In my previous post, I had explored GPT-3 in order to help me create some drum beats, chords and a synth lead. However, with the text only possibility of GPT-3, this resulted in a limited, not very audio appealing music track. In this post, I will not only dive into the creation of EDM tracks, but more on the impact of AI on the culture and presentation.
Let’s first start with discussing the meaning of EDM music culture by taking the techno scene as an example. With the involvement of computers in the late 80’s and early 90’s, the original techno sound had garnered a large underground following, growing in popularity with the emergence of the rave scene. This “rave” scene consists not only of the kind of music, but also the artists, clothing, hair styles, and (industrial) locations. The question arises whether AI generated music or artists will also gain a following and have their own culture.
To assess this, I explored an AI tool called splice, by asking the AI tool to create a “night rave” like techno sound. What I recognized was a typical sound known by a famous house DJ called Boris Brejcha, which is a great example of an artist positioning himself behind the DJ desks with a well recognizable demon-like mask. In my opinion, the AI tool took inspiration from Boris Brejcha, creating a different, but comparable sound:
The Ai generated music: https://youtu.be/rwrlwEpJqk8
Boris Brejcha: https://www.youtube.com/watch?v=TAxXRmwA40o and https://www.youtube.com/watch?v=BNe7OrleTlg
To extrapolate this, let’s imagine that an AI avatar, visible as a hologram, looking like the AI generated image below, is playing behind the DJ decks. He plays songs, sounding like the generated one by the AI-tool splice. Assuming that these songs eventually will increase in quality, does it matter that the DJ is a real person or not? If the AI avatar plays fire tracks that you like, why not follow it and create a culture?
Thanks for your interesting post! Personally, I don’t think AI is ready to create entire “good” songs and sets yet because it’s often quite simplistic. I do think that AI can be used by artists to enhance their ability to create songs and that this is a good development. Also funny to hear that the AI-generated sound is similar to Boris Brejcha’s sound, I hear it as well. Additionally, I think playing an entire set at a party or festival is also harder for an AI. Of course, you can play a pre-recorded set, but in my opinion, the best DJs listen to the reactions from the crowd and play songs according to that.
Finally, I think we should consider the ethical topic of ownership of AI-generated music. In order to create music “by itself”, the AI is trained with music created by artists. These artists don’t get any compensation for it, so is that really fair? I do agree that it doesn’t really matter whether the DJ is a person or an AI, but I do think it is important to make decisions about the ownership of music when it comes to AI-generated music.