Three Body Problem: Art, Artist And AI  

The AI-generated productions of songs in the voice of artists of a bygone era has given us with what I would like to term as “a three-body problem”, i.e., What is art? Who is the Artist? And, can AI produce original Art? In the context of this particular song – who is the real artist – the singer i.e., the one who actually sang it in the first place, or the composer who composed music using AI, or AI itself? 

In his pathbreaking essay, ‘The Work of Art in the Age of Mechanical Reproduction’, Walter Benjamin laid down a few rules for the art to be called as such. He argued that Art carried uniqueness and aura with it. The authenticity of Art came from the fact that it could not be reproduced. Although, any piece of art that was created borrowed and based itself on the elements of tradition, yet the impossibility of its reproduction or identical copy gave art its essence. 

This changed with technological revolution in the eighteenth and nineteenth centuries when printing and reproduction became commonplace. The aura of art withered at that time because of its reproduction. However, with AI, technological intervention has now reached beyond just reprinting or copying which still gave some value to the works of art even after their reproduction. Copying or recreating a piece of art that has been originally created by the artist is different from creating a piece of art that was never created by the artist originally. Now, in this process, the fundamentals of originality have been removed. The phrase Art has its own life has taken another meaning with AI. Art under AI has become a separate entity of its own. The Artist is dead. Do we have the consent of the artist? Do we know what variation or murki the late Muhammad Rafi would have taken while singing this song? We know how emotions of the singers/artists affect their art, which actually creates the originality of the piece. In this case there is no emotion -at least at the level of interchange with AI.  AI does not have the inner struggle to achieve perfection which is a hallmark of an artist. If not, then should we take this song as another copy of the artist's song/voice or does this have something more to offer?  

Singers such as Kishore Kumar and Rafi were people with a unique voice, a voice that cannot  be reproduced by anyone else. AI replication of people’s voices also gives rise to some ethical dilemmas. Firstly, when songs are made using thevoices of the departed artists, the issue of consent becomes extremely pertinent. Is it ethical/correct to use the voice of a departed artist when millions of people find the old songs as a connection to their beloved artists and as markers of the past? These reproductions take away the sense of time from people. There also lies the issue of the lyrics; some singers would absolutely refuse to sing vulgar or nonsensical lyrics, but AI has no such self-restriction or reservations. There was a sense of personhood which was/is associated with singers/artists. They were more than just their voices –they were people. Now, what happens to this sense of personhood in the mechanical reproduction of the song?   

In current times, the question – can AI think, feel, or cognise – is a question that bothers both philosophers and laypersons. The use of AI to make music, art, and videos, and the quality with which it produces such pieces of ‘art’ is unprecedented. Some may say that the production quality is perfect, having removed the human factor where some imperfections crop in. AI tools are much more dynamic. Using principles of machine learning, AI can quickly learn to replicate different art forms and artists, thereby making it more than just a mindless tool. When given a command, AI can write essays, generate fictional content, compose lyrics to a tune of its algorithm; and when it does so, it essentially takes over the role of the artist. 

We must ask ourselves the question: is the artist relevant anymore? Has the tool used to create Art has itself become the Artist? It is not just a question of copyright or royalties, but gives rise to a triangular philosophical dilemma– a three body problem. In the age of AI– What is Art? Who is the Artist? And what is AI for Art?  When we claim that Art has its own consciousness, do we then mean that AI has become conscious? Can we then give personhood to AI? 

There is an argument which proposes that AI, particularly Generative AI, carries consciousness, it has a body (in the metaphysical-digital realm) and mind of its own. Can we then treat AI as a living being-as a person? The Samkhya - Yoga schools of Indian philosophy believed that humans are a combination of matter and consciousness. According to the ancient Indian philosophical schools, meaning was produced when the matter evolved in the presence of consciousness. Apart from consciousness, a human or sentient being has a Chitta (roughly translated to mental makeup) which consists of intellect, ego and Mind (Buddhi, Ahamkara and Manas). This human mind goes through several stages of emotions to produce Art. The unpredictability of the fact that at what point of time, what combination of emotions gives birth to what form of art –is the basis of the beauty of Art.  AI does not work with unpredictability or ‘out of the box’ thinking. It produces, or reproduces the patterns that already exist in the digital world.  Human history has shown us that we are generally restless to break free from old patterns and always strive towards new possibilities. Even if Generative AI reaches a level of unprecedented levels of creation, the mother consciousness will always be human. Human world will always be marred by contradictions and ethical dilemmas.  

Until now, AI is the best commodity for a capitalistic society. It doesn't require wages; it doesn't get tired and doesn't face the problem of ageing as human artists face. It can reproduce like factory chains of assembly lines. However, as the famous German poet and playwright Bertolt Brecht once wrote in the context of war machinery, that it’s useless without (hu)man because (hu)man can think. AI cannot work without humans. If there was no late Mohammad Rafi, then there would not be the recreation of a song from the movie Animal in his voice. AI can sing, but only with the human voice. In other words– Humans are AI’s voice.