AI adoption, structural constraints and inclusive growth: Evidence from Greece

Artificial intelligence (AI) is increasingly characterized as a General Purpose Technology (GPT) capable of reconfiguring work, accelerating digitalization, and raising aggregate productivity. Yet the pace and inclusiveness of AI adoption depend less on technical potential than on complementary sectoral capacity, institutional quality, and human capital. This paper examines Greece as a peripheral EU economy marked by high projected AI gains but significant structural constraints, generating risks of delayed adoption and uneven distributional outcomes. The paper develops a structural conversion framework linking three dimensions: capacity for AI-driven sectoral upgrading; institutional and functional readiness beyond digital infrastructure; and the regeneration and retention of skilled labour. Drawing on micro-level data from the JustReDI survey, the paper constructs composite indices of institutional trust, digital confidence, and digital public service usability, documenting substantial heterogeneity across educational subgroups and regions. The findings reveal a pronounced structural mismatch: service usability rises sharply with educational attainment, yet institutional trust varies only modestly, and digital confidence is performance-driven rather than legitimacy-based. A further tension emerges in Attica, where high institutional trust and technological centrality coexist with comparatively poor user-centric service quality. These results point to a fragmented digital transition marked by persistent social and regional asymmetries, suggesting that inclusive growth through AI adoption in structurally constrained economies requires a policy reorientation - from deployment toward the institutional and functional conversion of technological resources into broad-based development outcomes.