On May 12, 2022, Mr. Morale was an unknown artist, with a Spotify profile featuring a single hip-hop demo. When the clock hit midnight on May 13, millions of people opened the streaming platform and clicked play on that demo for the first time.
Granted, most, if not all, of those people were really Kendrick Lamar fans on an eager hunt for the rap superstar’s latest album, Mr. Morale & the Big Steppers. But the fluke revealed something important about the power of Spotify’s search function: An artist doesn’t need real fans to make money; they only need to be streamed for a couple of seconds.
According to Professor of Communication Arts Jeremy Morris, it’s these conditions that have accelerated the creation and promotion of AI-generated media.
Unsurprisingly, the music industry hasn’t dodged the growing footprint of AI, and the structure of streaming platforms like Spotify has increased the pace at which the two have become intertwined. It’s this relationship that Morris is now studying.
Because there is limited space on a user’s homepage for Spotify to promote new music and shows, the search bar is often key to being discovered on the platform. Making mere cents on each stream, creators are desperate to optimize these kinds of tools. One easy way to make listeners search? Use AI technology that’s novel to listeners. Generate millions of songs in seconds. Upload them all.
Morris’ interest in media and the tools that make it developed in the 2000s, at the same time that music was becoming increasingly digitalized. Because of his research, he’s now a key part of the new L&S Center for Humanistic Inquiry into AI and Uncertainty.
Morris considers himself “curiously concerned” with the future of AI and music creation and empathetic to the musicians who fear the technology.
“What does it mean if somebody can just take a sample of a musician’s voice, put it through a model and create a version of them singing songs that they never would have wanted to sing? There are a whole bunch of issues about ownership and consent,” says Morris. “There are also valid concerns about AI-generated music redirecting royalties away from artists who are already getting paid very little from Spotify.”
At the same time, Morris is hesitant to be entirely negative. History says that a rough introduction to the industry doesn’t necessarily mean a technology is doomed.
Morris points back to Spotify. The platform is essentially a product of the legal legacy left by the early file-sharing service Napster. When it was first released back in 1999, Napster was sued for allowing users to share or download artists’ songs without directly compensating them, leading to its shutdown less than two years later. Spotify learned from the Napster crisis by building a similarly convenient streaming service that also pays royalties to everyone who generates a set number of streams. The technology, which many in the music industry once feared as immoral and irredeemable, has now evolved into a product used daily by listeners around the world.
Examples like this are the reason it is so hard to predict how new tools will ultimately integrate into the media landscape, and it’s one reason Morris hasn’t come down hard on either side of the AI debate.
He also quells fears with the point that AI’s influence in the industry isn’t nearly as extensive as many people believe. A fake band like Velvet Sundown, which went viral this summer for its entirely computer-generated discography, makes for a good headline, but those examples are few and far between. In reality, AI is more often used for minor tasks such as tweaks in the production phase of music that might otherwise take a human hours to do.
As for the future, Morris believes that any true AI takeover of music would most likely impact the genres that aren’t necessarily associated with human creativity anyways. While this would still impact people, it wouldn’t be as widespread.
The bottom line is that the importance of human experience remains undeniable. Morris references podcasting as an example of this. An AI chatbot can read news headlines if that’s all a listener is looking for. But there’s a reason millions of people tune into Call Her Daddy, the celebrity podcast hosted by Alex Cooper.
“You want to actually hear somebody work through conversations that take unexpected turns or hear them interview in a particular style. That’s valuable,” he explains.
Similarly, AI can draw from the history of hip-hop to write a rap song, but it won’t have the same specific personal anecdotes and reflection that make a Kendrick Lamar song great.
It’s true that the infancy of the AI music industry leaves many questions yet unanswered, but in the meantime, Morris is confident that there is no reason to mourn the market for humanity.
“AI excels at pattern recognition and replication, which allows it to mimic emotional qualities in music based on vast data sets,” he concedes. “However, the soul, the authentic human experience that drives original artistic expression, still primarily resides with human creators.”