Preface

I would like to preface this assignment to talk a little about how this topic surrounds the scene I am in and to acknowledge that it is really hard to remove current biases I have towards the issue. AI companies have inserted themselves into music in ways that are really harmful to the industry. Currently, streaming services like Spotify create their own AI artists, filled with AI generated music, and they promote them to make money off of them, taking away what little revenue artists got already on the platform and taking away the chances to get on playlists, editorials, or being discovered by the algorithm. Companies such as Suno AI market their product as an AI product that removes the need for skill in making music, and claims that it makes music accessible to people, when it removes the core element of making music that makes it so fun – the creativity! People prompt away and create songs that do not reflect them at all, and many people in the industry are being harmed because of this. Besides the unethical-ness of where the data was sourced, or how the music cannot be copyrighted, it is simply just killing the craft. Kids aren’t putting in the hours on an instrument or studying how to use music programs like when  I was a kid starting out. People used to spend hours watching tutorials and really learning about a lot of concepts and making art that really sounded like them. This just isn’t a thing anymore unfortunately. Its sad to see. For this reason, I am honestly very anti-AI. As an artist, I owe it to myself to fight against these things even if they will replace me. Its different in the case of music where the core “tedious” part of making music is the creative part of it. Sorry if maybe the questions I asked were a bit leading in that sense.

There also might be a lot of terminology that sounds a bit too “music-y” so I hope that I’m able to explain things 

Where Are We Now?

Alright, lets talk about how AI is currently being used in the music industry. There’s a lot of useful things that some AI tools currently do that are beneficial to workflows in music. A big example of this is AI stem splitting. As a basic concept, it splits your song into multiple separate audio files for each instrument. So, for example, if I chug in my favourite Sabrina Carpenter song, it would spit out the Vocals on their own, the drums on their own, the bass on its own, etc. This is quite useful! It allows artists to quickly be able to take specific parts and sample them (sampling means to take a small snippet of audio and mess around with it to make it your own) . There’s also some other tools that do analysis for your songs and suggest changes to them (in terms of how they sound, not composition and such, the timbre/tonal balance of it). These tools are useful. Sometimes they kind of suck, which is okay. I don’t really use anything outside of AI stem splitting really to do remixes for fun. There are other tools that are more specifically generative AI. These are used in order to generate full songs (or parts of them, depending on what you want). Right now some artists in the scene generate entire songs using AI and write vocals over them and sing over the instrumental (really common these days). There’s also a hybrid approach where you can ask it to generate specific sounds for you (e.g. a kick, a snare, some shakers, violins, etc). to fit in the current song you are working on. Currently these kinds of features are available in applications that arent really mainstream, so not a lot of people have used them. We’re seeing this use widely adopted across the industry, especially in Egypt where a lot of the film scene is abusing AI across all levels right now, and its definitely trending up. However, it is important to note that an overwhelming majority of music artists are against this movement, with many popular artists really being against it, especially indie artists. The ones that support it are the ones with money to make in it in a sense. I think as it gets more training data it will start being used more in a lot more things. The moment AI composition stops sounding boring I think it will start being widely adopted versus people composing themselves. Though this concept has honestly already existed (people made composition packs that you can use and insert into any song). I’m not sure what the trajectory will be in the future, but its currently looking more up than down from what I’m seeing.

https://suno.com/ Suno is by far the most popular AI tool, you can check it out and try it out if you’d like.

Asking ChatGPT, the response is about the same. It claims that its meant to be an assistant, not a replacement, but talks about how AI can write lyrics (the fun part) or generate demos (also the fun part). This isn’t really assistant level help, its just doing things for you. It also talked about how people are selling their likeness to be used with AI. Some singers license their voice to be used with AI vocal generators. It also talks about how the industry is coping with the change.

https://chatgpt.com/share/69f24f2f-b0b0-83ea-b157-03395560da5e The link to my chat 🙂

The Interview

I’m a music major, and with this perk comes access to the great Dr. David Rafferty. Dr David is a music technologist, which is a fancy way of saying he likes doing audio and music on a computer, and researching in ways in how music can be combined with technology. I had some interview questions, but a lot of free flowing dialogue followed after which had a lot of interesting ideas came up. I also talked with a lot of my musician friends (Some are adjunct faculty at AUC but also gigging musicians), and these are the general questions I asked:

What are your general opinions about AI in music? Do you think that it is overall a positive or negative thing?

Do you use AI in your work? If so, how?

What parts do you think AI shouldn’t be used, if any?

Do you think people who use AI for music are limiting themselves or actually opening themselves up?

These questions covered the general ideas that I wanted to talk about, and I asked follow up questions where appropriate.
Lets talk about Dr David’s responses first.

Dr David had a generally negative opinion about AI (generally). He loves it in the sense that it simplifies a lot of interactions, and removes a lot of bloat from a lot of things, but he acknowledged that we’re in an echo chamber where people are using AI to talk to another AI to then be read by the other person. He talked about how AI should be used to wash the dishes and clean the clothes, and give him more room to do the fun creative things that he wanted to do, rather than do the creative things for him. As a composer, Dr David obviously finds it fun to compose and such for various different instruments.

