Like so many, I have been researching AI, the many platforms it is presenting on and its many varied uses. There is no doubt its power and effectiveness is jaw dropping and this will have deep and forever changing ramifications across all industries. For me as a business owner it presents a very compelling opportunity to automate marketing and generate assets that before would have taken skilled labor and at a much greater cost. On the other hand, as is true with all technological evolutions, the music industry is yet again being forever changed. Given all the buzz around AI music creation and my recent interview with generative AI music company founder Yihao Chen of GenreX, I have made it my mission to dive deep into music AI to understand how it is impacting music producers, composers and what that will mean for those seeking to pursue careers in music.
Today I posted a track I’ve been working on for the past few months on Soundcloud. Given that I have a very busy schedule, finding time to work on original music can be challenging, but with renewed interest and determination I have pushed through to release new music under my artist handle Code369. Recently, I posted the original track Ghost Eyes to my Soundcloud account.
As an experiment I decided to use the generative music platform so many of my peers have been telling me about called Suno, to test out its music production prowess and do a first hand comparison between it and the original track I created. I was originally intending to produce several “remixes” of the track as is customary in the electronic music world and engage other artists to potentially remix my track in various sub genres but for this test case I decided to use AI. To be clear the original version of Ghost Eyes is 100% my doing and utilizes no AI, ghost writers or other producers in its creation. Yes, I used some samples I acquired off of Splice, the online sample library used by so many electronic music producers today but other than that the content is 100% original. Lyrics of spoken word poetry where written and performed by Joseph S. Plum with the chorus vocals and refrains written and performed by me.
Suno AI music generator is a platform that provides the function of generating songs based on text descriptions. It says that the perfect prompt is all about simplicity and not to over-explain or use unnecessary words. Based on their user agreement you can use Suno AI for free but any content generated using a free account can not be monetized. You still own any of the lyrics you enter, assuming you wrote them to begin with, but only paid accounts give you the license to monetize the audio and associated art files created. Copyrighting is of course a whole other issue here which brings monetization into close scrutiny. If you read the fine print of Suno’s user agreement they say the following: “The availability and scope of copyright protection for content generated (in whole or in part) using artificial intelligence is a complex and dynamic area of law, which is rapidly evolving and varies among countries. We encourage you to consult a qualified attorney to advise you about the latest development and the degree of copyright protection available for the output you generate using Suno.” So to put it succinctly: You may own the lyrics but copyright in the wild west of generative AI music remains a very significant grey zone.
So armed with this information I purchased the entry level ten dollar Suno monthly subscription plan and began prompting the AI with the style of the track I had produced and the very same lyrics used in Ghost Eyes. Based on my interview and the early tests I performed with Yihao Chen on the GenreX platform, I knew that organic and more exotic ethnic instrumentation (of which I am a big fan as evidenced by the the assorted instruments I use in my version of Ghost Eyes which include Hang Pan, Didgeridoo, Kanoun, Ney & Djembe) would be more challenging for the AI. None the less I moved forward with the following simplified prompt: Organic house meets melodic techno at 123BPM with a fusion of ethnic instruments and dance beats.
I then added the lyrics and began generating the tracks. At first I was surprised by the etherial quality of the female vocal but after multiple prompts to continue the track with all the lyrics what I ended up with was a very generic sounding EDM progressive house track. Not that this is necessarily a bad thing but if it weren’t for what I believe to be the interesting lyrical content the track itself was pretty generic in its quality.
This was proof enough that human creativity has yet to be replaced by AI but as a tool for inspiration and the ability to quickly generate music for sync and other commercial uses it is definitely a game changer. Most of all the speed that Suno was able to create at was indeed impressive.
What this ultimately brought into focus for me though was the true nature of what the creative process represents. I believe that as a society we are so motivated by profit and the short cuts AI represents that we forget that like life itself, creating art is about the journey.
AI is a powerful learning mechanism that will devour every artits style and then regurgitate it based on the prompts one uses to extract the desired result. The ability to mimic a popular artists sound and likeness is where this is already getting very complex. Where will the line be drawn? Is this a new licensing opportunity for established artists? If so how will it be policed and monitored?
Launching my DAW and diving into the all consuming process of building a piece of music from nothing can be a long and winding road but that road is the essence of the creative process. I’m not saying it has to be a gut wrenching experience but by removing the process of creation and handing it over to the AI what then does art become but a game of prompts for the purpose of commerce? And if that journey is replaced by an AI prompt are we still making art or has it become consumed by commerce?
As enticing as the short cut that AI represents is, I still believe that the creative process of toiling over a piece of art is ultimately what makes the art great and most important, original. Without originality we will be forever imerssed in the same sounding music that has already overwhelmed the streaming services with close to 100,000 tracks per day being uploaded. Now with AI’s power and speed to create finished tracks what will that number skyrocket to?
As a studio owner and an educator who operates a music production school it is my responsibility to understand and integrate these new modalities where they matter most.
The impact of AI on music producers and music production schools is multifaceted, with both potential benefits and challenges.
