This sounds promising for music and human creativity.
UMG and Udio just announced a deal that ends their copyright lawsuit. More importantly, it hints how AI-generated music could power immersive experiences without competing with or polluting the rest of the creative ecosystem. Big caveat: details are few at this point, so this is my interpretation of what I read in a Billboard article.
Udio has pledged to become a “walled garden” for music fans. You can remix tracks, swap vocals, play around with tempo. But nothing leaves the platform. No exports. No slop machine polluting the internet.
Artists and songwriters opt in voluntarily. They control whether their work gets used. They get compensated for both training and outputs. And most critically to my delicate ears, synthetic AI-generated slop won’t flood streaming platforms and cannibalize ear-time and earnings from real music.
What I think this deal proves is that licenses become viable and attractive when built around fair remuneration, access to high-value intellectual property, and joint business/product development between rights holders and AI companies. A lawsuit or two doesn’t hurt either when it comes to bringing people to the table.
If done right it should be the opposite of scraping all the copyrighted music you can get your bots to capture without permission to train models that compete directly with the artists they stole from. That is, however, still the standard practice for AI companies. But I’m increasingly hopeful that this approach is being smashed by rightsholders standing up for their rights, while also getting better at showing tech companies how to build superior products by using their IP the right way.
“The walled music playground” (sounds more fun than a music garden) could work because it treats AI-generated music as its own thing, separate from but hopefully also able to boost regular music consumption and the human artist. Technology companies continue to innovate within clear boundaries. Fans get to play and experiment with music. Artists get protected, paid and acknowledged for the important work they do.
No displacement. No flooding DSPs with generated content. No undermining the people who make the actual music.
This could be a model. Licensed. Controlled. Fair.
We’ll see how it holds up in practice, and how competitors like Suno and open source projects trained on scraped music will react.
But I allow myself to be hopeful for now.
Link: https://www.billboard.com/pro/ai-music-udio-settles-lawsuit-universal-music-group/