When generated music is the right answer
The comparison nobody makes explicitly
The argument about generated music usually gets framed as machines against composers. That is not the real comparison for most projects.
The alternative to generated music, in the overwhelming majority of cases, is library music — pre-recorded tracks licensed for a fee, chosen from a catalogue, used unchanged. A composer writing to picture is a different and much more expensive product that most projects were never going to buy.
So the honest question is: does this project need music that is specific to it, or music that is appropriate to it?
Where generated music genuinely fits
- Background beds for explainer videos, tutorials, corporate material, podcasts. Nobody is listening to the music; it is there so the room is not silent.
- Temp tracks for editing, where the final music will be licensed or composed. This is a real workflow improvement: cutting to a temp track that fits exactly beats cutting to something approximate.
- Loops and stings for interfaces, transitions and short-form content.
- Personal and hobby projects with no budget, where the alternative is silence.
- Rapid iteration on tone before commissioning — generating six candidates to show a client what "warmer" means is far more useful than a conversation about adjectives.
- Accessible music-making for people learning, exploring or with disabilities that make instruments difficult. This is a real and under-discussed benefit and it deserves to be counted.
Where it does not fit
- When the music is the point. A film's score, a song, a piece somebody will listen to for its own sake.
- When it must be memorable. Generated hooks are weak precisely where memorability lives, and a brand theme that nobody remembers has failed at its only job.
- When it must be defensible. High-value, long-lived campaigns where an infringement claim would be expensive. Licensed library music has a clear chain and an indemnity behind it.
- When there is a relationship to protect. If you work with musicians, quietly replacing paid work with generation will be noticed, and the damage is not recoverable by explaining it well.
- In documentary and journalism, where fabricated sound has the same disclosure problem as any other fabricated element.
Saying it out loud
Two conversations worth having plainly.
With clients. "The bed is generated, the licence permits commercial use, here is what it cost." Some clients care and most do not. The ones who care would care far more about discovering it later, and being the person who said so first costs nothing.
With collaborators. If you brought in generated music on a project where a musician expected work, say so before they hear it. This is professional courtesy and it is also self-interest, because the industry is small.
The wider judgement
There is a real cost to the working musicians whose income came from library and commission work, and it is being felt now rather than in some hypothetical future. Pretending otherwise, or framing it purely as democratisation, is dishonest.
There is also a real benefit to the enormous number of people who could never afford any music at all, and for whom this is the difference between a piece of work existing and not existing. Dismissing that as unimportant is dishonest in the other direction.
Both are true and no framing dissolves the tension. What an individual can do is narrow and worth doing anyway: use generation where the music does not need to be specific, pay a person where it does, disclose either way, and do not pretend the second category is empty because the first one is cheap.
That is not a rule anyone can enforce on you. It is a standard you can hold, and holding it is what makes the difference between a practice you can explain and one you have to avoid explaining.
There is a third route that gets forgotten between the two extremes, and it is often the best answer. A great deal of excellent music is already available under Creative Commons licences and in the public domain — community archives, national collections, artists who release deliberately, and every recording old enough to be out of copyright. It costs nothing, the licence is explicit, attribution is usually the only condition, and a real musician made it. For a project with no budget that still wants music somebody cared about, this beats both generation and silence, and it takes twenty minutes to search properly.
The one thing to keep
Generated music competes with library music rather than with composition, so the honest choice depends on whether the music needs to be specific, and saying which is which protects both your work and your relationships.
Before you move on
What is the most useful question when deciding whether generated music suits a project?
Pick the one you would defend. Nobody sees your answer.