AI music startups say copyright violation is just rock-and-roll

Several weeks after being targeted with copyright infringement lawsuits, AI music startups Suno and Udio have now accused the record labels that filed them of attempting to quell competition within the music industry. Both companies admitted to training their music-generating AI models on copyrighted materials in separate legal filings, arguing that doing so is lawful under fair-use doctrine.

The lawsuits against Suno and Udio were raised in June by the Recording Industry Association of America (RIAA), a group representing major record labels like Universal Music Group (UMG), Sony Music Entertainment, and Warner Records. Both cases accuse Suno and Udio of committing “copyright infringement involving unlicensed copying of sound recordings on a massive scale.” The RIAA is seeking damages of up to $150,000 for every work infringed.

Both Udio and Suno’s AI-music generation tools allow users to produce songs by typing in written descriptions. According to the RIAA, some of these tracks contain vocals that sound identical to those by famous artists like Bruce Springsteen, Michael Jackson, and ABBA. In May Suno said that its music generator had been used 12 million times since it was rolled out in December 2023.

In their responses, both Suno and Udio say the lawsuits highlight the music industry’s opposition to competition. “Helping people generate new artistic expression is what copyright law is designed to encourage, not prohibit,” Udio wrote in its filing. “Under longstanding doctrine, what Udio has done — use existing sound recordings as data to mine and analyze for the purpose of identifying patterns in the sounds of various musical styles, all to enable people to make their own new creations — is a quintessential ‘fair use’ under copyright law.”

In a blog post accompanying its own filing, Suno said that major record labels had misconceptions about how its AI music tools work, likening its model training to “a kid learning to write new rock songs by listening religiously to rock music” as opposed to just copying and repeating copyrighted tracks. Suno also admitted to training its model on online music, noting that other AI providers like OpenAI, Google, and Apple also source their training data from the open internet.

“Much of the open internet indeed contains copyrighted materials, and some of it is owned by major record labels,” Suno said in the blog. ”Learning is not infringing. It never has been, and it is not now.”

In a statement to Musically responding to Suno and Udio’s filings, the RIAA said that the companies failed to obtain appropriate consent to use copyrighted works before bringing their tools to market, unlike competing services like YouTube. “There’s nothing fair about stealing an artist’s life’s work, extracting its core value, and repackaging it to compete directly with the originals,” said the RIAA. “Their vision of the ‘future of music’ is apparently one in which fans will no longer enjoy music by their favorite artists because those artists can no longer earn a living.”