The Challenge
The Challenge
Generative AI today is biased towards stereotypical depictions of cultures and geographies. This can lead to biased designed outputs when designers utilise AI for concept iteration and inspiration in their creative practice.
Objective and Impact
The project aimed to create a culturally adaptable design process leveraging ethnographic research to craft generative AI output with a focus on culture-centric speculative design. The framework comes accompanied by a Fields Guide and digital platform, allowing designers to seamlessly integrate AI into their physical ideation.
Emerging technologies such as AI have the future capability to create an evolving and interactive feedback loop of codes between communities and design experts.
Generative AI tool exploration
Hands-on exploration of Stable Diffusion (SB) and Runway Gen2 to create “culture-based” potential futures set in Singapore. Results yielded images in which technology overpowered contextual factors turning depictions of future Singapore into “cultural-less” representations.
This is a major flaw if designers are looking to use generative AI as a prototyping tool because the output will seldom be inserted in a real contextual scenario.AI ethnography
We asked ourselves a simple question to test and guide our project. “How will people drink in Singapore in the future?”
We conducted ethnographic observational research, to understand how people drink in Singapore now, and create a criteria by which to evaluate our generative AI output, so as to be able to decide if it could be considered to be culturally robust.Image Reproduction in AI
We utilised SB plugins such as ControlNet and Style Selector to engineer hyper-specficic prompts to recreate accurate images of what drinking practices in Singapore look like now.Future trends research
Researched and collected future trends (global and local to Singapore) to utilise for future depiction of Singapore in AI
We selected Singapore’s move to a Sponge City for flood management, and the development of a Smart Nation for full digital transformation.
Singapore Speculative Futures
Used our learnings about stable diffusion plugins, as well as successfully engineered prompts, to create images for the two selected future trends producing a total of eight speculative design outputs.User validation
Gathered data on the use of generative AI for creativity from a generative AI class at NTU to understand how creatives use inspiration and how AI fits into that process.
Design Outputs
Crafted a framework to walk other creatives through this process.
Created an accompanying Field Guide for the usage of the generative AI in conjunction with ethnography and physical sketching of proposed ideas to seamlessly blend digital and physical creative work.
An accompanying Wikipedia-style, open-source platform for creatives to share their work.
