Folk AI
Commercials during the Super Bowl presented one vision of artificial intelligence (AI). The Genspark ad gave us AI as a tool that could speed up the cubicle warriors’ productivity so they could get back to managing their fantasy football teams and shopping on Etsy more quickly. The redundancy in the cubicle farm in the commercial makes it obvious that in real life a downsizing would soon follow, but the commercial sold the fantasy that a better work life is only a prompt away.
This past January, a Folk Alliance International conference panel saw AI as the problem, not the solution. Folk Alliance is the annual meeting of the folk music industry, and the panelists generally concluded that AI was bad. The conversation was muddled, and it was often hard to hear agency or motivations in the analysis. Who was doing what and why with AI was hard to nail down, but everything it and “they” were doing was wrong. “I don’t see the point,” artist manager Blair Clark added with a hint of exasperation.
Charles Alexander, CEO of ViNIL, tried to reassure Clark and the panel. “We’re at the Pleistocene Era” of AI development, he said, and argued that technology will catch up to AI and counter its more nefarious uses. That’s part of what his company and the streaming platform Deezer do. Many of the weapons already exist to help us deal with AI, he implied, if people only bothered to use them.
The AI conversation is really a discussion of a tool, and many of the questions center on how it’s used. American capitalism in the 21st century doesn’t give us much cause for optimism, and the Super Bowl ads were so rosy in the future they foretold that you couldn’t help but be suspicious. The amorphous anxiety in the Folk Alliance panel suggested that our collective discussion of AI needs more nuance and specificity.
At Folk Alliance, Blair Clark’s objection made sense. In the folk music world where songs are perceived as an extension of the artist, individuality is the coin of the realm, and anything that mutes it diminishes a song’s appeal. Sabrina Short of NOLAvate Black talked about her experiences working with Loyola music students who didn’t use AI as a creative shortcut and only employed it to speed up non-creative processes. The anxiety in the room about AI music was too jacked for such a caveat, and the conversation continued as if it had never been said. The concern that AI fabrications will replace songs by living, breathing artists makes particular sense at Folk Alliance, where the financial margins for folk music have always been rough and have only gotten worse. Every potential hit to the income stream, no matter how small, inflicts financial and emotional damage.
But if fans of folk music genuinely choose AI folk over an actual person’s song, the artists need to check themselves. What’s happening with your music when people, given the choice between generic, pre-fab songs and musicians’ IRL creations, pick the AI song? Is it possible that some musicians offer too many secondhand thoughts, phrases, and melodies to be clearly more interesting? Or that AI-generated song structures and instrumentation catch the listener’s attention better than a rehash of conventional chord progressions?
Is it also possible that we overrate the listener? We like to think audiences recognize, love, and validate authenticity, but the success of contemporary country music says we’re being very kind. We like the idea that there are a lot of people out there ready for our nuanced breakdown of [insert topic here], but the fact that seven-inch singles historically outsold albums suggests that more people have preferred the simple pleasures of pop songs they can sing and dance to over the deep dive that an album cut represents.
We like to think of the song as the place where the artist and listener come together, and it’s a connection listeners choose to make. I wonder if too often that relationship is part of a speed-dating event curated by Spotify, and listeners are letting the platform’s playlists choose their music for them.
The place where we most often see human folk competing with AI folk is on Spotify, where an estimated 25% of the music is AI-generated. If listeners choose to let an untrustworthy platform pick their folk for them, that’s on them, even if human artists suffer for their slack. Spotify began by marketing itself as the world’s largest record collection, but Liz Pelly’s book, Mood Machine: The Rise of Spotify and the Costs of the Perfect Playlist, drills into the company’s transition into Muzak 2.0 and the changes that came with that focus. Analyst Sam Taylor summarized Spotify’s Q4 2025 report, writing, “For Spotify, AI is as big an opportunity as the internet or the smartphone and, from where Spotify sits, AI generated music isn’t necessarily a problem, because the more content there is out there, the more important it is for a platform like Spotify to surface the right content to the right people at the right time.”
Algorithmically curated playlists that match moods, activities, and times of day have become the platform’s moneymakers, and they are often consumed passively. Listeners hearing music through Spotify playlists are more likely to experience it inattentively, which makes it less likely that consumers will notice the difference between human and AI product, much less have a preference.
If there’s a way forward, it may be for artists to strive to be as boldly individual as possible. To write songs that only they could write. To steer away from even casually familiar turns of phrase and observations others have made. To mine their own perspectives hard. The commercial marketplace will almost certainly become more muddled and less financially sustaining, but folk music has always spoken to eclectic communities, maybe micro-communities. The economics of that are no better now than they ever were, but that means there’s also less to lose by going all in on your own vision.
For more Alex Rawls musings, check out myspiltmilk.com and @myspiltmilk on socials.
