© Angela Weiss/AFP/Getty Images
A mural of the Notorious B.I.G. by Sipros Naberezny in New York City.
Dante Deiana’s all-time favorite artist is the Notorious B.I.G., the legendary rapper who was killed at 24 in a 1997 drive-by shooting. Deiana, a 39-year-old disc jockey and business owner in Chicago, has always wondered what Christopher Wallace — also known as Biggie Smalls — could have done if his career hadn’t ended so suddenly.
In late April, Deiana got his first taste — and he was blown away.
An artificial-intelligence-generated track featuring Biggie Smalls’s voice rapping “N.Y. State of Mind” by Nas landed on social media, its quality garnering rave reviews from fans. But it also sparked a debate: Did it represent a violation of the rapper’s legacy?
“Just hearing him on something new is like uncovering something from the vault,” Deiana said in an interview. But thinking deeper “brought it back down to earth and made me realize this isn’t real. … AI made this.”
The track is among a wave of AI-generated music to reach listeners in recent weeks, as a boom in generative AI technology has led to applications that can produce movies, fine art, novels and other works, upending the tech industry and captivating the public. Music has been no exception, and the release of widely consumed AI-created tracks featuring music’s biggest names has raised questions about the music industry’s future in a world steeped in AI tools.
The use of a dead artist’s voice adds another layer of legal and ethical uncertainty, and fans and artists are weighing a desire for more music from departed artists while also facing the discomfort of bringing their voices back from the dead. Some fans have aired concerns about creating artwork using these artists’ voices without their consent — and the possibility someone other than that artist or their family could profit from it. They’ve also questioned the music’s authenticity and whether it could be considered art at all.
Their voices are their livelihood. Now AI could take it away.
The Nas cover isn’t the only example of AI replicating Biggie. Influential hip-hop producer Timbaland, who has won four Grammys and has worked with Beyoncé and Justin Timberlake, previewed his own song featuring the voice of Biggie Smalls in a video viewed 1 million times on Twitter, as well as being posted on Instagram.
“I always wanted to work with Big, and I never got the chance to — until today,” he said in a video before playing an excerpt of the song that includes the Notorious B.I.G. impersonation.
Some listeners wrote that they were impressed, commenting with fire emoji and even describing the production as historic.
Others argued that the use of Biggie’s voice sounded inorganic — that certain lyrics seemed strange in the artist’s voice.
Timbaland, whose given name is Timothy Mosley, told The Washington Post that he understood the criticism. He recalled producing Michael Jackson’s posthumous 2014 album, “Xscape,” as at first “a little eerie” but also something that eventually “touched my soul.”
Mosley described a similar sensation when creating his Biggie track, and he noted that his intentions were to show what the new technology could do, and not necessarily release a track for money. He compared listening to the sample to viewing fine art infused with stories and “memory,” and he spoke about how the technology presented an opportunity to connect with the artists that fans miss most. Mosley said he had other similar productions in the works.
“It really was an art statement … to tell the world, ‘This is here — this is not really going anywhere,’” he said.
Listeners’ initial discomfort over new work relying on dead artists can be overcome, Mosley said. He added that he was brainstorming ideas for using performers’ voices respectfully, such as music for a “celebration” on a dead artist’s birthday — or producing a one-time release.
“It gives us memories,” he said. “Sometimes if it’s done right — and if it’s done with class and taste, the song was written right, and it was touching to the soul — I think we would love it.”
Patients were told their voices could disappear. They turned to AI to save them.
Justin Bernardez, 27-year-old music producer with 2 million TikTok followers, said he started using AI-powered music software a couple of months ago. He has used it to mimic living artists including Drake, Bruno Mars and Rihanna, but Bernardez has also created tracks replicating the voices of dead performers such as Michael Jackson, XXXTentacion and the Notorious B.I.G.
Bernardez said he’s wrestled with the ethics of such music production and has closely tracked feedback from his followers.
“It’s very, very tricky,” Bernardez said. “There’s the side of people that are like, ‘This doesn’t feel right to me — let them rest in peace. And then there’s an entire other side of the internet that says, ‘Oh, my gosh, thank you so much. It feels like their voice is going to live on forever, and you’re now allowing me to imagine it in a way that I never could have imagined.’”
Bernardez said he is not making money off content that mimics other artists to avoid legal complications. Right now, he said, he’s focused on creativity and getting a foothold in what he considers the future of music production. He also said he avoids using the voices of artists who have recently died “to take into account processing time.”
“I just try to wait as long as I can and make sure that my audience knows that this is out of respect and creativity and imagination,” he said.
In April, the release of “Heart On My Sleeve,” a generative-AI track in the style of Drake and the Weeknd, sparked debate over the definition of artistic expression and offered a glimpse into consumers’ appetite for music created using AI. Responding to the track, Universal Music Group, which works with Drake and the Weeknd, asked “which side of history all stakeholders in the music ecosystem want to be on: the side of artists, fans and human creative expression, or on the side of deep fakes, fraud and denying artists their due compensation.”
The music label may not like the answer. “Heart On My Sleeve” received 15 million views on TikTok and 600,000 streams on Spotify before it was taken down on those platforms.
After the “N.Y. State of Mind” cover was released, Reddit founder Alexis Ohanian predicted the music industry would devise a system for paying artists and their estates for the use of AI-generated voices.
“There’s gonna be a messy few months here, but record companies can actually not mess up this time by leaning into the tech shift,” he tweeted.
U.S. law remains largely unsettled when it comes to ownership and copyright of AI-created works, legal experts said, although they said it is clear that voices, in and of themselves, are generally not subject to copyright. Rather, the right to a person’s voice most closely falls under a patchwork of state laws aimed at protecting the use of a celebrity’s name, face and other aspects of their “likeness.”
But some experts noted another danger that has intensified in the age of social media and could balloon as AI-generated music becomes more popular.
“Historically, if we look at music, African American artists have rarely received the credit and compensation commensurate with their contributions,” said Olufunmilayo Arewa, a business law professor at Temple University’s Beasley School of Law. She added that the digital era has made it easier to deny Black artists compensation and recognition.
Arewa, who is also an anthropology scholar, noted that one recent incarnation of the issue is the use of music produced by Black artists in TikTok dances by White creators. And if that trend continues as generative AI becomes more popular, Arewa said, it could deny Black artists proper compensation and lead to misrepresentation or other unforeseen problems.
“That’s the world we live in today,” she said. “And I think it’s going to rear its head in all kinds of places.”