When connecting my essay and research work to the practical sound piece, my focus was set on the relationship between the ethical approaches to using AI and the subcultures this it sits in. RAVE enabled the foundations of this work, allowing me to create a piece that accesses a morally correct manner of using AI creatively, directly through the artists that inhabit said subcultural communities.
Originally I wanted to use RAVE to build my own data set which I would then use to compose the piece, however I soon realised this was overly ambitious of me in regards to the time frame of the project. In order to build a densely rich model you need to provide between 2-10 hours of recordings which does require some time. In order to combat this issue I found that the designers of RAVE actually made a version that can be put into a DAW and so this what I used to create the piece. The plug in itself came with 10 pre loaded RAVE models which I began exploring, however I wasn’t fully satisfied with the sounds they were producing. Therefore I decided to download other models that artists and designers had made, (see Process of Making blog for context) which I felt to be much more inspiring.
Through using these models within the DAW, I felt it representable in relation to the ethical concerns of using AI. The sources I used were ethically collected, directly from the artist to the users, their full consent has been given, therefore the ownership complications, that are so common within AI, are taken away.
Environmentally, regardless of the purpose, using AI does negatively impact the world, large scale data centres are what power our ability to use AI. And so using RAVE may be more ethically liable but does still have contradictable aspects to its operation. However, while a model like ChatGPT requires thousands of high power GPUs running for months, a RAVE model is designed to be efficient. It can often be trained on a single high end personal computer in a matter of hours or days.
Alongside the creative use of AI, my essay focuses on the importance of subcultural communities, and their impact on perspective. In order to connect this to my practical piece, I decided to use the sources of subcultures I found during my research.
IRCAM was a massive contributor to this, providing a space where practitioners are able to share and discuss current works in a welcoming, sympathetic environment. I wanted to have this association of subculture in my piece to tie it to the written work, but also to create a practice that builds on the sense of belonging that these communal spaces offer. I feel that from this project I have widened my knowledge and passion to create, realising these specific subcultural spaces are very attractive to me. Although this may not be made apparent to the listener, the connection is there, helping to create continuity to the work as a whole.
I used multiple sources from these communities, free and open to all, as well as reading and learning about the process of building a RAVE model. This information was not only useful for this specific project, but for my own personal practice as I eventually do want to build my own model.
https://forum.ircam.fr/search/?topics=RAVE%20Model%20Challenge%202025
Within this process of making I wanted to create a work that represents my current practice, something I can use to grow and adapt. From my period of research I acquired a lot of new knowledge on the topic of AI, and the practitioners working with it, and begin questioning why I was intrigued into their work. What parts of it attracted and inspired me? I took this inquisitive element and kept it relevant during the making of my piece, relating elements of my work to the factors of others that influenced me. Although this may not be noticeable to others, it has allowed me to delve deeper into the specifics of where I want to take this work, and how I can include components of previous projects with this to create a new, relevant practice for myself.