This song is nonsense — a grotesque mockery of what it is to be human.
An editorial on human creativity
No AI
No AI
Music.
A field guide to the debate over whether machines can make real music — and what's at stake for art, literature, and the human soul.
The debate
When a machine sings, who is it singing to?
Generative AI can now produce a plausible pop song in seconds. Nearly a third of new tracks uploaded to some streaming platforms are made by machines (The Atlantic). The question is no longer whether AI can make music — it's whether that music means anything, and what its flood will do to the people who have spent lifetimes learning to make it themselves.
This is a debate about creativity — the strange, costly, deeply human act of turning lived experience into sound, image, and word. It is also a debate about labour: AI models are trained, without consent or payment, on the creative work of millions of musicians, writers, and artists, then used to produce content that competes with the very people it learned from.
The pages below gather the strongest arguments from both sides. They are not a verdict. They are an invitation to listen carefully — and to decide what kind of culture you want to live in.
Voices from the debate
What musicians and thinkers are actually saying
Short excerpts — follow the links to read each argument in full. Stances are simplified for orientation, not as a verdict.
Perfect in its cynicism, magnificent in its emptiness.
There's no story, there's no DNA, there's no childhood — it's just a copyist.
AI is so far only able to create variations on genuine human artistic expression — pallid facsimiles.
I don't see how this can be acceptable in a society that built its creative economy on copyright.
I'll be interested in AI when it produces something that makes me cry.
We have to adapt or die. The future of music is a human soul fused with AI's technical depth.
We must protect against the predatory use of AI to steal professional artists' voices and likenesses.
The argument
Two ways to see the future
The debate rarely lands cleanly on one side. Here is the case each camp is making.
What could be lost
- Devalued labour. Models train on human work without consent or pay, then compete with the artists they learned from. A global study projects music-sector income falling nearly a quarter within four years (The Guardian).
- Diluted meaning. Songs, Cave argues, come from suffering and risk — the messy human material AI has never lived through and cannot possess.
- A flood of sameness. When output is statistically predicted from the past, the strange, the broken, and the genuinely new get buried under the average.
- Severed lineage. As Sting puts it, a copyist has no childhood, no story, no DNA. Remove the human and you remove the lineage that gives a song someone to sing it.
What advocates argue
- A new instrument. Treated as a tool, AI can lower the floor for experimentation — letting anyone sketch ideas the way a guitar or a sampler does.
- Human curation remains essential. Eno, a pioneer of generative systems, insists success comes only when people are meticulous about inputs and discerning about outputs.
- Adaptation, not extinction. New technologies have repeatedly reshaped music without ending it; resistance may matter less than learning to wield the tool well.
- Different, not worse. An algorithmic track can be serviceable, even enjoyable. The optimist's case is that meaning is supplied by the listener, not only the maker.
Beyond music
The same question, asked of art and literature
Music is the loudest front, but the argument reaches every creative field. Visual artists and writers face the same trade — their unlicensed work becomes training data for systems that may replace them.
Visual art
Image generators were trained on millions of paintings and photographs scraped from the web. In late 2024, more than 13,000 artists and creators — from Kazuo Ishiguro to Julianne Moore — signed a statement calling the unlicensed use of their work "a major, unjust threat" (Maginative).
Literature
Novelists and poets confront models that mimic voice without possessing a life to draw on. Writers and teachers describe AI prose as derivative — able to replicate genres that already exist, but not to produce anything genuinely original.
Live performance
The stage is the hardest place for a machine to fake. The communion between performer and audience — the risk, the breath, the shared moment — is exactly what AI cannot manufacture, and why it may be where human art retreats to regroup.
The manifesto
Make it human, or don't make it.
This site does not claim that tools are evil, or that every algorithm is theft. It claims something smaller and harder: that creativity is a human act, that it costs something, and that a culture which forgets the cost will forget how to make anything worth keeping.
01
Consent
No model should train on an artist's work without permission and fair payment.
02
Provenance
Listeners deserve to know whether a track was made by a person or generated by a machine.
03
Risk
The best art is dangerous to its maker. Defend the conditions that let artists take the leap.
04
Listening
Hearing is easy; listening is an act of attention. Refuse the flood. Choose what you give your ears to.
Curated reading
The pieces behind this debate
Opinion pieces, reporting, and primary documents. Read the originals — this site only quotes them briefly under fair use.
"This song is nonsense" — Nick Cave on a ChatGPT song in his style
→
A world that feels completely meaningless — Cave on AI song generators
→
Will AI ever write a great song? "It lacks the nerve."
→
Sting warns against AI songs: "The building blocks of music belong to us"
→
Sting: AI music "leaves a lot to be desired" — it's just a copyist
→
Thom Yorke: AI steals from artists and "devalues humanity"
→
An AI music pioneer quits over the "fair use" of copyrighted works
→
200+ artists call for protections against "predatory use of AI"
→
13,000+ artists sign statement opposing unlicensed AI training
→
1,000+ musicians release a silent album to protest AI copyright changes
→
Brian Eno on art, AI, and why human curation remains essential
→
Daniel Bedingfield: "We have to adapt or die" — AI is music's future
→
Musicians are deeply concerned about AI. Why are labels embracing it?
→
AI Is Testing What Society Wants From Music
→
Listening to All the Music AI Is Trained on Would Take Decades
→
"ChatGPT said I did not exist" — how artists and writers are fighting back
→
Music sector workers to lose nearly a quarter of income to AI
→
How Should We Create Things?
→