Hey, there! Log in / Register

Music giants sue Cambridge AI startup over its alleged use of their songs

Several of the nation's largest record labels today sued Suno, a Cambridge firm started by AI gurus whose offering lets users create their own songs - based, the labels charge, on their copyrighted work.

In the suit, filed in US District Court in Boston, the music companies allege the company sucked their offerings into its computers to build a music-based "generative AI" system. In December, Microsoft announced it would integrate Suno's offering into its Copilot laptop and desktop AI-based operating system.

Upon information and belief, and consistent with the basic facts of how generative AI works, the content Suno used to "train" its AI model includes reams of Copyrighted Recordings that Suno reproduced without permission from Plaintiffs. Suno could not have built a model capable of producing audio so similar to the Copyrighted Recordings without the initial act of copying those recordings. This explains why one of Suno's investors has publicly recognized that Suno's service is likely to spawn litigation and that defending lawsuits from music labels is "the risk we had to underwrite when we invested in the company."

The suit continues:

When a user prompts Suno's service with a text input (e.g., make a jazz song about New York), the service generates an audio output by making generalizations about what the audio output should sound like based on the prompt and the corpus of sound recordings on which it was trained. Certain features of the outputs from Suno's model betray that it was trained on particular data—in this case, the Copyrighted Recordings. In particular, Suno's product frequently generates outputs with strong resemblance to the Copyrighted Recordings, a telltale sign that such recordings were included in its training data.

In technical terms, by generating outputs that mimic sound recordings in its training corpus, Suno's model reflects the machine-learning phenomenon known as "overfitting." An AI model is "overfitted" when it is too closely adapted to the data on which it was trained, making it difficult for the model to generalize to new data sets. One symptom of overfitting is a model that replicates portions of its training data. To take a simplified example, if a user inputs the prompt "a jazz song about New York" into an overfitted AI model, the model may output a file that closely resembles one of the jazz tracks on which it trained. As the myriad examples discussed below reflect, Suno's model obviously was trained on the Copyrighted Recordings. This infringement cannot be cured by simply loosening the model's fit or by implementing technical guardrails that make it less likely that outputs will match excerpts of the Copyrighted Recordings. In other words, modifying Suno's offering in a way that better conceals its training data would not alter the fact that Suno infringed the Copyrighted Recordings the moment it copied them to create its model.

The basic point is that Suno's model requires a vast corpus of sound recordings in order to output synthetic music files that are convincing imitations of human music. Suno's corpus includes the body of recorded music that people listen to in their everyday lives. Because of their sheer popularity and exposure, the Copyrighted Recordings had to be included within Suno's training data for Suno's model to be successful at creating the desired human-sounding outputs.

The companies are seeking a formal legal declaration that Suno is bad and should feel bad, should knock off what its doing immediately and pay at least $150,000 per copyrighted song sucked into its system, or damages to be determined by the court or Suno's profits, whichever is more, plus attorneys' fees and interest.

The plaintiffs are: UMG Recordings, Capitol Records, Sony Music Entertainment, Atlantic Recording, Atlantic Records Group, Rhino Entertainment, the All Blacks, Warner Music International Services and Warner Records.

Complete complaint (1.6M PDF).

Neighborhoods: 
Topics: 
Free tagging: 


Ad:


Like the job UHub is doing? Consider a contribution. Thanks!

Comments

You don’t hate the record industry enough

up
22

this whole "AI" grift needs a judicial reckoning sooner rather than later. And I'm glad it's starting.

up
73

Suno's lawyers should easily be able to turn this around to point out that if the entire libraries of music from ALL of these vendors is supposedly causing overfitting of the data to the point that Suno is making music so similar to the Copyrighted Recordings as to be obvious knock-offs, then maybe that says more about the state of their industry than it does about whether Suno actually took anything from them that isn't allowed under copyright law.

up
20

these companies collectively own the rights to 70+ years worth of music across various genres so good joke but not really applicable

up
36

> maybe that says more about the state of their industry

So, an excuse to shut slop pushers down? Sounds good to me.

up
17

It takes a lot for me to side with music companies, but these "AI gurus" managed to pull it off. No one should be able to use artists' work without proper compensation. Now if only the music companies would fairly distribute whatever settlement they get to the actual artists... yeah right.

up
55

I hear a song on the radio. I like the style. I write down the chord progressions and setup my synth and drum machine to produce a similar sound and beat. I then create a new song, where only the style is derived from the original.

