AI in the Public Sector: An Exploration of Social Imaginaries

The use of artificial intelligence (AI) enables the transformation of both This study aims to investigate the imaginaries and narratives of AI in leading Italian newspapers, specifically evaluating the impact of ChatGPT on public discourse. organizations and institutions alike in many different fields. This paper explores the social imaginaries (i.e. set of imagined and collectively shared ideas, practices, and values people use to construct and legitimize reality) in relation to the use of AI in the public sector. According to Castoriadis (1975), human imagination plays a key role in the process of the “self-generation” of the current social order. Indeed, technological affordances and social practices are intertwined and mutually shape each other, and both contribute to the generation of what is called the “social imaginary.” Through a discourse analysis, the paper investigates the most popular narratives in European and Italian public policies framing and regulating the implementation of AI in the public sector. Then, the paper explores the AI imaginaries of Italian public sector through 15 interviews with AI experts (i.e., designers and managers) from the Italian public sector. Differences and similarities in the AI social imaginaries between policies and experts are highlighted. Experts advocate the need for a human-centered design approach in the implementation of AI to map users’ needs and to design useful public services. By adopting Castoriadis’ (1975) approach, we identify opposition between the heteronomous imagination, the one imposed by public policies, and the autonomous imagination, based on citizens’ needs and perceptions, including emotional ones. The paper calls for a human- centered design approach in the implementation of AI in the public sector.