Exploring the SPI using AI

Design exercise using DALL·E 2

cover image for Exploring the SPI using AI

Exploration the SPI domains

To explore potential futures that could arise from the SPI, I conducted a design exercise using DALL·E 2 image generation software. By inputting various domains of the index as text prompts, the software generated several thought-provoking images that I will share below:


In my previous post about the Social Prosperity Index, I highlighted the limitations of the approach. The generated images using DALL·E further underscored this, as they all featured similar high-rises, futuristic shapes, and transportation options such as roads, train tubes, or flying vehicles. These images reflect the rigid and limited conceptions of future cities that the algorithm was trained on. To truly move beyond these limits, we need to explore new language and more radical visions for shaping our ways of living.

Pushing beyond the bounds of the SPI

To delve into this further, I conducted a second round of iterations using more abstract language prompts with DALL·E. I also specifically focused on the rigid infrastructure of our cities, such as roads and transportation systems, and asked the AI to reconsider their roles and shapes in the city. This led to a wider variety of results:


When I used more abstract language in my queries to DALL·E, the generated results were less consistent in their type, but they were significantly different from the results produced by the SPI. This experiment highlights that our perception of the future has a significant impact on the creativity of our ideas for future legislation. This could be an intriguing starting point for this project.