The Punggol Digital District Mural

One of the projects Jeffrey discussed involved fostering co-creation and community engagement through the Punggol Digital District mural. Because the district is such a diverse area, they wanted to create something that was not purely aesthetic, but something that invited the community to participate in shaping the mural. Their goal was to use participatory design methods to bridge the physical and digital, while highlighting artists who may not typically be in the spotlight, including those with disabilities.

The mural development was approached in multiple rounds. In the first round, participants were given the same set of instructions, but were encouraged to interpret them in their own way, resulting in a wide variety of outcomes. In the second round, caregivers were invited to contribute, allowing them to become co-artists alongside the original creators. The third round involved a co-creation workshop where participants assembled different compositions using Ezra’s artwork, which became a foundation for larger collaborative pieces.

They later developed a prototype called Pelim, where people could create their own avatars which were then injected into the digital side of the project. These avatars were placed into the mural so people could later find their characters within the artwork.

Jeffery also mentioned that they worked with Tell Your Children Singapore while strengthening bonds between the artists and caregivers. The project became extremely process-focused, successfully bridging physical murals and digital participation.

Students were given pathways such as curators, discoverers, or storytellers, encouraging them to explore content in different ways. Inspired by “choose your own adventure” and Dungeons and Dragons, the team even created a custom GPT as a proof of concept, embedding historical data and artefact references into a conversational system. The main challenge was striking a balance between factual accuracy and historical imagination, which became a central point of discussion. They are still conducting focus groups to determine how students engage with these ideas

← This really inspired me!

While I know I can train GPT models myself, seeing it being used in this application was really cool to me. While there is a lot of opinions whether AI usage is good or bad, this project to me shows how we can utilise it in a way that benefits us with the right training of these models. I also really like the mode of application in education by gamifying learning to reach younger audiences.

Work Check!

Table Setup

For my table setup, I was initially worried about how to display my work since most of it is digital. One of my main concerns was my TouchDesigner pieces, as they sometimes break or fail unexpectedly. If anyone wanted a demonstration, I would need to be physically present to supervise, which is not ideal. Even if I were there, it wouldn't be easy to fix issues on the spot if anything went wrong.

Because of this, I decided that the safest and clearest way to showcase my TouchDesigner project was to create a video demonstration. I also included the small writing task probes I conducted earlier in the semester as another point of reference. Although I wished I had more substantial physical items to present, I felt it would not be a good idea to print unfinished work at this stage. Perhaps this documentation can grow into something more meaningful later on.

Andreas advised me to document my work in a video, which reassured me that I am heading in the right direction. He also shared a useful shortcut: Option + Command + D that hides the hotbar during recording, making the video cleaner. He then brought up a common problem designers face which is our technical skills sometimes lag behind our vision. To tackle this, he said it’s important to write out my ideas clearly, including the concept, scope, and target audience, so that even without perfect execution, others can still understand the direction.

He also suggested creating a moodboard to explore what each chatbot personality could look and feel like. Another suggestion was to build a paper prototype by placing different responses on a laptop screen to visualise the interactions. Once the ideas are clearer, I can work towards creating low-fidelity prototypes in Illustrator, which can eventually develop into a functional app later on.

Andreas encouraged me to find case studies for chatbot personalities from films, books, and existing work. He also mentioned that I could even ask AI how it envisions certain personalities, which might give me unexpected insights. Finally, he suggested that I further develop Experiment 1 as a way of gathering more insight and data before committing to more complex prototypes.