Are AI-generated songs a ‘grotesque mockery’ of humanity or simply an opportunity to make a new kind of music? | Jeff Sparrow

Earlier this week, a fan called Mark sent Nick Cave lyrics written “in Nick Cave’s style” by the ChatGPT AI system for reasons that aren’t entirely clear.

Let’s just say that Cave wasn’t satisfied with mimicking algorithms.

“With all the love and respect around the world, this song is a bullshit, a grotesque mockery of being human. Well, I don’t really like it.”

Fair enough: why would he do that?

But Cave’s response on the Red Hand Files blog raises a relevant question for all of us when considering what the AI ​​revolution means for our own lives and careers.

For Cave, ChatGPT could not write “real songs”, only “duplicate, burlesque stuff”. That’s because real songs, he says, come out of “the human struggle of complex, inner creation.”

This is what we humble humans can offer, AI can only imitate, the transcendental journey of an artist forever wrestling with his own faults. That’s where human talent resides, embedded deep within its limits, yet surpassing them. ”

Today, artists have been plagued by the stifling effects of technology since time immemorial.

Back in 1906, composer John Philip Sousa made a very famous rebuttal to a future invention called the gramophone.

“Historically, the whole process of music, from the first days to the present day, has been along the lines of making music the expression of the state of the soul. I propose again to reduce musical expression to a mathematical system of megaphones, wheels, gears, discs and cylinders.”

Similar denunciations can be found for electric guitars, synthesizers, drum machines, Auto-Tune, and nearly every new development in songwriting or recording.

But time and time again, people find ways to use this technology in exciting and creative ways.

Think of the golden age of hip-hop. Consider how producers have deployed sampling to create entirely new kinds of music.

That example, particularly legal constraints on subsequent sampling, shows how the likelihood associated with a particular technique depends on the social and economic context in which it emerges.


After all, most pop songs aren’t the result of individual geniuses, and they haven’t been for a long time. The New York Times, dating back to 1910 You can publish an article titled “How Popular Song Factory Creates Hits”.

“Today, the consumption of songs by the American public is as constant as the consumption of shoes, and demand is equally met by factory output.”

Then, as now, companies in fierce business adopted the method of making the most profit as quickly as possible.

AI doesn’t have to be genius to disrupt pop music and many other fields. It has to be good enough so that it can reverse any loss in quality that it makes by being cheap compared to human labor.

A few years ago, John Seabrook documented how Swedish producers like Denniz Pop, Max Martin and Dr Luke changed contemporary music in his book The Song Machine. To create iconic songs by Taylor Swift, Rihanna, Katy Perry, Beyoncé, and more, production wizards start with simple chord progressions on their laptops and then pass that file through to hundreds of thousands of singers, melody makers, and hook writers. , lyricists and tastemakers, then mix digital takes from multiple contributors into a seamless whole.

David Hajdu of The Nation describes the method as “industrialization of the method by mining vast digital repositories of past records, or by emulating or referencing them through synthesis, and manipulating and mashing them up.” We describe it as post-industrial rather than edgy.”

AI is perfect for this kind of songwriting.

Max Martin famously gave Britney Spears the amazing lyric “Hit me baby one more time” because the non-native English speaker misunderstood teen slang in text messages. It’s because Still, as songwriter Ulf Ekberg explained, “It was an advantage for us that English was not our native language, because we treated it very disrespectfully, and it seemed like we weren’t very close to the melody. Because you can find the words.”

Does anyone really think Martin and his team didn’t make use of ChatGPT?

None of this means that AI itself is an obstacle to music production. The problem lies less with technology than with a social system that directs all innovation to immediate profit, regardless of its impact on art or society.

If an AI-generated song “in the style of Nick Cave” makes money, no matter how poor the results, you get it.

Given the loyalty of his fan base, it probably won’t affect Cave much.

After all, AI doesn’t have to be a genius to put you out of work. It should be adequate and slightly cheaper.

Jeff Sparrow is a columnist for The Guardian Australia

Leave a Reply

Your email address will not be published. Required fields are marked *