Abstract for: Embracing the Complexity of the English Social Housing Sector: A New AI-Assisted Tool to Avoid Mental models’ Simplification

The UK housing crisis is a wicked problem, characterised by stakeholders with conflicting values, differing understandings of the issue, and the need for complex, interconnected responses. While participatory systems mapping can help capture diverse mental models, navigating this complexity is a daunting task. AI shows promise in supporting complex analyses, particularly in merging and simplifying CLDs. However, limitations inherent in aggregative processes, such as misrepresentation of complexity, remain untapped. We introduce an alternative approach to the analysis of multiple CLDs. In interviews with actors across the English social housing sector, we developed 38 CLDs depicting the issues surrounding housing provision and regeneration. To navigate their content, we designed an interacting, AI-assisted CLD Explorer tool; the tool generates a database that can be queried to extract and visualise stakeholder-specific cause-effect relationships around key variables (e.g. demolition, health, sustainability). By preserving tensions between stakeholders’ worldviews, this approach is critical to ‘embracing complexity’, fostering dialogue and understanding between the seemingly incompatible positions of different actors in the sector, and designing interventions that engage with them holistically. AI was used to develop the tool; the tool is an AI-assisted explorer