Conference Program

In accordance with the ‘Guidelines’ issued by the Italian Ministry of Education (MIM, 2024), the teaching of civic education in primary schools - structured around three interconnected conceptual pillars: the Constitution, Sustainable Development, and Digital Citizenship - is established as a cross-curricular and foundational pedagogical paradigm, aimed at the holistic development of the individual and the citizen. Its transversal nature implies that educators are required to integrate citizenship content within their specific teaching pathways, transcending the logic of fragmented knowledge to promote a unified and critical understanding of social and institutional reality.

The 'on-life' reality (Floridi, 2017) in which we are immersed is increasingly integrated with Artificial Intelligence (AI) devices that act as invisible mediators: through the massive analysis of behavioural data, they filter information to steer user choices (Susser et al., 2019). This generates informational isolation that reinforces pre-existing biases and restricts encounters with diversity. The use of subtle persuasion techniques (nudges) and probabilistic calculations enables AI to anticipate and co-create individual desires, transforming apparent freedom of choice into a pre-channelled path (Zuboff, 2019). This scenario delineates a paradox of individual autonomy and raises crucial ethical questions regarding the transparency of systems that, by operating silently, redefine the boundaries of human agency and access to reality.

Given these premises, it is evident that, from primary school onwards, there is an urgent need to structure Artificial Intelligence (AI) education within the framework of Civic Education. This aligns with the digital citizenship pillar, which addresses the necessity of cultivating informed technology users capable of navigating virtual spaces with critical discernment. An early approach allows pupils to be transformed from passive consumers of algorithmic suggestions into conscious citizens, capable of recognising the probabilistic and non-deterministic nature of such technologies, thereby ensuring the preservation of free and pluralistic access to reality (Panciroli & Rivoltella, 2023). Without specific algorithmic literacy, children risk growing up within "filter bubbles" which, by catering exclusively to prior preferences, drastically limit encounters with otherness and divergent thinking. Therefore, educating on AI from primary school means deconstructing this paradox of individual autonomy by making explicit the silent mechanisms that redefine the boundaries of human will (Ranieri et al., 2023).

The following contribution presents a didactic experiment conducted with children in Year 1 and Year 2 of primary school in the metropolitan area of Naples. The project was structured around a co-writing activity (Lowry et al., 2004; Tinajero & Rojas-Drummond, 2012; Storch, 2019; Coccimiglio & Ranieri, 2023; Tsatsou-Nikolouli, 2025; Martins et al., 2025), followed by a comparison with AI. The latter was utilised as an alternative device to enable children to detect the adherence (or lack thereof) of algorithmic choices to the proposed text and the extent to which these diverged from the choices made by the children themselves, thereby highlighting AI biases.