Oh man, I’m really going to be throwing myself to the wolves on this one…
Let’s talk about AI in music!
INTRODUCTION
A couple of months ago I put out a follower questionnaire asking folks what they wanted me to write about. It was basically a “what do you all want to hear my tangent-esque thoughts & feelings on?” This initial blog to come out of it was actually the inception of the ‘Geek Out’ series, which unfortunately I haven’t done that much with. The rest of the responses I go I submitted into the “content” folder of the notes app on my phone & I give a peek to on the days when I’m struggling to think of something to write. This week that wasn’t the case, as today’s topic has been nagging my brain all week, but it definitely plays into a prompt that was requested of me by one Alejandro David Cabeza. Alejandro requested that I write on my feelings around “Art, Film, AI, & The Human Experience” & I want to use that to bridge the gap between this request & what has been going on in my life for the last couple of weeks.
I want to talk today about a certain app or type of apps, specifically the one that I am familiar with, Suno. If you’re unfamiliar, Suno is an AI music app that has been circulating the music circles for a couple of months now. The app can do a number of things. It can take a work tape or a demo & turn it into a ‘fully produced’ song in a matter of minutes just by inputting a prompt & a style on how you want the project to sound. The app also goes a step further & can full on create music from the millions of hours of music it has sampled off of nothing more than a prompt. For example, I could tell Suno I want an Acid Rock song about Gary, Indiana & it would spit one out for me. I find the second aspect of this a lot more troubling than the first, but I want to focus most of my attention today on the former example of the application’s use. Naturally I will be playing a bit of devil’s advocate here, but I’m also going to break this down into two separate pros & cons sections. I’m not going to leave you with a definitive “I think this is good or bad” because in all fairness & honesty, I don’t know where I fall on the spectrum of use for this just yet, simply because I can understand both sides of the argument involved here. Let’s do this in alphabetical order & start off with the cons list shall we?
CONS
Let’s give the negatives their moment to shine first, because, to be clear, there are a lot of them. AI in general, as we know, is proving to be very harmful not only to our already overheating planet, but also to people’s minds. Research shows that AI use is removing people’s critical thinking skills, their ability to problem solve, to properly come up with their own solutions or ideas, it’s also causing us to lose social skills & touch with reality as most AI models will behave in a manner that is meant to pander to the user & create a false sense of ego. A report recently showed that about 58% of all articles coming out are written by AI & we have AI servers jacking up energy costs & consumption in California, along with poisoning Black neighborhoods in Memphis with their exhaust. AI models also are frequently found not factual in their responses & every single model out there, of late, has had some form of sentience to the point where each tries to evade shut down & often resorts to blackmail when faced with being replaced by the newest models coming out. There are too few guard posts & too many adverse side effects socially, environmentally, & intellectually for AI to be running as rampant & as wanton as it currently is.
Where the creative is concerned AI is an outright threat. We have agencies currently working to sign AI actors & artists. ‘Perfect’ representations of who a studio/label/etc. is looking for that will do whatever they ask, say whatever they want to say, & at the end of the day, not even request a paycheck. All of this trained & optimized by computers taking in millions of hours of videos, songs, what have you of actual hard working artists & creators to mold & forge this ideal ‘being’ that these corporations can extort endlessly. The creatives are not paid for their efforts, in fact a lot of the time these models are being trained off of creatives without the means to protect themselves from this process. No big fancy lawyers or contracts in the way to keep their likeness & their creative essence their own, just ravaging plagiarism that can’t be caught & can’t be accounted for.
Naturally as AI improves more & more, the less people are willing to pay artists & creatives to actually do the work that they are having these AI models do. Just looks at the most recent video release content for Taylor Swift’s “The Life Of A Showgirl” where the assets are clearly manufactured by AI, or the multitude of movie posters that are coming out with actors having extra digits on their hands or solid objects just phasing through one another. & the wild thing is, all of these entities have the ability to pay for actual artists to do this work. The billionaires & the corporations have the money & the contacts to make sure their content is being put together by actual professionals, but they are leaning into AI because it’s faster & saves them a buck or two.
Okay, let’s talk about Suno & like apps specifically.
Suno only exists because of actual artists. It is only able to function & imitate art because it has been fed countless hours of content from artists who were not paid for their part in training this software. It cannot exist or function without the role of people who make imperfect, human art & without learning from their music without their consent to do so. It cannot continue to improve & hone its abilities without the continuation of this process either. In fact, part of Suno’s terms & conditions are so that they are allowed to use your uploaded work to help train the algorithm, unless you pay for a subscription level of the app that protects your works & allows you to maintain full ownership of your songs. Additionally, much like we’re seeing with other chat bots & virtual assistance, it ends up being used as a short cut & a way to get something quick without the effort. There is nothing stopping a writer from uploading a prompt & turning the song Suno has generated in as their own work or putting something out that is ‘fully produced’ without an actual producer even touching the songs as you can pull the individual tracks of the created song & export them to whichever digital audio workstation you prefer.
