The future of music is cyborg

If science fiction has been speculating and imagining new technologies for 200 years, why wouldn't these also permeate artistic creation? Today, there is much discussion about the possibility of artificial intelligence creating art—not so much in the technical sense, as we have various examples of projects that have proven this is already a reality, but in the philosophical sense of what it means to make art and what creativity is. Although there are some possible paths to consider this question, a spoiler is that even in a doctoral program in visual arts, I still hear many people debating what art is. As Merleau-Ponty argued in his work, philosophy, like art, is a discipline that asks questions that are left open—its creations do not come into the world to provide answers, but to fuel more questions and spark new possibilities.

In 1924, Aldous Huxley introduced in "Brave New World" the suggestion of a musical genre called synthetic music reproduced by a machine. In an excerpt from the book, the author wrote:

“In the synthetic music machine the sound-track roll began to unwind. It was a trio for hyper-violin, super-cello and oboe-surrogate that now filled the air with its agreeable languor. Thirty or forty bars–and then, against this instrumental background, a much more than human voice began to warble; now throaty, now from the head, now hollow as a flute, now charged with yearning harmonics, it effortlessly passed from Gaspard’s Forster’s low record on the very frontiers of musical tone to a trilled bat-note high above the highest C to which (in 1770, at the Ducal opera of Parma, and to the astonishment of Mozart) Lucrezia Ajugari, alone of all the singers in history, once piercingly gave utterance.”

In an article published on the Museum of Imaginary Musical Instruments website, the author argues that this reference to synthetic music could be related to the expression "ether music," which was widely used for the ethereal tones produced by early electronic instruments, such as the Theremin. Consequently, by mentioning imaginary instruments like the "hyper-violin" and "super-cello," the author might have been relating the sound of original instruments to how they sounded when reproduced on gramophone recordings or radio broadcasts, which at that time, in the early 20th century, were seen as "substitute" tools for what would later be called "live music."

In this sense, could we think of artificial intelligences, robots, and virtual beings as "substitutes" not for "live music," but for music made by living beings? This also raises the controversy over the definition of life and the contrast between organic and synthetic life, but regardless, what companies like Auxuman are doing is precisely creating virtual singers from digital avatars. The premise is not new, considering the wave of virtual influencers that already populate social media, as well as the Vocaloid characters that give personality to music created in human voice sample synthesizer software. In the case of Auxuman and its avatars Yona, Mony, Gemini, Hexe, and Zoya, these are artificial intelligences that scan social media posts to then create their own compositions. Or something like that, since they still go through a process that is heavily supported by human intelligence.

Ininterview for the Digital Trends website, Ash Koosha, the music producer responsible for the project, says that Auxuman is creating "the next generation of virtual entertainment." On September 27, they released a first album that will be followed by monthly releases on channels like YouTube, Soundcloud, and other streaming platforms.

In his report on the case, Luke Dormehl raises the fateful question of whether these avatars are truly creating a work of art or if their compositions are genuinely creative. In his words, it is difficult to define creativity. There are few metrics to determine whether something is creative or not, but for now, we already have auctions where a painting created by an artificial intelligence art collective was sold for $432,500. But if artificial intelligence is created to take on tasks and jobs that are considered unhealthy or dangerous for humans, why create programs that would be capable of replacing us in activities we would like to do? Here, the reflection becomes more ethical than technological; after all, the ability to develop a technology does not necessarily justify the importance and relevance of developing it. This is not always thought of through that lens, however.

In the case of Auxuman, Luke Dormehl questions whether art made by machines could sell more than human artists, which he believes might not be the case. After all, it would be the same way that independent and obscure films are more appreciated by cinephiles who, at the same time, minimize the work of Hollywood directors who, nevertheless, generate much higher revenue than the former. Bringing this to our Brazilian context, perhaps artificial intelligences can replace the art of Romero Britto, but it might take a little longer to develop a more elaborate algorithm to create an arthouse film.

On the other hand, Dormehl also problematizes the ambiguity in what we understand by originality. “A human artist who has deep knowledge of their field and makes references to it in their work is a way of being considered creative. A machine is considered uncreative because it is not trying to express a particular emotion and is not using its creativity on deeper themes. But literary critics have long spoken about ‘the death of the author’ and how the work only exists in the reception of the audience.” It is a point to consider, especially after the deconstruction of the meaning of art by the modernists. After all, to what extent do we today understand an artistic work as something more linked to the classical and romantic model or closer to the modernist avant-gardes and contemporary artists?

Interestingly, Koosh believes that since artists are also governed by commercial and economic concerns, it is not always possible for them to innovate and truly think outside the box. For the CEO of Auxuman, artificial intelligence could be a way to make the exploration of mixtures and combinations of forms and styles cheaper, in order to generate different and interesting sounds. Dormehl finds this a bit of an exaggeration, especially considering the result of the first Auxumans album—nothing really far from the standard of what is pop today. But, in any case, machine creativity has already been used in the past to develop hypotheses for new medications or even new components for satellites. The journalist's provocation, therefore, is the possibility of allowing machines to create new, unimaginable genres and making it commercially viable for people to consume them. But even this statement is controversial; after all, to what extent is Koosh aware of internet microcultures and underground genres? On Last.fm, for example, there are tags for bands with fewer than 100 listeners and so on. There is an entire subculture of people who seek to listen to music and consume the work of little-known artists precisely because they are almost invisible.

But just as the Auxumans still depend on a large amount of human intervention in the curation of the content that these artificial intelligences aggregate, Dormehl comments on how this process resembles techniques like the “cut-up” of William S. Burroughs, with which the author, in the late 1950s, began to explore new text formats by cutting up written parts and rearranging them into a new text with a new meaning. “Artificial intelligence does this kind of remix in an even more interesting way. It adds a level of complexity that further removes control from the creator. It is as if another, almost intelligent entity were occupying a role in the production process,” the journalist writes.

Dormehl also cites the concept of “decaying sense” referred to by Thomas Hobbes when speaking about imagination, which would mean that it could be defined as an obscure memory of something, often mixed with other memories. “This is what makes artificial intelligence creativity so interesting. Artificial intelligences that create jazz choruses,computer games,imaginary monologues inspired by TV shows, and more fragments of something recognizable to us, and then mixed and rearranged using computational logic. It is a fascinating example of how humans and machines can work together to produce something new.”

From this, Koosh concludes that, in reality, machines are creative, but in a “supplementary” way. For him, “human creativity will be fundamentally different due to our biological nature, our needs, and our intentions. Our memories are recorded and retrieved in a very complex system in our brain, but machines can start with rules and compositions with tangents that human creators could take months to think of. Machine creativity comes to maximize possible results that we humans, as curators, can look at and choose whether or not to apply in a creative process.” In other words, once again we see the centaur logic applied to technological innovations: a human together with a machine will always be more powerful than the two agents alone.