Press skip on AI music

In a world where machines can produce perfect, functional output, the value of art lies in its humanistic creation and appreciation, says Jason Walsh

Blogs

Image: Andrea Piacquadio via Pexels

Oscar Wilde said that art was useless. Let’s hope that he was right, as it may just save us.

Wilde was articulating an argument known as ‘art for art’s sake’, one that has proved controversial, particularly since the 1960s, the nub of which is that, unlike design, works of art have no utilitarian function and instead have a purpose in themselves.

This can all become a little confused when you throw money into the mix. Despite laments at the decline of the recording industry, few would mistake record companies for benevolent societies: art, whether high or low, has long served capital. More precisely, it necessarily exists in relation to power and cannot be entirely separated from questions of who controls cultural production – and to what ends.

The question has recently been made all the more real following the success of the Velvet Sundown, an AI-generated music project that went viral on streaming platforms without disclosing its artificial nature, sparking a debate about transparency, authenticity, and the displacement of human artists in the digital music industry.

If this act was a conceptual joke it would put one in mind of the KLF – latterly known as K Foundation – it would have been a good one: the true shock of Velvet Sundown wasn’t its financial success; it was its philosophical implication.

This is not entirely new territory. The purpose of art is one of the most significant themes in philosophy and criticism. The disputes are enormous, but in one of those rare instances where thinkers across the political spectrum speak in chorus. Looking at just a few of the most important art critics of the 20th century, there is in their work, despite their disagreements, a thread that we would do well to pull on.

Critics across the political spectrum agree on art’s essentially human nature. Marxist John Berger saw art as a human cultural product that “becomes a meeting-place of the invisible, the irreducible, the enduring”. Robert Hughes argued art exists to “close the gap between you and everything that is not you”. Even John Carey’s critique of artistic elitism remained fundamentally humanistic.

Roger Scruton – a conservative philosopher who wrote extensively about aesthetics – argued that art was not “amusement”; rather it was our cultural inheritance and demanded our contemplation.

Rejecting easy consumption

So that’s what the experts said, but where does this leave the rest of us, whose cultural landscapes are now being altered by AI?

My argument is that art is not about the object. Art is not a fungible thing, be that a song, a painting or a poem. Instead, it is an expression of human experience and struggle, including the struggle to create itself.

I suspect some others may agree with me. Computer scientist, philosopher and inventor of the chatbot Joseph Weizenbaum was a harsh critic of AI as far back as the 1970s, with his book Computer Power and Human Reason arguing that some activities should remain fundamentally human, regardless of whether machines could perform them more efficiently. Weizenbaum argued that the dreary logic of computers creates its own reality, and that this reality is not at all that of humans.

In this we can hear the echo of 19th century critic John Ruskin, who wrote: “You must either make a tool of the creature, or a man of him. You cannot make both. Men were not intended to work with the accuracy of tools, to be precise and perfect in all their actions”.

Weizenbaum said in a 1987 interview with Der Spiegel: “The thought that a Mozart sonata, or a Hoelderlin poem, that perhaps describes the longings and delusions of the human spirit, would come from a computer is terrible.”

Discussing this with me, TechCentral.ie editor Niall Kitson said: “What if harsh atonality was the only thing separating genuine creativity from AI?” In other words, what if the very flaws and difficulties that AI would optimise away were all that was left to make art human and meaningful?

While atonality is not the only marker of human effort, its intentional inclusion is a clear sign of human intent. Ornette Coleman’s free jazz from 1959 could be argued to be, if not the only reasonable response to the threat of nuclear annihilation, then at least a true creation of humans stuck with living in that context.

A machine could be trained to produce atonal music, abstract paintings or even conceptual art, but such output would lack the existential weight of human choice. The decision to make something difficult, discordant, or challenging emerges from lived experience and conscious rebellion against easy consumption.

It is not that machines could not replicate the beautiful flaws of art, but that the act of replication lacks either the human struggle or the unique perspective resulting in mere simulacra.

We already knew this, though. Clement Greenberg distinguished avant-garde from kitsch by the former’s resistance to easy consumption and, later, when artists transformed kitsch into avant-garde, even there we see the human spark.

More contemporary critical frameworks, such as relational aesthetics, emphasise art’s role in creating human encounters and social bonds. The gallery becomes a space for people to meet through shared contemplation of human expression – something that loses its meaning when the expression itself is algorithmic rather than experiential.

It seems self-evident that the creation of art is proper to humans. Consequently, we already have the answer to how we should deal with AI art: we should resist it. The lawsuits and lobbying are underway, but what is crucial now is to remember that art is not wallpaper.

I am hardly the first to complain that the tech industry seems determined to use AI to automate cultural production and leave the rest of us with more time on our hands to do life’s drudgery. However, there is something more fundamental at stake: AI-generated ‘art’ is useful, useful only to capital and to fill time and space. And that is why it is a mistake.

Read More: AI Artificial Intelligence Blog Blogs Jason Walsh