Record companies sue AI music generators Suno, Udio

San Mateo, CA, USA - Feb 8, 2020: Office building at the Sony Interactive Entertainment Headquarters campus in San Mateo, California, United States.

SAN FRANCISCO - The biggest players in the music recording industry sued two fast-growing artificial intelligence music start-ups Monday, alleging that they used copyrighted songs to train their tools, adding to the pile of lawsuits the AI industry is already facing.

A group of record companies, including Sony Music Entertainment, Universal Music Group and Warner Records, brought two suits, one against Suno and the other against Uncharted Labs, the developer of Udio. Both companies let people generate songs with simple text prompts.

Subscribe to The Post Most newsletter for the most important and interesting stories from The Washington Post.

“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,” said Mitch Glazier, CEO of the Recording Industry Association of America, the industry group that Sony, UMG and Warner are all members of.

Generative AI tools like chatbots, image-generators and song-generators are built by ingesting huge amounts of human-created content. The record companies allege that Suno and Udio used songs they didn’t have the rights to when they trained their AI algorithms. Spokespeople for Udio and Suno did not return requests for comment.

As interest in AI exploded over the past year, authors, artists, graphic designers, musicians and journalists have begun pushing back against the AI industry’s use of their work to train its tech. Lawsuits have been filed against AI companies such as OpenAI by authors, comedians and newspapers.

AI leaders generally say the use of books, news articles and art to train AI falls under “fair use,” a concept in copyright law that allows the re-use of copyrighted content if it is substantially changed. But many creators disagree, saying that their work is being stolen to train tools that could be used to replace them.

Suno and Udio allow users to generate full songs by typing in a description that might include the desired genre, lyrics and the kinds of instruments being used. Suno blocks requests asking it to generate a song mimicking a specific artist. Asking it to create a song “in the style of Dolly Parton” leads to an error message saying it’s not possible to generate something with a prompt that mentions an artist’s name, according to tests done by The Washington Post. Udio doesn’t appear to have the same restriction, readily generating a mournful country song with lyrics sung by a voice that sounds similar to Parton’s when given the same prompt.

Some musicians have asked for new laws specifically protecting their likeness or the style of their music. In Tennessee, home to the Nashville music industry, legislators updated an older law earlier this year to specifically ban mimicking a musician’s voice without their permission. A bipartisan group of federal senators proposed a similar national law last year.

Related Content

How the migrant crisis tested schools 2,000 miles from the southern border

The pope’s right-hand man is reshaping the church, becoming a target

GOP pick for N.C. governor downplayed Weinstein allegations, assault by Ray Rice