How AI is understood: perceptions among italian citizens, teachers, and public administration

The chapter explores how artificial intelligence (AI) is understood, interpreted and governed in Italy across three domains: citizens, schools and public administration. It adopts the perspective of science communication as a ‘social conversation around science’. Among Italian citizens, a persistent gap emerges between widespread exposure to AI and a low level of perceived knowledge. Symbolic representations, such as humanoid robots, remain central, while the association with generative chatbots is growing. Demand for regulation is also increasing, and acceptance of AI appears selective: high in safety-related domains, much lower in journalism, artistic production and workforce management. Teachers show cautious openness towards students’ use of AI for independent study. Yet its use in teaching remains marginal and perceived preparedness is limited, with significant differences by gender, school level and geographical area. In public administration, AI emerges as a technology ‘in suspension’. It is tested mainly in back-office functions as a tool for assisted automation, and its use remains subject to the human-in-the-loop principle, consistent with the logic of the European AI Act. Overall, the Italian case reveals a pattern of critical ambivalence and ‘conditional adoption’, in which trust appears as a relational and context-dependent construct. The chapter argues that understanding the different forms of cultural, institutional and communicative mediation of AI is essential for developing communication and governance strategies that can align technological development with societal values and expectations.