Do you want AI to take your music and use it to help generate… new music?
The futures of music creation, copyright law, and the entertainment industry itself are being shaped right now. Because a massive lawsuit is about to decide whether AI companies have the right to train their models on copyright-protected content.
The three major labels (Sony Music, Warner Music Group, and Universal Music Group) have filed a lawsuit against the music-generation services Suno and Udio. The suit claims that the defendants used music from label catalogs to develop and train their AI models.
In response, the two AI companies hired Latham & Watkins. This is the same lawfirm who has represented Anthropic and OpenAI.
What do the labels argue?
Lawyers for the three major labels say that allowing AI companies to train their models on label-owned music would:
”… saturate the market with machine-generated content that will directly compete with, cheapen and ultimately drown out the genuine sound recordings on which [their services were] built.”
Are they alone in this fear? No. And that’s partly why a bipartisan group of US senators introduced the COPIED Act, which aims to protect intellectual property and prevent deepfakes.
What are the AI companies arguing?
In response to the lawsuit, the AI companies are not admitting they DID train their models on copyright-protected music. But rather that if they did, it would be lawful under the doctrine of “fair use.”
According to Wikipedia:
Fair use is a doctrine in United States law that permits limited use of copyrighted material without having to first acquire permission from the copyright holder. Fair use is one of the limitations to copyright intended to balance the interests of copyright holders with the public interest in the wider distribution and use of creative works by allowing as a defense to copyright infringement claims certain limited uses that might otherwise be considered infringement.
I’m no legal scholar, so take this with a grain of salt. But I often hear “fair use” cited when copyright-protected content is used in news coverage or documentaries, educational material, or satires and parodies.
In the case of AI models, some very smart people argue that the tech is just doing what humans do: Listening, learning, recognizing patterns, and transforming that knowledge into something measurably different from the source material.
Of course that argument leaves out… the humanity. And to paraphrase the image below:
“We wanted AI to wash the dishes so we can make music, not the other way around.”
What happens next?
This case will take a while to get settled. And there’s a good chance it’ll work its way to the Supreme Court.
If the AI companies win, rights-holders may have to figure out how to monetize the humanity of their music in a future of endless machine-generated choices.
If the labels win, the AI companies could probably never afford the damages, so they’d get taken over by the major labels.
And at that point, the labels can authorize themselves to have fully-licensed, fully-legal AI-generated spinoffs of ABBA songs and club hits by deepfake Drake.
