Dustin Ballard fell in love with playing the violin as a child in his school orchestra, and making music has remained his lifelong passion, only now he does it using artificial intelligence.
When the coronavirus pandemic hit and he could not play in bands, Ballard started making music parodies, swapping out the audio of famous songs with his own for comedic effect. He then started his channel, “There, I Ruined It,” where he “lovingly destroys your favorite songs.”
A few years later, Ballard discovered he could streamline his creative process by using AI to mimic the voices of artists, allowing him to make his parodies even more realistic.
“This is weird,” said Grammy Award-winning artist Ed Sheeran in response to Ballard putting his song “Perfect” into the style of the Super Mario Bros video game theme song.
Ballard now has 1.31 million subscribers on YouTube, and he’s not the only one who has recognized the applications that AI technology can have in making music.
In September, Xania Monet became the first AI artist to be signed to a label, and even made it onto the Billboard Adult R&B Airplay chart. According to Billboard, in early November at least one AI artist had “debuted in each of the past six chart weeks, a streak suggesting this trend is quickly accelerating.”
The songwriter behind Monet, Telisha Jones, said in an interview with CBS Mornings that she uses AI as a tool and views it as an instrument. Her success has spurred debate about whether AI can be used ethically, especially because her name and sound bear a resemblance to human singer Victoria Monét.
According to Vanity Fair, the artist can’t prove that her work was used to train the AI model, but she feels uneasy about its success.
Xania Monet has raised concerns among other aspiring artists and Grammy-winners as to how this revolutionary technology could change the landscape of the music industry and impact artists’ livelihoods.
“It really pisses me off. Actually, I feel like we should be pretty pissed off. I feel disgusted by it,” said Sarah Negahdari, who has been a professional musician for 20 years.
She asked if this is the kind of creative world people want to live in: “a bunch of vanilla, soulless, same not moving forward, slop, puke, AI, crap.”
“Go for it,” she said. “It’s not the world I want to live in.”
Doug Petty, a professor of popular music and music technology at University of Southern California, as well as a composer and producer for award-winning artists, said ignoring AI is not an option. In his advanced music production course, one of the projects is to create a song from scratch using AI.
He said the goal of the project is to force students to confront AI, as many are concerned about it threatening their future careers and are hesitant to use it. According to Petty, this perspective is common among musicians who are not interested in putting a prompt into an AI model to do their job for them.
For Negahdari, the guitarist and singer of the bands Happy Hollows and Pisces, and former touring bassist for Silversun Pickups, the path she chose has not always been easy. Still, she has remained dedicated to her art.
“Musicians, we sacrifice a lot for our passion,” she said. “[W]e do it because the joy of the journey is everything for us.”
Negahdari spent years developing her craft as a songwriter and multi-instrumentalist. She said most artists are not born great, but their passion drives them to continue making music.
“[It] is the most fulfilling thing as a human being you can experience. And so when somebody, some tech bro or AI bro, says, ‘Oh, they just want to push buttons,’ it’s like, you are missing the whole point of what makes us an artist and what makes us, essentially, human,” she said.
Though Ballard has found online success using AI, he has a complicated relationship to it as that very technology is threatening his day job in advertising. AI music has begun to be used in commercials, troubling musicians who depend on that steady source of income.
“The priority for our clients … is to save money. It’s not a creative endeavor. They’re just trying to sell a product or a service,” Ballard said in an interview. “So it does bother me that you can cut out these music houses, these talented musicians that are just trying to create better work.”
While artists are concerned they could lose their jobs to AI, it is being trained on their work, raising questions about the legal implications of its use in terms of copyright infringement.
“As human beings, why would we allow this to happen? I hope that we can rein this in and be really careful. AI doesn’t create anything new. It can’t. Only humans can,” Negahdari said.
Judith Finell, an adjunct professor of musicology at UCLA and an expert witness on musical copyright, said there are several pending lawsuits involving AI models being trained on copyrighted materials.
“Both the recording industry and the music publishing industry have this strong advocacy when they believe that copyrights have been disrespected and ignored,” Finell said.
At a congressional hearing on AI and intellectual property, one of Ballard’s videos depicting Johnny Cash singing “Barbie Girl,” was played. Then-Rep. Matt Gaetz said that such deepfakes should not be protected under the First Amendment as they are created by robots. In his Ted Talk, Ballard said Gaetz’s statement illustrates legislators’ confusion when it comes to regulating AI, as Ballard said that he, not a robot, created that track.
Lawmakers have proposed the TRAIN Act which would require transparency as to whether AI models were trained on copyrighted works, and the NO FAKES Act, which seeks to prohibit musicians’ voices from being used in AI tracks without their consent. Negahdari is a vocal advocate on social media of regulating the use of AI.
“It shouldn’t be allowed to make these people billions of dollars by stealing from us,” Negahdari said.
Despite the complicated legal issues that AI raises, Finell said a potential positive outcome of AI music is that it allows anyone to be able to play the role of musician, eliminating educational, financial or physical barriers to entry.
“It’s a great opportunity, but that means that someone else can come along and lift it and use it without permission, so it’s both. It eliminates the protection in a way, but extends the potential of the music,” Finell said.
What separates AI from human-made music is the emotion behind it, and the ability to create a relatable experience for listeners, according to Negahdari.
Petty said this ability to communicate is an advantage over AI because human artists are able to create connections. He said the music business is one of building relationships with other musicians, as well as fans.
“However you may feel about Taylor Swift, there’s something about her, to a lot of people, that people are drawn to her, and they want to go and see her live, and they want to understand her. I don’t think you’re going to have that from an AI-generated artist,” said Petty.
He said a lot of musicians seem pessimistic about the increased use of AI and its possible ramifications, but that he himself remains optimistic.
“There’s a lot of doomsdayers about AI and [people who say] that this is the end of music, and this is the end of everything,” said Petty. “But we have been making music since humans have existed, and we’ve been watching humans perform, and there’s been a drive to create this music. I just don’t see that changing. It’s a basic part of the human condition.”
Going forward, if AI music continues to be embraced by the industry, Petty said he doesn’t think it will lead to widespread job loss. Instead, musicians will have to learn to adapt and use the new technology to their advantage.