My children are starting to outgrow kids’ meals, and the last time we got dinner from McDonald’s, they both asked for Big Macs. Of course, they didn’t actually know what was on McDonald’s signature sandwich, and asked me if I did. Without missing a beat, I sang it to them: Two all-beef patties, special sauce, lettuce, cheese, pickles, onions on a sesame seed bun. A huge win for the power of music in advertising—coupled with the fact that this jingle was everywhere when I was growing up.
Music never stopped having that power in marketing. Just about everyone has a collection of jingles they still remember decades later, the songs that trigger memories of notable scenes from movies and TV shows, the music that helps set a mood. And as AI is disrupting everything around us, it’s also breaking into both marketing and popular music. Most recently, AI-generated videos featuring an aging rocker named Michael Bennett finally getting his time in the spotlight on America’s Got Talent duped fans, writes Forbes senior contributor Leslie Katz. And late last year, the AI-generated band Breaking Rust actually reached the top of the Billboard charts.
And while AI can produce appealing lyrics and melodies, people largely do not want to hear them. A study late last year by Bain & Company found that 62% of people did not want to listen to AI music—only surpassed by the rejection of AI news (65%) and AI books (72%).
I talked to Josh Collum, vice president of music at licensing company Soundstripe, about the future of AI-generated music in marketing—which is getting a late start because of sprawling copyright litigation against two of the larger practitioners, Udio and Suno. Once the lawsuits filed by music labels and artists are resolved, the copyright rules of the road will be set—and marketers will have a host of questions to consider about audience sentiment and preference. An excerpt from our conversation is later in this newsletter.
ON MESSAGE
Why AI Music Isn’t Quite Pitch-Perfect Yet
Music has always been important to marketers, and is a vital part of the short-form videos, streaming shows and podcasts that have a growing role in entertainment today. But as AI is taking over, AI-generated music may be moving in on what humans have created—though likely not until sprawling copyright litigation by major music labels and artists against AI music creation companies Suno and Udio is resolved.
After that, what impact will AI-generated music have on marketing? I talked to Josh Collum, who is vice president of music at music licensing company Soundstripe—and a big fan of human-created music—about what that future looks like. This conversation has been edited for length, clarity and continuity.
What would you tell marketers about the larger topic of AI generated music? What to know about it, what to look at best practices, just what would you tell them from where you are?
The first thing I would say, and it’s what we tell our customers right now as well, is that the ground rules haven’t been set yet. Using generative AI music in any form or fashion exposes you to some risks and [possible] litigation right now. We need to wait 12 to 24 months, until the rules around copyright and who owns what is settled.
Beyond that, I think we can tell when something is human made. We relate to each other in a way that we don’t with gen AI music. If you’re a marketer that’s trying to connect with a consumer, they’re going to feel that as well. Music is the most authentic, emotional way to connect with your customer. You’re taking a big risk if you’re wanting to use music that was created with no human emotion, to try to connect with a human that does have emotion. I think in the future, we’ll have enough studies to show that matters.
Any conversation about AI gets to the topic that AI is getting better and better at everything it does. Is there a point in the future where humans will have to do more to defend their music as higher quality than what comes from AI?
I think they will in certain areas… mood music and background music for reality television… will be the first two battlegrounds. There are artists and composers that have built a career on yoga music, on sound bath music. They’re going to have to differentiate. They’re going to have to really battle and make sure that their music can still be found and heard and bought in the TV space. That’s going to hit a certain group of composers really hard. Their unions are going to have to make the case to networks and studios: Why those composers should continue to be hired for those jobs versus using generative AI.
Beyond that, I think production music in general—any kind of music that competes in socials for creators and influencers—they’re going to have to make the case as well, and differentiate in some way why you should use my music in this TikTok video versus this generative AI piece. There’s cases to be made to a marketer that working with this artist that’s real has more value than this generative AI music that has no face, because there’s a larger partnership to be had.
In general, why would you say human-created music is superior to AI-created?
Collum: I’m just a firm believer in there’s something magical about humans. I also firmly believe that music is magic, especially when it’s paired with picture. You can watch a million different scenes and ads, then remove the picture, and you’re like, ‘I remember the music.’
I think there’s an inspiration point with any song. It’s deeply personal in some form or fashion. It was created in a room with humans, and I think we can detect that. I don’t think that’s going to go away. It might get harder, it might get blended, but when we hear something unique and special, I think we know it.
The biggest thing that makes human-created music superior is that we’re not perfect. So if we miss a note, if we play a wrong chord, if there’s an extra sound in the background that they didn’t remove, those are unexpected things that happen in a recording or a writing session that AI can’t fake, because it inherently is perfect. It tries to be. The short answer is imperfection.