The Recording Industry Association of America—along with a coalition of music industry titans—has filed a lawsuit against AI developers Udio and Suno, the group announced on Monday. The lawsuit stems from what the RIAA described as “mass infringement of copyrighted sound recordings.”
In lawsuits filed in the Southern District of New York and the District of Massachusetts, the RIAA alleges that Suno and Udio illegally used recordings to train their respective AI models, including recordings by Mariah Carey, Jason Derulo, The Temptations, and The Jackson 5.
“As a preliminary matter, only Suno and Udio know the full scope of what they’ve illegally copied,” an RIAA spokesperson told Decrypt. “They have both taken steps to hide the scope of their widescale infringement, which we expect to uncover in the litigations. But we know more than just that the snippets ‘sound like’ copyrighted recordings."
Others joining the lawsuit against Suno and Udio include UMG Recordings, Capital Records, Rhino Entertainment, and Warner Music International.
Both Suno and Udio are generative music services, which allow users to create music based on descriptions or uploaded samples. The RIAA lawsuit is similar to several filed by book and news publishers against AI companies for allegedly training their text models on their content.
According to the RIAA spokesperson, simple prompts mentioning an artist and their work produce recordings that “overwhelmingly mimic and replicate specific recordings” by those artists or their voices, style, and sound.
“There is no way basic text prompts could produce such sounds if the models themselves had not copied and ingested those artists’ copyrighted work,” they said.
“These are straightforward cases of copyright infringement involving unlicensed copying of sound recordings on a massive scale,” RIAA Chief Legal Officer Ken Doroshow said in a statement. “Suno and Udio are attempting to hide the full scope of their infringement rather than putting their services on a sound and lawful footing.”
The lawsuits, Doroshow added, are necessary to reinforce the most basic rules for the responsible, ethical, and lawful development of generative AI systems.
In court documents provided by the RIAA, the attorneys said that when Suno was approached about the copyright issue, the AI developer claimed fair use.
In addition to a jury trial, the RIAA is demanding $150,000 per infringement incident, a declaration that Udio and Suno infringed on copyrighted recordings, and an injunction barring the services from infringing copyrighted recordings in the future.
“Unlicensed services like Suno and Udio that claim it’s ‘fair’ to copy an artist’s life’s work and exploit it for their own profit without consent or pay set back the promise of genuinely innovative AI for us all,” RIAA Chairman and CEO Mitch Glazier said in a statement.
Udio and Suno did not immediately respond to a request for comment from Decrypt.
While neither company discloses how their AI models were trained, documentation published by both services prohibit the use of copyrighted material by users.
“We explicitly forbid the use of copyrighted material and any other third-party intellectual property to generate content using Udio,” according to its published published FAQ.
Suno's FAQ also requires users to obtain permission for material uploaded to the service, including lyrics. It also adds a disclaimer regarding the legal status of music generated by the service.
“The availability and scope of copyright protection for content generated (in whole or in part) using artificial intelligence is a complex and dynamic area of law, which is rapidly evolving and varies among countries,” Suno wrote. “We encourage you to consult a qualified attorney to advise you about the latest development and the degree of copyright protection available for the output you generate using Suno.”
While large language model AIs like OpenAI’s ChatGPT and Google Gemini draw the majority of scrutiny, generative music AI use has also come under fire.
In 2023, an AI-generated song featuring Drake and The Weeknd, titled “Heart on My Sleeve,” went viral on social media. “Heart on My Sleeve” was so popular that there was speculation that the AI song could win a Grammy despite neither artist giving consent.
In April, over 200 artists, including Billie Eilish, Pearl Jam, Nicki Minaj, and Peter Frampton, joined the Artist Rights Alliance in signing an open letter to AI developers demanding that they not release music generation tools, that the group said devalued artists’ work.
“We call on all AI developers, technology companies, platforms, and digital music services to pledge that they will not develop or deploy AI music-generation technology, content or tools that undermine or replace the human artistry of songwriters and artists or deny us fair compensation for our work,” the letter said.
In a separate letter in May, Sony Music Group issued a notice to over 700 companies, including OpenAI, Microsoft, and Google, that its vast library of music was off-limits to AI developers, releasing what it called a “Declaration of AI Training Opt Out.”
Lawsuits and threats of lawsuits have not slowed the release of unauthorized AI deepfakes of musicians, however.
In May, another AI-generated song, “BBL Drizzy,” appeared online as a diss track aimed at Drake during his very public feud with fellow rapper Kendrick Lamar. The song’s creator, Metro Boomin, posted “BBL Drizzy” on Twitter and encouraged fans to create content using the song in exchange for a prize.
Edited by Ryan Ozawa.