For Drake and the Weeknd, two of the most popular musicians on the planet, the existence of “Heart on My Sleeve,” a track that claimed to use A.I. versions of their voices to create a passable mimicry, may have qualified as a minor nuisance — a short-lived novelty that was easily stamped out by their powerful record company.
But for others in the industry, the song — which became a viral curio on social media, racking up millions of plays across TikTok, Spotify, YouTube and more before it was removed this week — represented something more serious: a harbinger of the headaches that can occur when a new technology crosses over into the mainstream consciousness of creators and consumers before the necessary rules are in place.
“Heart on My Sleeve” was the latest and loudest example of a gray-area genre that has exploded in recent months: homemade tracks that use generative artificial intelligence technology, in part or in full, to conjure familiar sounds that can be passed off as authentic, or at least close enough. It earned instant comparisons to earlier technologies that disrupted the music industry, including the dawn of the synthesizer, the sampler and the file-sharing service Napster.
Yet while A.I. Rihanna singing a Beyoncé song or A.I. Kanye West doing “Hey There Delilah” may seem like a harmless lark, the successful (if brief) arrival of “Heart on My Sleeve” on official streaming services, complete with shrewd online marketing from its anonymous creator, intensified alarms that were already ringing in the music business, where corporations have grown concerned about A.I. models learning from, and then diluting, their copyrighted material.
Universal Music Group, the largest of the major labels and home to both Drake and the Weeknd, had already flagged such content to its streaming partners this month, citing intellectual property concerns. But in a statement this week, the company spoke to the broader stakes, asking “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.”
Artists and their labels are confident, at least for the time being, that the social and emotional component of fandom will separate the work of the real Drake from a fake one, even if an A.I. version can nod at his emotional preoccupations and musical tics.
But whether superstars could have their pockets picked, or become altogether obsolete in favor of machines that can imitate them, is only one side of the equation. Royalty-free music generators can be used now to compose a rap beat, a commercial jingle or a film score, cutting into an already fragile economy for working musicians.
And as generative A.I. booms and rapidly improves across text, images, sound and video, experts say the technology could reshape creative industries at all levels, with fans, artists and the systems that govern them having to adjust to new norms on the fly.
“It is now possible to produce infinite media in the style or likeness of someone else, soon with little effort, so we all have to come to terms with what that means,” the musician Holly Herndon, who has studied and used A.I. in her work for years, wrote in an email.
“The question is, as a society, do we care what Drake really feels or is it enough to just hear a superficially intelligent rendering?” she asked. “For some people that will not be enough. However, when you consider that most people listening to Spotify are doing so just to have something pleasant to listen to, it complicates things.”
The breakthrough success of “Heart on My Sleeve,” uploaded by a user called ghostwriter, has helped bring music to the forefront of a conversation that has intensified lately around other mediums, especially since the release of Open AI’s ChatGPT language model and image generators like DALL-E. Commenting under the track on YouTube, ghostwriter promised, “This is just the beginning.”
Courts and lawmakers are only beginning to sort out questions of ownership when it comes to A.I., and copyrights in music can be complicated as it is. For now, protected intellectual property can only be created by humans, but what about when musicians collaborate with the machines?
Martin Clancy, a musician and the chair of a global committee that seeks to explore the ethics of A.I. in the arts, said the music industry was more organized than some other fields grappling with the rise of A.I.
“What’s at stake are things we take for granted: listening to music made by humans, people doing that as a livelihood and it being recognized as a special skill,” he said.
It was unclear exactly which elements of “Heart on My Sleeve” — the lyrics, the instrumental beat, the melody, the vocals — were created by A.I. (Ghostwriter declined to comment.)
Some songs have been written by real people and recorded with real human vocals, before being replaced by A.I. imitations of brand-name artists using tools that had “learned” from existing music and produced a similar effect. Those could invite one form of legal challenge: Artists and photographers, for instance, have sued image generators for creating derivative versions of their work.
But a human creator passing off her own song as being performed by a famous artist, or promoting it commercially using that singer’s name or likeness, could lead to a different kind of legal threat. In the past, musicians including Tom Waits and Bette Midler have successfully argued in court that they had a right to not just their musical compositions or recordings, but their voices, in the face of sound-alike imitators in advertisements.
In this case, getting “Heart on My Sleeve” removed from services where it could have earned streaming royalties — and even charted on Billboard — may have been even simpler for Drake, the Weeknd and Universal Music. The track appeared to use a popular vocal snippet from the rapper Future that implied the song was produced by Metro Boomin, a sample of a master recording that was not cleared for use.
Drake, the Weeknd and Metro Boomin declined to comment. (Last week, in response to another track that used an A.I. Drake voice to perform Ice Spice’s “Munch,” Drake wrote cheekily on Instagram, “This is the final straw AI.”)
Aside from raising questions of legality, such technology can introduce knotty ethical concerns regarding race and identity. Last year, Capitol Records apologized and dropped the digital rap avatar FN Meka after critics said the project amounted to a form of blackface. Among the recent explosion of A.I. imitations, rap has emerged as the most common playground.
“It’s another way for people who are not Black to put on the costume of a Black person — to put their hands up Kanye or Drake and make him a puppet — and that is alarming to me,” said Lauren Chanel, a writer on tech and culture. “This is just another example in a long line of people underestimating what it takes to create the type of art that, historically, Black people make.”
But for musicians like Herndon, who has provided her own A.I. voice as a tool for other musicians — complete with a system for compensation — and created a company, Spawning, to build consent guidelines for A.I., there can be magic in harnessing the future fairly and ethically.
“There is more opportunity in exploring this technology than trying to shut it down,” she said.
While meme art like “Heart on My Sleeve” may quickly become “a real cultural force,” she added, “the novelty will eventually be exhausted.” What will remain are the artistic possibilities “when anyone can assume the identity of someone else, even just for a second, as an expressive tool.”
As the technology continues to advance and is adopted in novel ways, someone may eventually do for A.I. voice models — part of what Herndon calls “identity play” — what producers like Prince Paul and J Dilla did for sampling.
“As an artist I am interested in what it means for someone to be me, with my permission, and maybe even be better at being me in different ways,” Herndon said. “The creative possibilities there are fascinating and will change art forever. We just have to figure out the terms and tech.”