The backlash over Antonoff’s AI comments points to a deeper anxiety about creativity, authenticity, and whether generated art can ever mean anything.
In late 2024, producer Jack Antonoff — the man behind some of the decade’s most commercially successful records, from Taylor Swift’s Midnights to Lana Del Rey’s Did You Know That There’s a Tunnel Under Ocean Blvd — made headlines for something he said offstage. Describing the wave of developers and entrepreneurs building AI music generation tools, he called them “godless whores.” The phrase spread fast, as provocations do. Supporters of AI-generated music pushed back. The discourse settled, as it always does, into two camps talking past each other: one side citing democratization and creative possibility, the other citing soul and authenticity.
But Antonoff’s word choice was more interesting than it first appeared. He didn’t say inferior or cheap or even lazy. He said godless. And that, I think, is the more precise complaint — not a question of quality, but of origin.
I want to be careful here. The history of technology and art is not a history of degradation. Every generation has its panic about the new medium — photography killed painting, the synthesizer killed real musicians, AutoTune killed the authentic voice — and every generation, some artists eventually find in the new tool something genuinely their own. AI music, as it matures, may well produce works of real artistic weight. I don’t think that’s impossible. What we’re debating, in 2025, is what the technology can do now, with what it is now, in the hands of the use cases that dominate now.
And right now, what AI music systems do is extraordinarily sophisticated pattern completion. They have ingested an almost incomprehensible volume of human musical output, found the shapes inside it, and learned to produce new objects that fit those shapes convincingly. The output can be melodically sound, harmonically plausible, emotionally legible. What it cannot do — yet, and perhaps structurally — is be about anything in the way that matters.
This is where Jung becomes useful, not as a mystical appeal but as a psychological one. He described the creative process as the unconscious activation of an archetypal image — something that rises from below the threshold of deliberate thought, not something assembled according to a plan. He wasn’t romanticizing artists. He was making a more uncomfortable claim: that the works which actually move people tend to originate in a kind of contact the creator doesn’t fully control. He called the quality of that contact numinous — a word borrowed from the theologian Rudolf Otto to describe encounters with forces larger than the ego. Something is felt before it’s understood. The art is the attempt to make it legible.
What this means, practically, is that art in this sense begins with a wound, a revelation, an obsession — something specific to a person inside a specific moment of their life. The formal choices — the chord change, the line break, the silence — are downstream of that. They’re accountable to something real. Not to a dataset. To an experience that demanded form.
An AI system has no experience that demands form. It has demands derived from training objectives and user prompts. The difference is not trivial. It is, I’d argue, the whole thing.
Which brings me to the other word in Antonoff’s insult: whores. Strip away the provocation and what remains is a complaint about transaction — about producing something intimate on demand, for anyone, in volume, without the conditions that make intimacy real. AI music tools are, at their core, demand-responsive. You describe the feeling you want; it generates the sonic object that approximates it. The feeling is specified in advance by the person who wants the product, and the system optimizes toward it. This is the opposite of how the creative encounter works. In the creative encounter, the artist often doesn’t know what they feel until they’ve made the thing. The work is how they find out.
That capacity — to discover something true in the act of making — is not a feature AI music tools currently possess. Whether it’s a feature they could ever possess is a question I’ll leave open. Consciousness and its relationship to creativity is not a problem anyone has cleanly solved.
One more caveat, and it matters. When I say “art” here, I’m not talking about most of what the music industry produces. The mainstream cultural economy has always run on product that is competent, targeted, and emotionally generic. Jingles, trend-chasing pop, background music for content — none of this is what Jung or Antonoff are really pointing at. A lot of what currently passes for music in commercial contexts is already closer to functional design than to art in any serious sense. If AI replaces that output, the cultural loss is minimal.
What I’m describing — what Antonoff, however crudely, is defending — is a specific and irreplaceable form: the work that arrives from somewhere the maker didn’t fully expect, that carries in it the mark of a real person’s real encounter with something they couldn’t otherwise say. That form of creation has always been rare. It will remain rare. The question is whether we’re building a culture that can still recognize it when it appears — or whether the sheer volume of generated content makes the signal impossible to hear.
That’s the actual risk. Not that AI music exists. But that we lose the ear for the difference.
Enjoyed This Story? Subscribe to EntertainLens for more entertainment insights.
