Rethinking Inner Areas Through AI: Hybrid Approaches for Inclusive Planning

This paper critically examines the implications of Artificial Intelligence (AI) for inner area regeneration, serving as a vital testing ground for AI’s contradictions in territorial planning. As non-neutral technologies, AI systems prioritize quantifiable data, often rendering invisible those local resources that resist simple encoding. In these regions, territorial value is deeply rooted in relational, cultural, and ecological dimensions that risk being overlooked by quantitative models. As spatial planning increasingly relies on AI-driven predictive models, there is a danger of standardizing territorial narratives and flattening local specificities. Moreover, predictive models trained on historical data often perpetuate existing inequalities. Inner areas face a “double disadvantage”: scarce, low-quality data produces distorted, self-referential forecasts that reinforce marginalization. To counter these risks, engaging local communities in dataset creation and the rediscovery of cultural specificities is essential. Drawing on the PRIN INSITE project, this paper proposes Hybrid Intelligence (HI) practices where AI systems and local knowledge coexist through participatory processes. The objective is to shift territorial development from an efficiency-centric model toward an inclusive, contextual paradigm that accurately interprets the unique potential of inner areas.