Following the Recording Industry Association of America (RIAA)’s recent lawsuit against music production startups Udio and Suno, Suno admitted in court filings Thursday that it had indeed used copyrighted songs to train its AI models, but argued that doing so was legal under fair use doctrine.
The RIAA filed suit against Udio and Suno on June 24, claiming the companies used copyrighted music to train their models. Suno investors have previously hinted that the startup didn’t get permission from music labels to use copyrighted material, but that’s not directly mentioned in the lawsuit filed today.
“It is no secret that the tens of millions of recordings on which Suno’s models were trained will include recordings in which the plaintiffs own the rights,” the filing states.
“We train our models on mid- to high-quality music found on the open internet… Much of the open internet actually contains copyrighted material, some of it owned by the major record labels,” Suno CEO and co-founder Mikey Shulman said in a blog post published the same day as the legal filing.
Schulman also argued that training AI models on data from the “open internet” is no different than “a kid listening to that genre and writing his own rock song.”
“Learning is not an intrusion. It has never been an intrusion, and it is now,” Schulman added.
The RIAA countered with the following: “This is a significant concession to facts they have tried to hide for months but are now forced to admit due to this lawsuit. Their industrial-scale infringement does not constitute 'fair use.' It is not fair to steal an artist's life's work, extract its essence, and then repackage it to compete directly with the original… Their vision of 'the future of music' is one in which fans will no longer be able to enjoy the music of their favorite artists because those artists will no longer be able to make a living.”
Fair use issues are never straightforward, but even established doctrines may not apply when it comes to training AI models. The outcome of this case, which is still in its early stages, is likely to establish an influential precedent that could define the future of more than just the two startups mentioned in it.
