Lisa Röper
Remember how AI was supposed to hit the more mundane, mindless jobs first?
Recent developments suggest that prediction may be more wrong than anyone had imagined. Unveiled last month, a “Fake Drake” song, as just one example, shows that the creative class – artists, inventors, and innovators of all kinds in business, science, and the arts – are at high risk of replacement by bots, at least in part. Indeed, the music business, because of its fast innovation cycles, may be a particularly good case for forecasting how other innovation-related professions will be affected.
In this article I’ll consider implications of AI-generated music for what I see as the inevitable creative class wars against AI, and what it means for us all.
To set the stage here, let’s take a closer look at recent AI music innovations. In the case of Drake, an AI emulator was used to create the singer-songwriter’s tone, inflection, and style for the song “Heart on My Sleeve,” along with the vocal style of fellow mega-star The Weeknd. In just days the song racked up millions of streams, before being pulled for infringement on at least one other creative work. Beyond its popularity – which may have been chalked up to novelty – many expressed fascination about how well AI pulled off its Drake impression, with some noting they liked the song more than most by the two human versions of the artists.
Within days of the AI Drake/Weeknd debut, the technology was used to drop another piece of music: this time it took “New,” a song written and performed by Sir Paul McCartney 10 years ago, and reimagined it using the sound of McCartney’s voice when he was a Beatle. Fans loved it. Next, the creator layered in the AI-generated voice of John Lennon from his prime, though the McCartney bandmate has been dead for over 40 years now. Sure, it felt a little creepy and dishonest, but again fans loved it.
More broadly, what’s happening in the music business, a creative industry that moves faster than others – like books and film – because each creative production can be done relatively quickly, gives us insight into the battle lines of a future fight among creative-domain stakeholders now that AI is in the mix.
Three parties control the music business, at least for now: artists create content; record companies (aka “labels”) finance and market artists, and own the right to an artist’s music catalog; distributors like Spotify get music to the consumer. AI is new to this triumvirate, and can potentially disrupt everything.
For instance, if consumers like Fake Drake better than Real Drake, does a record company even need the real version? Once tracks are recorded in sufficient numbers, AI can be trained to make Fake Drake music perhaps in endless quantities, which cuts artists’ bargaining power at the knees – labels don’t need the artist once they’ve “bottled” their sound and style from existing tracks.
Distributors could get in on the game, too. What’s to stop them from creating their own AI-based Fake Drake or Fake McCartney, or resurrect Billy Holiday, Frank Sinatra, or Elvis, or mashup artists like Madonna and Little Richard or combine their voices and styles into “Little Madonnard?”
The Hollywood writers’ strike is the first big, public battle of the creative class wars.
—
Brian Uzzi
It’s not a big stretch to say where things go depends on who benefits and who holds the power to decide, as is almost always the case. For music, it could be argued that consumers get what they want most – immediate gratification because new songs can be spun off AI with a click of mouse (including by consumers themselves, as ChatGPT, DALL-E 2, and other generative AI apps have shown us for books and art, respectively), and business executives get what they want – financial returns – because AI generates more profits more quickly than human artists, experiences no creative blocks, and avoids messy publicity issues (ahem, R. Kelly). No surprise who the most likely losers are – the ones who create in the first place.
But music is just one front of AI’s encroachment into creative domains.
Indeed, what’s happening in Tinsel Town right now gives us a window into what might be ahead. The Hollywood writers’ strike is the first big, public battle of the creative class wars – and don’t miss the irony that Hollywood writers have long imagined machines as our overlords.
Recently, the Writers Guild of America stated, “AI can’t write or rewrite literary material; can’t be used as source material; and MBA-covered [contract protected] material can’t be used to train AI.” While that sounds reasonable, it’s not clear every creator shares that sentiment. Jumping back to music, the artist Grimes tweeted that anyone can use her AI-generated “voice” as long as she gets 50 percent of the royalties – “Same deal as I would with any artist i collabed with” (sic).
Those developments in two big creative fields suggest it’s largely about both money and job security. Of course both of those are important, as they’re critical to our survival. But remember that even such essentials fall at the bottom of Maslow’s famous hierarchy of needs; they keep us alive and able to strive, but what exactly we strive for and how much opportunity – and incentive – we have to achieve it presents a much bigger problem to consider as AI wraps its tentacles around creativity.
Continuing with Maslow, he argues that human creativity needs positive emotions to flourish. And positive emotion is fed largely by what we find at the top of the psychologist’s pyramid: status, recognition, and self-actualization. So it’s no surprise creative artists, including musicians, have often traded off money for exposure, and that innovators within corporations put in long hours and fight hard to have their name listed on a patent even if the company keeps all related profit.
So the long game in understanding AI’s impact on jobs and innovation shouldn’t be just about money and power, but the potential death of human creativity, one AI neural network at a time. If consumers just want immediate gratification, and businesspeople want profits, Fake Drake and its ilk are the logical future across creative fields. And if that’s the case, what will be the motivation for the next Mozart, Faulkner, or Curie to step forward? If innovators and artists come to realize that their future exists only as long as it take to copy them, why bother trying at all? Ironically, the faster AI changes things, the faster we will be coming to a creativity halt.
This problem is not trivial. Innovation is what keeps the human race ahead of its own problems. If humans don’t exercise their creativity or fail to be rewarded for it, creativity will be lost – we lose what we don’t use, just like with muscle.
So what do we do about it? Stemming this creativity crisis will require thoughtful leadership. Patent law was created to protect the little guy and incentivize innovation. As a start, the legal profession needs to update our conceptions of intellectual property, along with the rules and restrictions for using someone else’s intellectual property – others’ creative work can’t just be “fair game” for AI, as the Writers’ Strike is arguing.
Business leaders, too, need to think differently. The mantra for decades has been “Listen to the consumer.” But does that make sense if the vast majority of consumers are focused on short-term outcomes? Henry Ford understood this issue well, as his quote suggests: “If I had asked people what they wanted, they would have said faster horses.”
Like Ford, we need to recognize and embrace forward-looking leadership that scrutinizes the large costs and benefits of AI-driven creativity for multiple stakeholders: consumers, creators, and others. Let’s start the dialogue now, before AI does it for us.
*
This article originally appeared in Forbes.
