Week 1 Thesis Proposal
Week 2 Sketch & Prototyping
Week 3 Play with TouchDesigner
Week 4 Send + Receive Data
Week 5 Production Process
Week 6 Demo Day
Week 7 Midterm Feedback
Week 8 Reflection for the Demo Day
Week 9 Documentation
Week 10 Production Plan
Week 11 Production Progress
Week 12 Thesis Archive & Progress

Play with TouchDesigner

---

Last week, I shared my first prototype with the class and got some useful feedback. It made me rethink about the sequences of the experience, and whether showing the research findings before the real-time data visualization would influence the answer and make it biased.

Taking the class feedback into consideration, I also found myself more curious about the audience’s first thought of AI. Therefore, I reorganized the whole experience. Instead of showing the research insights first, I will start right away with the question, “If AI can be anything in this world, how would AI look like? (Type a word or a phrase)”

Then it comes to the process of making visualization. As I will be using TouchDesigner, there are some steps I need to figure out:

• How to receive real-time data from the audience
• How should the interactive interface look like?
• How should the data visualization look like?

I found helpful tutorials that show how to send and receive messages from a website to control TouchDesigner, using WebSockets, so I decided to experiment with that route. In terms of visual, I wanted to explore how AI image generator would portray the imaginaries of itself. However, I don’t want to use tools that are not transparent about the data they collected in order to train their models. Although that issue is not yet resolved and there are many artists who are not happy with the process of AI image generator, I found that Stable Diffusion took a little more steps toward being transparent about their training data. Luckily, I also found a tutorial that shows how to include real-time diffusion in TouchDesigner. The only limitation is that this requires a really good GPU (CUDA 11.8 or CUDA 12.1), which I belive I should be able to accessed at IDM Grad Lab.

For this week’s class, the main discussion will involve how my work could contribute to what’s already out there and how it is unique. For my thesis, there are two main themes including “AI imaginaries” and “immersive experience.” I think my work will contribute to the discussion on AI perceptions in a more creative and humanistic view. I hope it could leave some impacts on how the current AI technology is biased toward some groups and is seen in a homogenous way.

In terms of the experience, my work would be fitting in with other works that employ similar techniques to visualize sociotechnological issues through immersive art. I found my work builds off of many similar works like Queering the Map (which I mentioned in my thesis proposal as one of the key inspirations) and other immersive exhibits I had experienced such as Refik Anadol’s Machine Hallucination and the Mystic Universe (Wisdome LA). My goal is to create a space where the audience can immerse themselves in the world of AI imaginaries — to make them watch, concentrate, and reflect on their own and others’ perceptions of AI.