Musicians, Machines, and the AI-Powered Future of Sound
Can AI replace musicians and such?
There are alarming moods about possible replacement of humans by computers. However, some musicians make successful attempts for productive collaboration with AI.
In 2022, a human and an AI-based algorithm composed music together in Stockholm. David Dolan was playing piano into a microphone. A computer system, produced by Oden Bel-Tal a scientist from Kingston University, processed the music, extracted information on pitch, rhythm and tonality. Then the machine added its own attribute. It started improvising like a real person would. Sometimes it transformed the Dolano’s piano, some sounds were synthesized on the go. The performance sounded icy, diffusive, bizarre and structural. .
This time the scene appears to be surreal, considering the current artists-versus-robots rhetoric. One might have heard rumors about AI replacing editors, writers, producing SEO-written texts with errors. AI is also known for stealing illustrators, that sue Stability AI, DeviantArt and a few others for copyright violation. Additionally, AI-based FN Meka was taken off by Capitol Records after negative reviews, stating that the machine was too biased. Noam Chomsky recently stated that, AI was “the devil in electronic form”.
These concerns neighbor those, that AI will displace humans or that people, in control of other machines, will displace everybody else. Artists are more interested in supplementing human creativity, but not in the format “AI plays Metallica”. Collaboration is more promising, than competition.
Ben-Tal says, that creativity has many forms and spects to it. Particularly, it is about inspiration, revolution, ability, style. So, this is where AI might be helpful.
Belief that computer is capable of music production has been trendy for as long as computer has. Ada Lovelace was one of the first people to theorize that Charles Babbage’s steam machine could be utilized for other purposes. She thought that should science of harmony and music be used by the Babbage’s machine, something elaborate, scientific of any degree of complexity might come out.
Lejaren Hiller and Leonard Isaacson wrote the first book on the subject in 1959. It was about music experimentation with electronic computer. Ash Koosha, Arca and Holly Herndon used AI for their productions. Holly spoke to Wired about her available free of charge AI tool Holy+. She said, she preferred to think about it as an chance as opposed to it being scary.
Musicians have their opinions about ChatGPT and other AI-powered tools. Bogdan Raczynski after reading chatbots’ discussions with humans says that he detected fear, bewilderment, grief, alertness in the model’s answers. According to him, it’s not robotic emotions, but rather “human emotions, it evokes, are real”. He says, those feelings are more about concern and sympathy.
Bel-Tal states that his work is about alternative to :”human vs machine” discourse. He understands people’s reaction, since normally AI shows creativity level similar to a human, however he thinks of it as another tool, similar to the one going back in the past, the flute. For instance through the use of turntables, artists found out that those could be used for scratching, sampling. New genres were invented.
In this context, intellectual copyright might be rethought - Google refused to issue its MusicLM model (turning text into music) due to possible copyright infringement issues. In 2019 Ben-Tal offered the readers to imagine a situation, when AI would store all recorded music and could retrieve or generate any sound on request. Is there a place for songwriters in a world like that? Can music producers protect themselves from plagiarism? Should the audience be told, when AI is used in the articles?
Despite all this, AI offers great opportunities. They say, AI can be used to improvise with musicians or AI production can be used as an inspiration, even in genres, a pro musician might not be familiar with.
In the long run, machines can execute a wilder vision: execution of an artistic perception of music and reality. Composers might want to use it, since music production is a resource-intensive task. AI might help, because it can a be a plug, which one puts in and takes it out.
Fairly recently, day to day and omnipresent algorithms already distort the industry. Cory Doctorow recently wrote about Spotify impact on music: how artists are encouraged to stop producing “chill vibes” music and train Spotify to tell people what to listen to. In this context AI will become an enemy of musicians. Destiny of the industry becomes unknown should Spotify launch its own AI artists.
Raczynski says he would rather stay trendy rather than be consumed by it. He says, that since he can’t escape the trend, he would rather develop a relationship with one (AI), hoping it is of reciprocity, not self-centeredness.
AI Catalog's chief editor
Share on social networks:
