Recently, GEMA’s announcement of a proposed AI licensing model has sparked a heated debate in the music industry. The idea behind this model is to compensate songwriters when AI systems use their music for training purposes. On the surface, it seems like a noble pursuit to ensure creators are paid for the value their music generates, but the reality of this proposal is far more complex—and, frankly, ludicrous.
Let’s break it down.
Music Itself is Built on Training
The history of music is one of evolution and influence. Every musician learns, consciously or unconsciously, by absorbing the music of others. Whether you’re a rock guitarist inspired by Chuck Berry or a classical composer borrowing motifs from Bach, music itself is the result of a continuous process of “training” on past works.
So, if we’re going to claim that AI systems need to pay royalties for “training” on existing music, shouldn’t we apply this same logic to human musicians? If your music has been inspired by decades of past work, shouldn’t you be paying royalties to the estates of The Beatles, Beethoven, or even ancient folk musicians?
The entire idea of levying royalties for AI training ignores the fact that all creative work is derivative. Every musician, artist, and writer builds on the foundations laid by those before them. If we apply this logic to AI, it seems only fair that every human creator should also pay royalties for their influences. Of course, this would be an impossible and ridiculous burden—so why apply it to AI?
A Slippery Slope of Royalties
If AI training is going to be taxed, where do we stop? Do we start charging authors who write novels inspired by Shakespeare? Should painters pay royalties to the estates of the artists whose techniques they studied? The slope is a slippery one, and it ultimately leads to stifling creativity across all fields.
This isn’t to say that artists and songwriters don’t deserve to be compensated fairly. They do. However, trying to squeeze royalties out of AI training is not the way to do it. The focus should be on ensuring that musicians are paid for the actual use and distribution of their music—for streams, sales, and live performances—not for the passive act of being “listened to” by an AI system that’s learning patterns in music.
Cash Grab by Royalty Organizations
At its core, this proposal by GEMA and similar organizations feels more like a cash grab than a legitimate effort to protect creators’ rights. Royalty collection agencies are in the business of monetizing intellectual property, and they’ve found a new frontier to exploit with AI. However, the argument they’re pushing doesn’t hold up to scrutiny.
While it’s understandable that these organizations want to advocate for their artists, this isn’t a sustainable or logical way to ensure compensation. The real issue is fair distribution within the streaming economy, where artists often earn only fractions of a cent per stream. But diverting attention to AI training royalties doesn’t address the more pressing issue of modern music economics.
The Real Problem Isn’t AI
At the heart of this debate is a misunderstanding of where AI fits into the creative process. AI isn’t replacing musicians—it’s simply a tool, much like digital audio workstations (DAWs) or synthesizers. Training AI on music is about teaching it to recognize patterns, genres, and trends, much like a human artist learns by studying their influences. The AI doesn’t take the music and monetize it; it learns from it, just as human musicians do.
If we were to charge AI systems for learning from music, we would also need to charge human musicians who learn from the greats of the past. This is clearly impractical. If AI-generated content is sold commercially, that’s where royalty structures should apply. But charging for the learning process itself? That sets a dangerous precedent that could disrupt not only AI development but creative work across the board.
Conclusion: A Flawed Proposal
While the idea of levying royalties on AI training might sound like a way to defend creators, it’s ultimately flawed. Music has always been an evolutionary art form, built on the influence and inspiration of past works. If AI has to pay for learning from existing music, we would have to impose the same royalties on every human artist.
Instead of chasing AI royalties, royalty organizations should focus on ensuring artists are paid fairly for the actual use of their music—whether through streams, sales, or performances. This would provide a sustainable model that benefits musicians in the long run without stifling creativity or technological progress.
GEMA’s proposal might be well-intentioned, but it’s a misguided cash grab that risks opening a Pandora’s box of unnecessary complexity. Rather than taxing learning, we should be supporting creativity in all its forms—whether it’s from a human or a machine.
This push for AI training royalties is ultimately shortsighted, and if it plays out, it could set an unsustainable precedent. Let’s focus on the real issue of ensuring fair compensation for artists where it matters—through the distribution and use of their music, not the inspiration and training behind it.