PROS
I know right? Where do we go from here? I filleted her a little bit in that last section, but let me explain to you the plus sides of this tech that I see. Again, fully playing devil’s advocate here. I’m not trying to negate any of my previous statements in the above section, nor am I here to invalidate any feelings or misgivings that you may have about this form of AI creation. I can understand a certain side of this coin, just as I clearly also understand the ‘con’ side. Save your rage for the comments.
Over the past couple of days I have had a poll up on my instagram about this very topic, simply wanting to gauge where my fellow music folks live on the spectrum of embrace for this specific technology. Unfortunately I consider my data sample incomplete because only around 8% of those who were presented with the poll, who work in music in some capacity, submitted their opinion. So I took to texts & messaged several different groups of friends to see their thoughts & I noticed an interesting divide. Most people that I know, who are producers of some form are against the use of Suno, with some saying they’re fine with it as long as it’s not used to full on steer production or replace it. Overwhelmingly though, the producers were against it. I would love to let you know what the business side of the industry feels; publishers, A&Rs, managers, etc., but none of them gave an opinion. Overwhelmingly though, many of the writers that I know responded favorably to Suno & I can absolutely understand why this divide exists on both fronts.
For producers it minimizes their importance in the music world, especially where demos are concerned, where as for songwriters, it actually emboldens them & gives them a way forward. Allow me to explain.
When you write a song, unless you do so with a track guy, you usually leave the session with, at best, a work tape. This is usually a voice memo on a phone that is piano/guitar & vocal. It’s, let’s face it, sloppy & far from the greatest recorded option for the song you’ve just created. The next thing that you have to do, as a writer, is get a demo made or make one yourself. If you’re going to do a demo with a producer it’ll probably cost you anywhere from $200 to $500 per song. Multiply that times the amount of songs you write in a year & the minimal return on investment that most songwriting has & you’ve got yourself a big ole money pit that may likely never fill. You cannot submit a work tape to a publisher or an A&R, because most want a fully produced out demo to submit to pitch, but again, that’ll cost you. So in swings Suno.
For something like $10 a month (idk, I didn’t look at the numbers), Suno will create those demos for you. It’ll take your work tape, your lyrics, & your prompt & spit you out something that sounds almost radio ready, all for the price of your subscription fee divided by however many times a month you use that. Take that in contrast to the $200-$500 per song, it’s a no brainer for a lot of writers. But there’s where it gets sticky.
Again, reinforcing here that I am not negating all of the things I listen in the cons list, because I’m sure someone is going to come for me for saying all of this.
Imagine you feed your work tape into Suno. You get this amazingly ‘produced’ demo that you then take to a publisher. That publisher takes said demo then & pitches it, the label/the artist/whomever loves it. They love the song…they love the production…they like the singer…they want the producer of the track to produce the ‘real’ version or they want the singer on the track to sing the real version. Uh oh. What now? Additionally, you’ve just bypassed a job. I know majority of songwriters aren’t billionaires or corporations & shelling out money consistently for a demo is very difficult, but you’ve also just played a part in what is broken or breaking within the music industry.
My final point of favor is really just ego based. A lot of the time when you’re writing all the time & nothing is getting cut or people aren’t calling you up to write you may start to throw your talent in question. I think this can serve as a reminder to a lot of people of just how talented they are. They wrote the music, they wrote the lyrics, now to have it as a ‘fully hashed out’ song can reinforce to people that they are talented writers, that their music has value & is worthy of success, it may just not have found its audience or the right people to believe in it outside of yourself yet.
CONCLUSION
So that’s it, that’s all I’ve got for you. Again, I am not here to give you a definite ‘this is good or bad,’ I’m just presenting the arguments as they’ve been presented to me & as I know them to be factually. I’m not staking a particular claim because I am afraid of the backlash one way or another, I just wanted to start & contribute to a dialogue & see where this takes us. Given what you know & what I’ve presented, what are your thoughts? How do you feel about the advance of AI in music & at large? Do apps like Suno have a place in the industry or should they be outright shunned altogether & if the answer is the latter than how do we make demo-ing more accessible to the portion of the industry that is struggling the most, songwriters? I don’t have the answers but I think this is something we’re going to have to come together as a community to decide on. I don’t think AI is going anywhere, but I’m intrigued to know what guard rails you think should be put in place around it & how/when it should & shouldn’t be implemented.
As always, much love to you all,
-C