How The Music Industry Can Embrace A.I.

By now, you’ve likely heard the viral A.I.-generated song from voices that sound like Drake and TheWeeknd. The song isn’t that good, but it’s an improvement from 2020’s TravisBott and other generative music attempts in recent years.

Drake had already his “final straw” days before, after the Ice Spice “Munch” mashup. Universal MusicGroup has asked streaming services to block these songs and protect their music.

This feels like Napster in 1999. New technology is here and the industry’s protocol is to resist. But what if we could fairly compensate Drake, The Weeknd, Ice Spice, and the various content owners for any likeness or derivations of their work?

So far, these A.I.-generated songs have only gone viral because they’re based on popular artists. People are intrigued by the novelty. The response is “Wow, this is crazy it sounds just like The Weekend,” not necessarily, “Wow, let me add this to the rotation from now on.”

That familiarity works in favor of the most popular artists. Music itself is subject to the power law, which means generative music is also subject to the power law. In other words, it’s in the superstar artist and record label’s best interest to enable experimentation—as long as there’s a fair way to compensate the artist and rights holder. Their work would be the most-accessed music for generative songs anyway, so why not lean in?

In Stratechery, Ben Thompsonwrote an article about how YouTube’s approach to music licensing for user-generated content could inform how the industry handles A.I:

“Today, though, the videos can stay on YouTube: any monetization of said videos, though, goes to the record labels. This is very much in line with what I am proposing: in this case, the authenticity is the music itself, which YouTube ascertains and compensates accordingly, while in the future the authenticity will be name, image, and likeness or artists and creators.”

On YouTube, TikTok, and most social platforms today, the user just needs to tag the song associated with the artist and it’s cleared. The rights holders will be compensated. (Whether they’re compensated enough depends on the platform, but that’s another debate that we’re not gonna get into today).

Similarly, a user could tag the artist whose voice or likeness was used to create an A.I. song. Ideally, the record labels and streaming services would have agreements in place to count song’s plays as a “stream” or some other measurable unit that can compensate the underlying owners.

The music industry and rights holders are protective of their music, and rightfully so. Most of them don’t want it used in extreme cases, like pornography or controversial political campaigns. But these A.I. songs aren’t quite on that same level of problematic association. Other interactive media formats like gaming, like Fortnite, have found ways to make their environments engaging for fans to build on while protecting their IP from extreme cases.

The music industry’s innovators are eager to expand the pie for the overall music industry. They want to think beyond zero-sum games like competing for market share and revenue splits. They want music to have its Instagram moment that increases the total addressable market for song creation. And they know that A.I. is an opportunity to do that.

There are issues to work out, like any new technology, but the floodgates are already open. If the protocol continues to be “ban, not embrace,” then it may repeat the same mistakes it made at the end of the CD era.