The first encounter I had with AI music is listening to an AI Kurt Cobain cover. At first, I found it cool and interesting that a computer could recreate his vocals so well. During that time, I wasn’t thinking of the impact this type of technology could have on the music industry.
A few months later, my father told me about Suno and Udio, two similar platforms which could create a song according to a prompt you inputted. It was quite amazing, because you could create your own lyrics and add parts you wanted, like a solo, by using [solo].
Although it was impressive, the vocals sounded artificial, the chord progressions were very basic and uninteresting, and all the songs sounded somewhat familiar. This is no surprise, as what these companies did is just collect songs from each genre and feed them to a machine learning algorithm. The music industry considered this theft, and both companies are facing lawsuits.
Nonetheless, people nowadays are not replacing real artists with AI music. However, corporations that need basic music for their ads don’t really require unique compositions, making these platforms attractive to them.
The Copy Right Issue
There is this YouTuber that I watch called Rick Beato. He is a musician and music producer, as well as a former jazz professor. Rick has amazing interviews with artists like Sting, Alan Parsons, Rick Rubin, and many more. As an expert on the music industry, working both as a producer and a musician, Rick understands how these companies operate. Rick was even called before the Senate to testify on AI in music and copyright.
In one of his videos, Rick predicted that with the way AI is developing and how record labels are reacting to it, artists will start embracing their AI songs. For example, an artist like Drake would have his real songs, and then there will be an AI version of Drake. I think that’s a fair way to do it, as artists should have copyright over their voices.
New Capabilities for Music Producers
AI is creating new possibilities for music producers. For example, the plugin Genopatch allows users to insert any synth sound they found, and an AI tries to tweak the sound waves to match the original sound. These types of technologies will keep becoming better and better, opening many doors for music producers.
A major song that used AI technology is The Beatles “last” song, Now and Then. They had found a recording of John Lennon singing with piano, but the vocals were too low in quality to use. With AI, they were able to filter out the sound of the piano and get the pure sound of John Lennon, ending an era with an emotional music video featuring all of the Beatles members over the years.
In the end…
I believe that the Internet and AI will have a similar impact on music. Issues of copyright will become more pronounced, but the tools available to people who make music will increase. The Internet has given us a lot of bad songs, but it has also enabled people to create really good ones. AI will do the same—there will be a sea of generic and formulaic music, but it’s worth it for the few that are actually worth hearing.
I personally look forward to seeing the first people that truly harness AI in music in a creative way. While i’m slightly fearfull of AI and arts in general, i do think that AI can be used as just one tool amongst many and thus can be used in creative ways. I remember reading somewhere that when synthesizers and drumcomputers were first introduced there was a lot of fearmongering about it taking musicians jobs. What happened of course is that it opened up an incredible array of creative, new music. Let’s hope that AI has the same effect. Cool article!