Work and artificial intelligence in Italy: between opportunities and risks

Purpose This study aims to analyse how artificial intelligence (AI) may affect the Italian labour market by measuring workers' potential exposure to AI-driven substitution and AI-enabled complementarity. In particular, it identifies which groups of workers, occupations and sectors are likely to face the highest substitution pressure and which are more likely to benefit from the integration of AI. Design/methodology/approach Building on the USA-based AI Occupational Exposure Index, this study develops two AI exposure indices for Italian occupations, AIOEit and CAIOEit. The indices combine AI capabilities, measured through the MTurk survey, with occupation-level abilities and working conditions from the Italian Survey on Occupations. They are then linked to the Continuous Labour Force Survey to examine regional, sectoral and socio-demographic patterns of AI exposure. Findings The results suggest that, although routine and low-criticality jobs remain the most vulnerable to automation risk, many highly AI-exposed Italian occupations may benefit from AI integration, as their exposure is predominantly complementary rather than substitutive. The analysis of employees shows that around one in four Italian workers faces a potential risk of automation, while a similar share could benefit from AI as a productivity-enhancing technology. Originality/value This study develops the first Italian-specific AI exposure indices based on the latest available Italian occupational data. It provides both an empirical and theoretical framework to monitor the impact of AI on the Italian labour market and design targeted innovation policies.