Melodies in transition: What influence does Generative AI have on Music Creators?

1. Please introduce yourself and your community briefly.

I’m Jesse Josefsson, a TV and Film music producer and host of Sync My Music, the largest online sync licensing-based YouTube channel and educational platform. I got my start creating music for TV shows, movies, and commercials in 2008 and have been sharing my knowledge and insights into this side of the entertainment industry for the past seven years.

2. How have the recent advancements in generative AI been received by your community?

To many musicians, AI feels like a direct threat to not only their livelihoods but also their very identities. For the longest time I think we all assumed that music was so complex and ‘mysterious’ in its creation that there was no way a computer program could ever perform creative tasks – or if they could, it would be decades or even centuries before that was possible.

Apps like ChatGPT and Midjourney quickly stunned the world in showing just how capable some of these neural networks can be. Then, when Google’s MusicLM and Meta’s MusicGen applications came online it was one of those ‘jaw on the floor’ moments. We could see that these models are getting much better at making generative music at a much faster pace than we all anticipated. This leaves music creators with the undeniable question, “Will these models totally replace the work that I’ve been doing?”

I believe the top ten percent of music composers probably won’t face too much pressure from AI competition, for now. Our community is essentially making it a priority to “AI-proof” our music by creating more human-oriented elements in our tracks such as vocals, live instruments, creative genre-blending, and other techniques.

3. How do you see AI potentially threatening the jobs of musicians in the next few years?

AI has already begun to enter the stock music licensing side of our business. Stock sites such as Pond5 and Shutterstock have already begun to allow generative AI models to train on their catalogues. While this has earned those submitting composers a few extra bucks, it begs the question if these generative models will eventually become the direct competitors of these stock music sites. My guess is that they will; and without the cost of paying out the human creators, they stand to have a competitive advantage in a marketplace that’s really aimed at providing fast, cheap, and easy music to license. In the next two to five years, I can see some of these models becoming useful in the TV and Film marketplace, but it’s unclear as to how disruptive that will be.

4. Are you concerned that AI could eventually replace human musicians altogether?

I’m personally not concerned that AI would replace all human musicians. Creating music is a fundamental and ancient social function of human beings. So, it seems highly unlikely that some new software is going to destroy all of that. Yet it’s certainly going to change a lot of our current assumptions about creating music. For example, these new AI tools likely allow anyone to be creative musically with very minimal effort. So, the very term ‘musician’ might not stand as it once did. But I believe we will always want to connect human to human via music. So, while some commercial uses of music will be replaced by AI, meaningful artistic music will always remain a primarily human endeavor.

5. What can creatives already do now to put themselves in a better position for this AI transformation?

I’ve just released a new series of tutorials on our Sync Academy platform that showcases how producers can put themselves in a competitive position in the sync marketplace for the coming months and years. Essentially, I believe producers and composers need to start adding more human-focused elements such as vocals, live instruments, and more creative genre blending. In doing this, I believe they will be able to become the “premium, organic” choice of human-created music for music supervisors and editors to select from. With generative AI music, it’s likely that most background music, what I call ‘transactional’ or ‘functional’ music, can be easily replaced since most audiences don’t really care if that music was created by a human. So, the strategy is now to secure more prominent, premium placements on TV such as themes and jingles with music that contains a lot of high-quality human creativity. Here lies the potential to develop new forms of musical creativity with AI instruments.

Moreover, I have no doubt that AI will birth new sources of income for creatives of all types. One possibility could be that content creators decide to license their works to generative models in exchange for upfront fees and perhaps a percentage share in backend royalties or blanket licenses. In the music sync licensing marketplace, I could see a play where music publishers train a private AI model on their composers’ assets, then license that model out to particular TV and Film clients.  This would give content creators virtually unlimited abilities to create extremely high-quality music cues, while the composers and publishers remain included in the income benefits. But the future remains uncertain, so time will tell how our industry will incorporate AI technology.