Spotify says its AI remix tool is for superfans, but I’m not convinced

AI covers and remixes of songs are already a blight on the internet. Spotify, YouTube, TikTok, and Instagram are awash in flat reggae versions of “Smells Like Teen Spirit,” dinky country renditions of The Weeknd, and monotonous Motown reimaginings of AC/DC. Now, a new tool from Spotify will make them even easier to generate and share. Spotify and Universal Music Group (UMG) signed a licensing deal that will allow users to generate remixes and covers from UMG’s catalog. How exactly it will work, beyond being “powered by generative AI technology,” or how much it will cost, is unclear. They’re positioning this as a premium subscription add-on service for superfans. According to UMG’s CEO Sir Lucian Grainge, it’s supposed to “deepen fan relationships.” There’s no denying that learning to play your favorite song on guitar or dissecting a track to create your own remix can teach you a lot about songcraft and help you appreciate your favorite artist more. But those benefits don’t exist when you just prompt AI for a bluegrass version of Beyoncé’s “Break My Soul.” There’s also a tinge of narcissism at play here. Learning to play or sing a song creates a connection to a work and helps you develop a skill. An AI cover is just about shouting, “Look at what I made.” Or, more accurately, “Look what I asked a machine to make for me.” You can see this at play in the Suno subreddit, where people frequently say they only listen to their own music. People there proudly proclaim that they don’t listen to artists on Spotify or other streaming services anymore, they only listen to what they generate using Suno. Those are the people who will pay for Spotify’s remixing tool. Not Swifties looking to build a deeper connection with Taylor. It will be people who think that, somehow, whatever they generate will be better than what a skilled remixer can create. They will convince themselves that they can somehow improve on the work of an army of the most talented songwriters in the industry, with some clever prompting. But, they’re not actually engaging with the art in any meaningful way, and they’re certainly not creating art themselves. Frankly, the whole thing feels disrespectful to the concept of human creativity and to the artist serving as the source material. And what superfan wants to disrespect their favorite artist? At best, people prompting AI covers are simply having a laugh and churning out genre mashups. Which you could argue is a harmless use of AI, but it’s also not a particularly valuable one. Obviously, I can’t speak to the quality of Spotify’s specific generative AI output, as the tool hasn’t been released yet. But I’ve spent enough time with Suno and other generative AI music tools to tell you that what they spit out is dull and lifeless. Is the idea of a fiddle-driven version of the Dead Kennedys’ “California Über Alles” amusing? Sure. But Suno’s execution somehow sucks the fun out of it. It makes no unexpected choices. It sands down any rough edges. (It also generated cover art featuring a swastika, which is… something.) I’d rather hear a person play and sing a fiddle cover of the song on their own in a bedroom, recorded on an iPhone, than listen to the Suno version ever again. For whatever an amateur recording might lack in production value, at least it would have charm. Now, to be clear, taking a song and covering it in an unexpected style is a time-tested recipe for success. It can be played for laughs, as with The Gourds’ cover of “Gin and Juice.” Or it can reveal unappreciated beauty and depth, like on Travis’ “Baby One More Time,” or the Flaming Lips’ take on Kylie Minogue. But turning “I Wanna Dance with Somebody (Who Loves Me)” into a black metal song demands careful thought about instrumentation, arrangement, and a genuine appreciation for the original. Creator Mac Glocky frequently reimagines songs as if they were created by other artists. Where an AI might be able to make a version of “Mr. Blue Sky” that superficially resembles the Deftones, Mac demonstrates a deep understanding of the source material. He doesn’t just add distortion and scream the lyrics, he makes melodic and arrangement choices that genuinely feel like something Chino Moreno and Stephen Carpenter might do. He transforms the song in a way that is distinctly human. [Media: https://youtu.be/_Rp7duGN7V4] The same is true of remixes. The best of them recontextualize a song, ratcheting up certain traits or recasting them for play in a different venue. The steady dance punk groove of Bloc Party’s “Banquet” becomes a full-throated dancefloor banger, Missy Elliott’s slinky “Get Your Freak On” morphs into a glitchy punk rock rager, and La Roux’s “Bulletproof” goes from ’80s tinged pop to a moody slowburn. But these covers and remixes were made by people who knew their craft and developed an understanding of a song. Whatever creative value there is in fan-made remixes is diminished when the level of engagement is reduced to text prompts.