcendent as any of the canonical works.

This is huge because we all associate machines and artificial intelligence with speed and mathematical computations, while keeping the artistic creations in a quite distinct category, where it is assumed machine learning cannot really compete. But Cope shows that this might not quite be true.

The more I listen to Cope’s algorithmic compositions, the more the sharp lines I had implicitly drawn between humans and machines quiver. I share the cognitive scientist Douglas Hofstadter’s befuddled reaction of listening to Cope, which he pithily and hilariously summed up in the poem:

Is music a craft
Or is it an art?
Does it come from mere training
or spring from the heart? 
Did the études of Chopin
reveal his soul’s mood?
Or was Frédéric Chopin
Just some slick “pattern dude”?

It is simply no longer feasible to think that the coming AI revolution will simply be an increase in speed of already existing processes, since it is by now clear that deep changes will have to follow in how we think about the world and the role of humans in it. Many creative pursuits are already being influenced and impacted by AI, and this trend is only going to intensify. Every person and every field needs to take seriously the writing on the wall, and try to find constructive ways to work alongside the coming technological changes.

Gain better visibility and control over your entire processes.
Retain control over your content; archive and retrieve at will.
Achieve a 20% cost-saving with our AI-based publishing solution.