Artificial Intelligence as a Lever for Reducing Cognitive Biases in Public Investment Evaluation: An Exploratory Institutional Analysis of the Moroccan Case

Résumé

Public investment assessment is a core component of achieving efficient allocation of public resources. Existing research in this field has two core flaws. First, the rational decision-making assumption that underpins traditional assessment models has been widely challenged in behavioral economics, and current studies have confirmed that cognitive biases can severely distort the assessment outcomes of public investment projects. Second, while the academic community has gradually recognized AI’s decision-support attribute, very few studies have tested AI’s role in mitigating cognitive biases in public investment assessment, and relevant research is particularly lacking in developing-country contexts. This study takes Morocco as its research setting, proposes a comprehensive analytical framework that integrates behavioral economics, public management, and AI decision support systems, and conducts exploratory empirical assessment using secondary institutional data from internationally recognized databases and official Moroccan sources. The study finds that AI can optimize information processing and improve assessment consistency, but AI’s effectiveness is constrained by three conditions: institutional quality, governance arrangements, and access to reliable data. Therefore, AI should serve as a supplementary decision-making tool rather than a replacement for human judgment. This study also puts forward the concept of augmented rationality under institutional constraints, which expands existing research frameworks by integrating multi-dimensional perspectives.

Classification JEL: D91, H54, O38, C83

Paper type: Exploratory Empirical Research