Spacelab
TECH CREATORS FESTIVALS  MUSIC
GEAR AI SOCIAL MEDIA GUIDES  

CYBER SECURITY

 

  BITCOIN
 
     
     
 
     
 

AI Music Startup SUNO HITS BACK HARD At Record Labels, Saying Training on Copyrighted Music Is LEGAL

 
     
   
     
  By Spacelab  
     
 

Just five weeks after the big three record labels hit AI music startups Suno and Udio with lawsuits, accusing them of disregarding copyright owners' rights, Suno responded this Thursday. 

 

In their legal filing, Suno admitted to using copyrighted music to train their model, but they stood firm, arguing that this falls under the “fair use” doctrine and is perfectly legal.

 

Which is of course the exact type of pushback everyone has been waiting for, a challenge from an AI company to push against the definition of copyright law. Suno just hit the copyright music industry square across the jaw, and a fight is brewing.

 

“What the major record labels really don’t want is competition,” Suno’s attorneys wrote in their answer to the label’s complaint.

 

“Where Suno sees musicians, teachers, and everyday people using a new tool to create original music, the labels see a threat to their market share.”

 

Suno lets users create songs by typing descriptions and can now enhance uploaded sound files with its AI. In a recent legal filing, Suno admitted to training their model on “tens of millions” of recordings, including copyrighted songs from YouTube and other sites.

 

They acknowledged that some of these recordings likely belong to the plaintiffs, a fact previously hinted at but never outright confirmed by the company.

 

I've always thought it was interesting that most musicians learned by training themselves on other people's music, but don't want to let AI models train their models on their own their own music. 


This seems like a sort of duplicity to me, musicians think they should train themselves on other people's music but don't want to allow AI to train from their own music. There’s not a consistent ideal there.

 

 

     
  This seems like a sort of duplicity to me, musicians think they should train themselves on other people's music but don't want to allow AI to train from their own music. There’s not a consistent ideal there.  
     

 


The other thing to watch out for in this case is the vulnerability of the recording labels. The music that gets outputted does not violate copyright or lead to copyright infringement. This is a vulnerability of copyright law as copyright law is written.


So, we're in a whole new age here.

 

The kind of thinking that's governed the history of copyrighted music doesn't really have a good fit for this new modern Challenge on training AI bottles to create its own unique content music.


This will be an interesting case to watch and see how it plays out. The outcome is totally unpredictable. 

 
 
 
     
     
 

 

 
 
Spacelab
A community for music festivals, creators & influencers! A music festival platform, online store and digital magazine.
 
Creative Commons Copyright, 2024. Some Rights Reserved.
Spacelab is licensed under a Creative Commons Attribution-Share Alike 3.0 United States License. MORE >
         
FESTIVALS NEWS STORE CONNECT SPACELAB
USA TECH AMAZON FACEBOOK ABOUT
CANADA CREATORS ETSY INSTAGRAM CONTACT
UK     TWITTER ADVERTISE
AUSTRALIA     RSS PRIVACY
EUROPE       ETHICS
ASIA       FTC DISCLOSURE
2024       SEARCH
2025        
COACHELLA