Barriers to Institutional AI Integration: A Systematic Approach Using FUCOM and Segmentation Analysis

Artificial Intelligence (AI) technologies promise revolutionary transformations in many areas, from decision-making processes to service delivery within the public sector. However, despite this potential, the effective utilization of AI at institutional and societal levels in the public sector faces significant barriers. Multi-layered challenges such as deficiencies in technological infrastructure, concerns about data security and privacy, ethical uncertainties, employee resistance, inadequate legal and administrative frameworks, and limited organizational learning capacity constitute major barriers to the widespread adoption of AI applications. This study analyses the primary barriers to AI utilization in public institutions through a holistic and systematic perspective. It identifies and prioritizes these barriers using the Full Consistency Method (FUCOM), a Multi-Criteria Decision-Making (MCDM) approach based on expert judgement. After weighting the barriers, decision-makers within public sector organizations are clustered using segmentation analysis based on their professional profiles and contextual characteristics, thereby revealing resistance patterns and varying levels of institutional readiness in more detail. The findings indicate that, beyond technical and infrastructural limitations, cognitive, ethical, and institutional factors also contribute to delays in AI integration. Drawing on both the literature and empirical data, the study proposes strategic recommendations to overcome these barriers and discusses the governance structures necessary for the sustainable dissemination of AI applications. In addition, by combining methodological rigor with practical insight, this study offers evidence-based recommendations to address the challenges hindering the systematic deployment of AI and highlights the governance approaches essential for overcoming future adaptation constraints.