That's fair use. The radio station was responsible for compensating the artist (via the label).

AI basically does the same thing. It breaks down the inputs into component parts and predicts what follows what and what sounds good together. It pieces together parts based on what it has learned are well-traveled paths through the parts based on the syntax of the request and then glues them all together within a final output. It's the same process I went through up above in my own brain when it was fair use and I paid nothing for it.

But that's not fair use now? Odd.

The music industry is just mad that their human industry of churning out the same 4-chord schlock for the past 25 years has been usurped by a computer model that can do the same thing in much less time.

Also, I'm glad to share some songs I had Suno generate that don't sound anything like what you'd find in the music libraries of all those plaintiffs.

up
26

Yeah, for all of human existence, people have created songs based on existing songs. It's only in the last few decades that lawyers have found a way to extract huge amounts of money from people who do this.

up
21

By all accounts, he made a ton of money off of companies blatantly ripping off his style and using it to sell cheetos.

Ooh. found the report https://www.mentalfloss.com/article/79648/when-tom-waits-sued-frito-lay-...

Also, sorry but I have no interest in listening to the musical output of AI, just like I have no interest in viewing the execrable AI generated "movies" people have been sharing.

up
32

The music industry is just mad that their human industry of churning out the same 4-chord schlock for the past 25 years has been usurped by a computer model that can do the same thing in much less time.

But this perspective seems to come from a place of disdain for popular music. Your taste is your own, and that’s fair, but you're looking at a very small piece of the picture. For example, many of us make our daily bread by writing music for film and television libraries. That industry stands to be decimated by AI. When there’s no longer a middle class of music creators, all you’re going to get is more of the slop you dislike.

Edited for clarity

up
49

It's with the fact that the schlock peddlers have so much centralized power in your industry that they own you too...AND don't care about you. Don't kid yourself. This suit isn't about defending your copyright. It's about defending their profit machine...which isn't whatever you've written for TV and film.

Even if the AI model isn't overfitted, it's still only fitted to the data it's given. There's still room to innovate and do new and different things as creators. AI won't be doing that for quite a while. We've tried letting learning models "do their own thing" where they basically aren't bounded by training data. The output is the equivalent of 1000 monkeys with 1000 instruments waiting for something of value to come out of it.

Coming up with new and interesting material that widens the boundary of art is still something very human. Maybe that's where the profit margin should be. Maybe that's the problem here....not the AI.

up
18

It's with the fact that the schlock peddlers have so much centralized power in your industry that they own you too...AND don't care about you. Don't kid yourself. This suit isn't about defending your copyright. It's about defending their profit machine...which isn't whatever you've written for TV and film.

Like everyone else, I am loathe to side with record companies but this is a rare instance where their interests and mine align. I am well aware that music libraries and record labels will actively find ways to harness AI for their gain as soon as it is feasible and not a moment later.

Maybe that's the problem here....not the AI.

I understand what you're intimating, but I tend to disagree when people say there are no new ideas in music. Your taste is your own though, of course.

up
25

You hear a Beatles song, borrow a few elements from it, and write your own song. That 's fair use.

You program a computer to analyze every Beatles song ever written, to develop an algorithm that can be used to generate an infinite number of new "Beatles" songs. That's rather different. Is it a violation of copyright law? There's probably no precedent for stretching the law to cover such a case, but a judge may be persuaded that the law does cover it. Stay tuned.

up
22

More specifically, George Harrison.

up
22

You can't always get what you want.

up
10