We asked ourselves, can we use AI for gigamapping?
We conducted a series of experiments to find out if using AI in gigamapping makes any sense.
Work done by Ruben Sevaldson (AI) and Birger Sevaldson (SOD)
The purpose of systems Oriented Design has for 20 years been to help designers to become better at coping with complexity. For that we developed Gigamapping as a flexible qualitative and mouldable practice that served many different purposes in the design process
We asked our selves can AI be helpful in these processes. Obviously for many purposes the answer is no. AI cannot help with sharing internalised knowledge across a group. It cannot build embodies spatial memory of a complex system. But for other purposes it seems very powerful.
We designed a two-step process. First step was to generate a systems map with nodes and connectors. We prompted the AI to use SOD techniques like ZIP analyses and the Library of Systemic Relations.
The second step was more difficult: To make AI “paint” a visual gigamap based on the systems map. The “Prompter” and “Skill writer” has a role similar to a gardener. Many tests and tweaks were needed to develop it to a point where the gigamap felt useful.
Now we imagine that AI gigamapping can become very useful in any complex situation where we need to get fast access to complex and detailed information about a system to deepen our understanding to avoid simplistic decision making.
Below a AI generated video that explains how we do it. (Creds Ruben Sevaldson)