My argument isn't that AI should never be used to make music. My argument is that music platforms shouldn't normalize fully AI-generated music as if it were the same thing as human-created music. Human musicians spend years developing skills, creating original work, and building careers. AI music can reproduce those creative patterns at enormous scale, often using training data created by those same artists without their consent. If platforms flood their catalogs with cheap AI-generated songs, the problem isn't simply that AI exists, it’s that it can make it harder for the humans whose work inspired these systems to actually make a living. Innovation shouldn't require replacing the people whose creativity made the industry valuable in the first place.

The fact that AI music is already widespread doesn't prove that it belongs on equal footing with human-created music. That’s an appeal to inevitability, not an argument for what the rules should be. We’ve allowed plenty of technologies into society and then created regulations when their consequences became clear. And I don't simply ‘hate the spam.’ My concern is the economic incentive AI creates. A human musician has to spend years developing their craft, recording, performing, and creating original work. An AI system can generate enormous amounts of music at almost no marginal cost. If streaming platforms give unlimited AI-generated content the same access to their catalogs and recommendation systems, humans aren't competing on equal terms, they're competing against an effectively infinite supply of cheap content. You also can't reduce the issue to ‘judge the output, not the algorithm.’ The algorithm matters when the output may be generated from the accumulated creative work of thousands of human artists, potentially without their consent or compensation. If the technology depends on existing human creativity but can then flood the market with substitutes for those same humans, that's not merely a curation problem. So I'm not arguing that we can magically erase AI from music. I'm arguing that ‘the ship has sailed’ isn't a reason to stop regulating the ship. If AI-generated music is going to be on streaming platforms, there should at minimum be meaningful disclosure, safeguards against impersonating artists, and rules preventing AI-generated content from overwhelming human creators. Innovation deserves a place in music, but it doesn't automatically deserve the same treatment as the humans whose work built the industry.

I agree that mandatory labeling and compensation should be part of the solution, but you haven't actually answered my argument, you've just changed the subject from whether AI music should have unrestricted access to what regulations could accompany it. Labeling tells me that a song is AI-generated; it doesn't tell me whether the model was trained on copyrighted music without permission, whether an artist's voice or style was replicated, or whether AI-generated content is being mass-produced to overwhelm human musicians. And saying ‘training data becomes a licensing issue’ doesn't kill the consent argument, it proves my point. If artists should have the right to consent to their work being used for AI training, then unrestricted AI music platforms are already problematic until those licensing rules exist and are enforced. You can't use a future regulatory system as proof that the current system is acceptable. Finally, I'm not arguing that AI music should magically disappear. I'm arguing that platforms shouldn't treat unlimited machine-generated content exactly like human-created music while the rules governing consent, training data, royalties, impersonation, and market flooding are still inadequate. If your solution is mandatory labeling, licensing, and royalties, then we're actually closer than you think. The disagreement is whether platforms should be allowed to scale AI music first and figure out the consequences afterward.