Last month, “Walk My Walk” reached No. 1 on Billboard's Country Digital Song Sales chart. The moody stomp-clap song, with lyrics like “Every scar is a story of my survival, I've been through hell, but I'm still alive,” has been played more than 8 million times on Spotify. This song was not performed by a human artist. Breaking Rust, whose Spotify profile looks like a handsome, strong man in a cowboy hat, is an AI project. However, there are real people who claim that Breaking Rust's work is a rip-off. Blanco Brown is an artist who mixes country and rap, and the author of this song claims to have used AI to emulate his style. The author of Breaking Rust did not respond to a message I sent asking about the origin of the sound.
This is the latest example of how AI-generated music, with its opaque origins, can just as easily create a hit as it can create confusion about who actually wrote the song. AI-generated music that looks a lot like your favorites but was created with a few prompts is going viral, spreading far and faster than music labels can always take it down.
Two years ago, when some of the first AI-generated tracks started garnering ratings, music labels entered the fray, threatening and filing lawsuits to prevent AI generators from training and using artists' voices and musical styles. Universal Music Group (UMG) has asked for the removal of a YouTube video featuring Eminem's voice rapping about cats. Spotify removed AI slop songs listened to by bots to generate streaming revenue, and UMG had its viral “Drake” song removed from its streaming platform. The song wasn't written by Drake or The Weeknd, but by Ghostwriter, an anonymous artist who uses AI to create music and only appears publicly when wearing black and white glasses.
Now, labels are shaking their fists and shaking hands with AI music generators.
Warner Music Group announced last month that it had settled a lawsuit against its AI music generator Suno (the plaintiffs tested the service and found that Suno produced music similar to ABBA or Chuck Berry when prompted) and had formed a partnership with the company. WMG CEO Robert Kinkle called it “a victory for the creative community that benefits everyone.” The announcement comes just weeks after UMG, the world's largest record label, settled a copyright infringement lawsuit with AI music generation company Udio, and announced that the two companies will partner to build a new subscription service scheduled to launch next year, powered by Gene AI and licensed music from the label's artists.
For every minute you spend listening to generated AI tracks, you spend one minute less listening to artist tracks.Mark Mulligan, founder and senior music analyst at research firm MIDIA
AI companies are facing a flurry of lawsuits for using copyrighted material to train their models. This battle is being waged in Hollywood, the news industry, and the visual arts world. The music industry is the latest company to decide that it might be better to play nice with AI than to engage in a tense and drawn-out legal battle when rights are in murky territory. “AI is here to stay and is transformative,” said Chris Wares, associate dean of music business at Berklee College of Music. Record labels are “future-proofed,” he says.
The proliferation of AI-generated voices and musical styles is flooding streaming platforms and social media with new songs, some good, many less catchy. There are over 100 million songs on Spotify, Apple Music, and Bandcamp, many of which are rarely played or never played. French streaming platform Deezer announced in April that people were uploading around 20,000 fully AI-generated tracks every day, accounting for almost a fifth of all new content. As more playlists and artists creating gen AI music are uploaded, human artists have to fight for your ears. In July, a group called The Velvet Sundown rapidly racked up 1 million views with two albums, something many indie bands struggle to do on streaming platforms, but the band's photos on social media are AI-generated, and the real person and person behind the project remains unknown. In November, Billboard identified at least six AI or AI-assisted songs that charted on various charts. Amid the deluge, Spotify announced in September that it would update its impersonation policy and remove songs that use someone else's voice in an unauthorized manner.
If someone creates a banger AI cover song or viral mashup, such as reuniting Stevie Nicks and Lindsey Buckingham for Fleetwood Mac's new album or pitting Kendrick and Drake in a generated collaboration, the novelty factor is likely to increase the number of views that cut into that artist's original work. However, the purpose of these partnerships is to create new sources of income for artists. The deal with Warner stipulates that Suno will only allow downloads of generated audio to paid accounts, and requires artists to opt-in to having their names, voices, songs and likenesses reproduced. This could be a tailwind in an industry where streaming has dramatically reduced royalties. Creating music that fans love can lead to passive income for musicians. But it also means that an artist's original work competes not just for your ears, but for streaming revenue with the artist's own derivative works.
“The reason generative AI music can't be artist-first is because we live in a finite attention economy. Every minute spent listening to a generative AI track is one minute less spent listening to an artist's track,” says Mark Mulligan, founder and senior music analyst at research firm MIDiA. “We are definitely in a world where more and more consumers are creating and competing with their entertainment time.”
Part of that creative process may take us back to the old roots of how people interact with music. Music has been shared for hundreds, if not thousands of years. People played with it and used it to tell stories. With the advent of recording technology and radio, music became widely distributed, “creating a moat between artists and fans,” Mulligan says. “We came up with the idea that music is a creative endpoint, and that the audience doesn't help shape the music except when they go see a band live.” But now, AI tools are becoming the ultimate form of fan expression. “We’re expanding the funnel of creativity,” Mulligan says.
That's if artists allow their voices to be used on the platform. Some forward-thinking musicians, like Grimes, have already cloned their own voices and encouraged listeners to experiment. It's less clear whether a typical pop star would agree to use her voice to sing words she's never seen before, or whether the potential new revenue streams of such an experiment would prove worth the risk. On Friday, Brown released a “Trailer Trap” remix of “Walk My Walk,” which was like a return to his own style after seeing it imitated. So far, it has only 2,000 streams. “If anyone is going to sing like me, it should be me,” Brown told The Associated Press last month. From now on, we'll probably need to listen to the music multiple times to find out if it's being performed by the artists we've grown to love.
amanda huber I'm a senior correspondent for Business Insider, covering the technology industry. She writes about the biggest technology companies and trends.
Business Insider's Discourse articles provide perspectives on the most pressing issues of the day, powered by analysis, reporting and expertise.
