Smart Territorial Governance: The Moderating Role of Artificial Intelligence between Institutional Capacity and Regional Performance in Morocco

Abstract

Context. Advanced regionalization grants Moroccan regions growing autonomy, but their institutional capacity to turn it into performance remains uneven, and its articulation with the integration of artificial intelligence (AI) remains little explored.

Objective. This article addresses the following question: to what extent does AI integration moderate the relationship between institutional capacity and the performance of territorial development strategies in Moroccan regions? It proposes a model in which AI acts as a moderating variable, not as an autonomous determinant of performance.

Method. The study draws on the territorial management framework (2016-2022) and a recent literature review (2023-2026). The theoretical propositions are illustrated through a cross-checked documentary analysis based on about ten academic, institutional and specialized-press sources. Three contrasting Moroccan configurations, Casablanca, Rabat and Benguerir, are placed in perspective with two international experiences, Catalonia and Rwanda.

Results of the qualitative analysis. The three configurations reveal differentiated modes of technological integration: still sectoral tooling in Casablanca, a gap between ambitions and appropriation in Rabat, and advanced integration weakly articulated with regional steering in Benguerir. These qualitative results suggest, without demonstrating it statistically or measuring its magnitude, that AI integration may strengthen the translation of institutional capacities into performance when these are sufficiently consolidated. They also underline the importance of data quality and organizational capabilities.

Contribution and limitations. The article conceptualizes AI as a moderating variable and proposes an operationalization framework for later empirical validation. This contribution remains exploratory: qualitative illustration not statistically validated, uncontrolled confounding variables and heterogeneous units of analysis. The findings should therefore be interpreted as evidence of theoretical plausibility rather than as a causal demonstration.

Classification JEL: H70 ; O33 ; R58

Paper type: Theoretical Research (with qualitative illustration)