As the start of the year finds me with lots of early-stage ideas for my project but no strong impulse toward any one in particular, I thought it would be helpful to write down a bit about the disciplines and modes I do want to include in my project, if possible, then briefly discuss some half-formed ideas which might eventually grow into something equal to my vision for the overall project.
Interdisciplinarity: ideally, this project is something that challenges me as a writer, a designer, and a computer scientist and helps to prepare me for further research in any or all of those areas.
Multi-faceted approach: I’ve found that, in addition to drawing on multiple disciplines, I have the greatest understanding and engagement with a topic when I can approach it simultaneously in creative, analytical, and applied ways. I want to create something experiential and artistic, critically engage with relevant theories and scholarly practice, and build something which can and will be used or experienced outside of academia. This is the most daunting aspect of the project as I currently picture it, but if I can manage it, I think it will be more than worthwhile.
Potential topics/interests:
- Dig deep into the idea of AI ‘authorship,’ particularly using reader-response and post-structuralist theory to question whether generative systems like GPT-3 might more rightly be considered as participatory readers rather than writers
- Play formally with the idea of the infinite scroll in both physical and digital media—I have some rough ideas for a tangible “reel” created with ink-and-paper printmaking methods that could be digitally augmented to create a layered mixed-reality narrative
- Try to pin down better ways to look at and respond to our so-called ‘digital identities.’ I’m especially interested in object-oriented ontology not as a philosophy but as a framework for rhetorical analysis, allowing us to speak about software the software objects which were made to represent us as things which exist in and of themselves, closely intertwined with us but unable to be fully controlled or comprehended.
- Work with fine-tuning LLMs. I’m particularly interested in training a model on my own writing—every bit of it I can find, whether posted online or stored privately—so that I can then converse with it and see what sort of gaps there are between myself and the version of myself that can be predicted/extrapolated based on the things I choose to share or express in writing.
