Administrative traditions and barriers to AI implementation in public administration: an exploratory study with case studies from Italy and Spain

The aim of this thesis is to advance the understanding of barriers to Artificial Intelligence (AI) implementation in public administration by comparing two countries with similar administrative traditions: Italy and Spain (Sotiropoulos, 2006; Kickert, 2011; Kullman & Wollman, 2019). Although they belong to the Southern European administrative model, performance differences in digital transformation have been observed, as highlighted by e-government indexes (European Commission, 2023; United Nations, 2024; Zehiali & Mussari, 2025). Recent indicators, such as the AI Global Index (Mostrous et al. 2024) and the Artificial Intelligence Preparedness Index (AIPI, 2024), rank Spain ahead of Italy in AI readiness, while the Oxford AI Readiness Index, which focuses specifically on the public sector, shows Italy outperforming Spain. This inconsistency underscores the need to investigate how historical and cultural factors influence AI implementation within public administrations. To address this, the thesis employs a comparative case study approach (Agranoff and Radin, 1991) and applies the Technological-Organisational-Environmental (TOE) framework (Tornatzky & Fleischer, 1990) to categorise barriers and enabling factors and operationalises characteristics of the Southern European administrative tradition to explain cross-country differences. Empirical evidence derives from semi-structured interviews with civil servants across local, regional, and national levels in both countries. Findings reveal that in Spain, major barriers include data quality, limited funding, and privacy compliance constraints, while in Italy, organisational challenges such as internal resistance to change and data quality issues dominate. Regarding administrative traditions, Spanish civil servants predominantly perceive legalism as the main obstacle, whereas Italian respondents emphasise bureaucratisation, especially at lower levels of government. The study makes conclusions on the influence of historical and cultural elements on AI implementation and contributes to the growing body of empirical research on AI implementation in the public sector (Mergel et al., 2023; Merhi, 2023; Tangi et al., 2023).