Artificial intelligence and musicians don't have to be enemies

Artificial intelligence and musicians don't have to be enemies

Last November, at the Stockholm University of the Arts, a human being and an artificial intelligence (Ai) played together. The performance began with musician David Dolan playing a grand piano into a microphone. Meanwhile, a computer system designed and supervised by Kingston University composer and researcher Oded Ben-Tal "listened" to the piece, extrapolating data on the pitch, rhythm and timbre of the sounds. Then he added his own accompaniment, improvising just like a person would. Some sounds were elaborations of Dolan's piano, others were new synthesized creations for the occasion. The result was a performance that felt ambient, with a chill and creepiness, but also with its own structure.

The scene – an Ai and a human being peacefully cooperating – seems irreconcilable with the Current Artists vs. Machines Debate. You've surely heard that artificial intelligence is already replacing journalists, churning out articles full of errors. Or stealing the work of illustrators, who are in fact suing companies like Stability Ai, DeviantArt and Midjourney for copyright infringement. And more of computers taking to rap (or at least trying to: "robot rapper" Fn Meka was dumped by his record label, Capitol Records, following criticism that the character was "an amalgam of crass stereotypes"). In the most recent intervention in the debate, the famous linguist Noam Chomsky declared that ChatGpt shows the "banality of evil".

These anxieties fit neatly into fears about automation and the possibility that machines will replace people, or, rather, that the people who control these machines will use them to replace everyone else. Yet some artists, first and foremost musicians, are quietly taking an interest in how these models can complement human creativity, exploring how Ai and humans could collaborate rather than compete.

Alternative narrative

" Creativity is not unique – Ben-Tal explains in connection on Zoom -, but it includes many different aspects. It includes inspiration, innovation, art, technique and work. And there is no that's why computers can't be profitably involved."

Hypotheses about whether computers could compose music have been around forever. Mathematician and writer Ada Lovelace once theorized that Charles Babbage's steam-powered analytical engine, considered by many to be the first computer, could be used for something other than calculations. According to Lovelace, if a way had been found to adapt the “science of harmony and musical composition” to Babbage's machine, it could have "composed elaborate and scientific pieces of music of any degree of complexity or extent".

The first book on the subject, Experimental Music: Composition with an Electronic Computer, written by American composer and professor Lejaren Hiller Jr. and mathematician Leonard Isaacson, was published in 1959. In popular music, artists such as Ash Koosha, Arca and Above all, Holly Herndon have used Ai to enrich their work. Last year Herndon spoke to about his “AI-enhanced vocal clone,” Holly+, succinctly illustrating the tension between tech and music: “There's a scary narrative surrounding a lot of this stuff. and dystopian – he said -. I am trying to show a different aspect. It's an opportunity".

Musicians have also reacted to the general unease generated by ChatGPT and Bing's new Ai chatbot. Reading the transcripts of the discussions that have gone viral between chatbots and humans, Bogdan Raczynski tells by email about having recorded "fear, confusion, regret, shyness" in the model's responses. He doesn't think the chatbot has feelings, but that "the emotions it evokes in humans are very real". In response, he posted a "serious performance live comforting for AI”.

Ben-Tal says his work presents an alternative to “man versus machine storytelling.” He admits that generative AI can be unsettling because, at least on a superficial, displays a kind of creativity normally ascribed to human beings, but adds that it is simply another technology, another tool.For the composer, generative Ai is basically no different than turntables: when artists i discovered that they could use them to scratch and sample their sounds, they created whole new genres.\

Challenges and controversies

In this sense, copyright may need to be rethought in a way substantial: Google has decided not to distribute the MusicLM model, which turns text into music, due to the "risks associated with generating music, especially the potential misappropriation of creative content". In a 2019 research article, Ben-Tal and other researchers asked readers to imagine a holodeck for musicians, capable of storing all music ever recorded and generating or retrieving any possible sound on demand. What place do songwriters have in this future? And in the meantime, can songwriters defend themselves against plagiarism?

However, these templates still have interesting creative capabilities. For now, Ben-Tal continues, musicians can use AI to improvise with a pianist and come up with compositions that are beyond their capabilities, just as he did. Or they can draw inspiration from the compositions of an Ai, perhaps for a genre they are not familiar with, such as Irish folk music .

In the long run, however, artificial intelligence  could satisfy a more daring (albeit controversial): effortlessly realizing an artist's vision. " Composers, you know, have ideas about the music they would like to create, but then translating them into sounds or scores, realizing those ideas, is a rather laborious task – says Ben-Tal -. If it were enough to connect a cable to achieve all this, it would really be awesome and wonderful.”

Meanwhile, still pretty basic but pervasive algorithms are already turning the industry upside down. Author Cory Doctorow wrote about Spotify's stranglehold on music: Playlists, for example, encourage artists to ditch albums for collections of select songs and train audiences to let the platform tell them what to listen to. Introduced in such a context, artificial intelligence will be the enemy of musicians. What will happen when Spotify starts distributing and promoting their own Ai artists?

Raczynski hopes to ride the wave rather than be swept away by it: "Perhaps indirectly, whether I like it or not, I'm recognizing that unless that does not leave the network, I have no other choice but to develop a relationship with AI – he admits –. My hope is to build a mutual and non-univocal relationship ".

This article originally appeared at UK.

Powered by Blogger.