A thought here's a concrete framework for how AI music platforms could compensate artists without killing the technology
The current situation with AI music platforms like Suno is untenable, and everyone knows it.
Users are generating full tracks powered by training data that includes decades of real artists' work. Labels and artists aren't compensated. Platforms operate in legal grey zones. And the loudest positions on both sides "ban it all" vs. "information wants to be free" aren't going anywhere constructive.
Here's a structured framework that could actually work for both sides.
The core idea: a tiered copyright buyout system
Rather than treating all AI-generated music the same, classify tracks based on originality:
- Copy (blocked): Outputs too derivative of identifiable artists/styles → personal use only, no commercial pathway
- Derivative Work: Inspired by existing music but structurally distinct → user can purchase rights through a buyout fee
- Original Work: Clearly unique and user-directed → full ownership pathway with minimal friction
This mirrors how copyright law already thinks about derivative works it just formalizes it for AI outputs.
The revenue model
For any track that enters commercial use:
- 70% → artists and rights holders (distributed via existing label structures)
- 30% → the user who directed the creation
The buyout fee (to convert a generated track into a licensed asset) is also split between the platform and rights holders.
Why this works for the industry
Artists get ongoing compensation. Not a one-time settlement, not a lump-sum licensing deal actual passive income tied to tracks their influence shaped. This doesn't require dismantling existing label structures; it routes through them.
Labels get a legitimate seat at the table. Instead of litigating platforms into the ground (with uncertain outcomes), this creates a system where the industry is structurally embedded in the AI music economy.
The 70/30 split reflects reality. Typing prompts into Suno isn't the same as mastering an instrument, training your voice, or learning production. That's not an insult to AI users it's just honest. The split acknowledges what each party actually contributed.
The alternative is worse
If this kind of framework doesn't emerge through regulation, industry negotiation, or platform initiative the likely outcomes are:
- Continued litigation that drags on for years with no clear winner
- Platforms operating offshore or in jurisdictions that don't enforce copyright
- A black market for AI music that compensates nobody
A structured system isn't a concession. It's the version of this future where the industry has leverage and income.
Open question for the industry side:
What would a fair licensing framework actually need to look like for labels and artists to accept it? Is 70% enough? Does the classification system make sense? And critically who administers it?
Would genuinely value perspectives from people closer to the legal and licensing side of this.