Spotify plans to introduce a tool that will allow users to create AI-generated covers and remixes. My first instinct was to raise my hackles up, too, but it turns out there’s a smidge more nuance to get into.
Namely, the planned feature is pitched as a ‘groundbreaking responsible AI tool.’ How does a massive corporation plan to use AI ‘responsibly’? Well, for a start, the whole venture is being facilitated by a “landmark recorded music and music publishing licensing agreement” between Spotify and Universal Music Group.
That means that users will only be able to AI-ify “their favorite songs from participating artists and songwriters.” Spotify tends to pay artists a pittance, but the platform is pitching the AI tool, which will launch as a paid add-on for Spotify Premium users, as an additional source of revenue for songwriters and artists.
In short, the whole thing is an attempt to offer a generative AI tool that plays on the right side of music copyright law. Spotify co-CEO Alex Norström said, “What we’re building is grounded in consent, credit, and compensation for the artists and songwriters that take part.”
Not that the devil needs any more advocates, but if I was to indulge ol’ Beelzebub, I would say something about remix culture being a pretty central part of all things media—both offline and online. A generative remix tool on one of the biggest music platforms out there feels like a natural progression, and I definitely don’t hate the attempt to avoid lifting songs without permission.
However, after The Velvet Sundown debacle last year, I can’t help but feel a little wary about whether this will actually improve anything for artists or fans in the long term. For one thing, unlike other music platforms such as Deezer, Spotify is still not great about labelling AI content.
Introducing a generative AI tool could be seen as sanctioning AI music more broadly, potentially inviting more unmarked AI music onto the platform. This AI music might have nothing to do with the licensing agreement with Universal Music Group, and may be cobbled together from a plagiarised data set. I hope I’m wrong, but I’m also not hopeful.