He then talked about how he used AI in order to do a lot of the programming for a lot of projects he did. It makes it much easier for the AI to do a lot of the grunt work in terms of coding all of the algorithms that he specifies or doing a lot of the visual effects that he wants, and it makes a lot of his ideas possible. From what I see it falls under that “wash the dishes and clean the clothes” category of use, which I think is good. It still allows him to be creative in what he does. He didn’t mention using tools like Suno or that kind of stuff, but he actually mentioned how I used AI last semester in one of my assignments (I needed a female voice actor and I couldn’t find one, so I recorded a voice over myself and chugged it into an AI to make it female) and how he was surprised that I did that, and how he would have penalized me if it was against the rules or something like that.

Dr David also talked about our current graduation thesis (5 music majors are working on albums right now) and how he’s glad that none of us have used any AI in making it, and that we are trying to be creative on our own. He agrees that using AI to generate music for you outright isn’t really helpful and doesn’t help us musicians really improve our craft. You have to do the hours and think and struggle in order to get better! There’s no way around it. Again we tied it back to the point of it being used as a tool and not necessarily as a replacement, and we talked about how many people are genuinely becoming slaves to it in a sense by how dependent they are on it and in a sense atrophying their skills in the process. People lose critical thinking skills when using too much AI, and people are genuinely getting dumber because of it, and music isn’t the limit for that. He strongly suggests that we shouldn’t use generative AI in our major and that we should instead go through the process ourselves to discover our own sound, our own composition, likeness, and such. Using AI would generate something that doesn’t sound like you.

I also asked Prof. Youssef Moustafa (who plays in a lot of bands and gigs across Cairo but also teaches here) and he has a much more negative opinion about it. He sees AI as possibly the worst thing that has happened to music. He’s more in touch with the Egyptian music industry, so he can see first hand the effects of it and how people are using it to make “slop music” and how its being used in films and TV shows and such. He talked about how it looked like people were using it as a tool for instant gratification and as a way for them to pretend that they’re musicians when they really lack the skill for that earned title. He doesn’t use AI in his work (he just plays an instrument and does production for his own music, so there’s no way to really fill it in there) but he does use the useful non-genAI tools of course. He thinks genAI/Suno should simply never be a part of the music making process and how it genuinely makes people worse at what they do and their skill, and just acts as an easy way out for them. There wasn’t much to talk about after that, we simply talked about how we wish this trend would end soon.

My Thoughts

To be honest, I wasn’t really surprised about the results I got. Most if not all people in music are against AI, with varying degrees. Some definitely acknowledge how you need to “get with the times” in order to stay ahead in the industry. One of my professors, Ahmed Refai, made this argument to me. He talked to me about how film composers are given really short deadlines to compete for tasks with other people and get ahead in the industry, and how many of them have resorted to using AI in order to meet those deadlines and stay competitive. The industry is heading to a dark place where quality isn’t the metric, but speed and productivity is. The moment we kind of reach that is the moment where the art dies, and I kind of hate that. I think AI in general can be really useful for a lot of things, but one of them honestly isn’t music. It just doesn’t belong there. I also think this about things like drawing, creative writing, etc. These are fields where genericism and mediocrity really shouldn’t be celebrated. I’ve come to realize that a lot of the text we read throughout our day is AI generated, and reading some readings in this semester like the Mounzer one (hope I remember the spelling) really made me remember how much I love reading human written text that has careful attention to detail, or how I love seeing artist pouring hours into drawing awesome paintings. People who type a poorly thought out 30 second prompt to get a piece of art out are not artists, and their work is not art, and it seems like my professors agree with this notion. I think in these fields it has surpassed the notion of it being a tool and has become a instant gratification machine. People who don’t like writing use it to churn out 2000 word essays, people who can’t draw churn out realistic drawings using them, and people who cant make music churn out music with them too. The responses were expected, really. Most of us have this same opinion about AI. I’m not sure if I have much more to say on the topic in terms of my response to the interviews, it moreso felt like a conversation where we both agreed with each other the entire time. I use AI. Everyone uses AI. There’s a limit, and I think we’ve definitely gone past it in our society. I feel like there’s mass psychosis in our society at this point, like everybody is zombified. Posts in Rate AUC Professors are all AI generated now, with people debating with each other using ChatGPT in order to find a gotcha. In classes that I take where we have to write a reflection, people use AI to write it. Tools that provide such power with nothing to stop people from abusing them turn into addictions like drugs. Legislation is needed. Control is needed. So many steps that we should have done 2 years ago still haven’t been done, and the next generation of people will turn out much worse. I want to say that I feel like one of the last people clinging onto any form of critical thinking at this point. I’ve drifted off a bit from the music point to more of a societal perspective, but even for music, I don’t want to interact with this industry if this is how it is going to be, and I would rather shift my career to something else than to participate in this vicious cycle that is ruining the thing I love. I’d rather practice it as a hobby in that case.

I’m honestly really sorry if a lot of this feels unorganized. I’ve read through it a few times and I think the topic hits too close to home and I get really emotional about it. I think though it might be more powerful to leave a lot of my unfiltered thoughts in there especially in the reflection sections.