A Machine Is Allowed to Make Music Too

Robert Moog died in 2005, long before anyone had coined the term “prompt engineering,” but the panic he sparked in 1964 differs very little from the panic we see today. When he introduced his synthesizer, established musicians warned that an army of electronics was arriving to murder their craft. No army arrived. What came instead was Kraftwerk, Wendy Carlos playing Bach on oscillators, drum machines driving dance clubs across the globe, and ultimately an entire generation of producers who made platinum records with a laptop and a MIDI controller without ever setting a single note onto sheet music. The guitar survived. The piano survived. They all survived, and music was not impoverished by it, but enriched.

Now history is repeating itself. Within thirty seconds, Suno and Udio generate a fully realized track—complete with vocals, arrangement, and mix—from a simple text prompt. And within those same thirty seconds, the entire music industry is back on the barricades. Musicians, record labels, copyright collection societies—everyone chanting in unison that this isn’t real music, that it lacks soul, that it is an insult to the craft. It is a familiar sound. It is the exact same refrain we heard with the synthesizer, the sampler, Auto-Tune, and the drum machine. Only this time it sounds louder, because there is far more money at stake.

That is where the crux lies. Not in art, but in cash registers.

The “Soul” Argument Doesn’t Hold Up

The claim that AI music “has no soul” is seductive because it sounds so poetic, but it collapses the moment you hold it up against history. A synthesizer has no soul either. A drum machine has no soul. A sample of a broken piano from 1962, cut and pasted into Ableton, has no soul. The soul was never in the instrument. The soul was in the choice—which timbre, which moment, which context, which intention behind it. Anyone who prompts an AI model, rejects an output, regenerates, cuts, mixes, rearranges, and ultimately decides what the world gets to hear is making the exact same choice that a producer once made with a sampler. The tool is newer. The process of selection and intention, from which music has always derived its meaning, remains identical.

Furthermore, we have had this debate before: with photography, which supposedly wasn’t “real art” because a camera did the work instead of a paintbrush; with film, which wasn’t “real theater”; with electronic dance, which wasn’t “real music” because there was no guitarist sweating on stage. Every technological leap in the arts has begun with the same indictment: this is an imitation, this is cheap, this doesn’t count. And every single time, that indictment proved to be little more than a defense mechanism for those who felt threatened by what had just become easier.

Where the Real Fear Lies

This is not meant cynically; it is entirely human. A session musician who spent a decade perfecting a bassline now watches a system generate a bassline in fifteen seconds that is plenty good enough for a commercial, a podcast intro, or a background track on Spotify. That fear is real and deserves serious attention—as a labor market issue. As a question of fair compensation, copyright on training data, and how to rebuild an economic model where creators still get paid. Those are fierce, legitimate battles, and they must be fought through legislation, collective bargaining, and clear rules governing what an AI model may and may not ingest from existing human work.

Yet that is a fundamentally different fight from the one currently being waged. The current resistance wraps an economic interest in an aesthetic judgment, and that is dishonest toward the public. Audiences are told that AI music is intrinsically inferior, while the genuine issue is that there will soon be fewer paid gigs for whoever is currently holding the microphone. The second issue is legitimate. The first argument is a smokescreen.

History Repeats Itself—With a Catch

There is one distinction with Moog that sharpens the parallel rather than dulling it. The synthesizer added a new timbre to the palette of the human playing it; the human remained the creator. AI models, on the other hand, have been trained on decades of existing music—often without consent, often without compensation to the artists from whom they learned. That is not an aesthetic objection; it is a property objection, and it is the single objection in this entire debate that truly holds water. Recognition for AI-generated music should therefore never be unconditional. It must come with obligations: transparency regarding training data, a compensation model for the creators whose work made the training possible, and a clear label whenever a track has been (co-)generated by a model—so that listeners know what they are hearing without prematurely dismissing the value of the work itself.

With those conditions in place, no principled argument remains that AI music cannot, by definition, be art. What remains is purely an economic challenge: how to restructure an industry so that creators are not wiped out by the very material they produced. That is a problem we can resolve through policy, licensing frameworks, and collective agreements—not by writing off an entire category of sound in advance as unworthy.

Moog eventually got his own museum. His synthesizer now sits alongside Steinway grand pianos in conservatories. The day will come when we look back on this debate with the same bewilderment we now feel toward those who once dismissed electronic music as a passing fad that would soon blow over. The question is not whether AI-generated music deserves recognition. The question is whether we shape that recognition fairly—with respect for the creators upon whose legacy the technology was built—or whether we wait until the money wins anyway, forcing recognition through without any conditions at all.