If you can’t beat ’em, join ’em.
The use of artificial intelligence continues to be a real concern for the songwriting community. In recent years, AI-generated music has become more and more commonplace, and creating it has become more and more accessible. Music created with artificial intelligence is already running rampant on social media platforms like TikTok.
And as AI continues to advance, music made solely by machines is becoming harder to distinguish from songs made by real people. A study shared last year stated that 97% of people can’t tell the difference between music made by real people and music made by artificial intelligence. Considering AI is still somewhat in its infant stages… that could spell trouble for the songwriting community.
Many songwriters are choosing to buck against the trend and condemn AI at every turn. Others, like Kevin Griffin, are embracing the presence of AI. If that name sounds familiar, you probably know him as the lead singer of the band Better Than Ezra, an alternative rock band that had some big hits in the ’90s and 2000s.
Fast forward to today, and Griffin has used his extensive music knowledge – and his concern with artificial intelligence’s role in music – to launch a new platform called “Soundbreak.” What is that, you might be asking? It’s actually an AI music-generating software that Griffin invented and many other songwriters in Nashville are getting behind.
Griffin told WPLN that part of his motivation for creating Soundbreak was the fact that labels have benefited the most from the streaming era of music. So when he noticed that AI was feeding on already released songs to train their AI-generating models, and found out that major labels were already working with AI companies, he decided to do what he could to fight back:
“I can throw up my hands and say, ‘Here’s another thing that’s going to cannibalize my songwriting career.’ Or I can say, ‘I want to do something that advocates for me as a songwriter, and my peers.’”
So he came up with Soundbreak, which basically includes a menu of songwriters, some from Nashville including The Cadillac Three frontman Jaren Johnston, who have opted in to choose from. All of their most well-known songs are plugged into the software, and users can give us much guidance to Soundbreak as they would like. If they prefer certain tones, instruments, or vocal styles, those can be requested through a prompt.
To test out the model, Griffin fed Soundbreak a prompt for a song called “Bright-colored kicks,” based on Johnston’s writing style, which was a song about wearing nice shoes to brighten up a gloomy day. The AI software quickly provided two, southern-rock-leaning songs, and Johnston was very happy with what it produced:
“It’s actually really on brand. If I was in a writing session with somebody and I came out with that, I’d be like, ‘Baby, you gotta hear this!’ And my wife hates everything. And she’d be, like, ‘Hell yeah, that’s great!'”
And it all came to be because Johnston was chosen as a “virtual co-writer.” After the songs are generated, they are uploaded to a place where the actual co-writer behind the virtual co-writer can give a listen, and provide feedback. That allows for Griffin and other virtual co-writers to still have some control in the process.
In Griffin’s eyes, Soundbreak is more about having a say in how artificial intelligence is being used than anything else. In the past, AI could have generated songs based on a songwriter’s music without them even knowing it. With this new software, they are at least included in the process… and though I’m personally still not a fan of it, I don’t blame the songwriters in Nashville for trying to find a way to move forward alongside artificial intelligence.
