Digital Exclusion in the Age of Artificial Intelligence: Competences, Algorithmic Narratives, and the Reproduction of Social Inequalities

Digital technologies, and more recently artificial intelligence, have reshaped not only the possibilities for social and political participation but also the mechanisms of exclusion. This paper addresses digital exclusion in its contemporary, expanded form, considering the growing role of AI as a cognitive, communicative, and political infrastructure. It examines how findings from research on digital competences and exclusion in Poland can inform a critical reflection on emerging forms of marginalisation associated with the diffusion of AI. In contexts marked by unequal access to technology and uneven levels of digital literacy, focusing solely on those already excluded is insufficient. Equally important is the analysis of those at risk of exclusion as AI systems increasingly mediate access to information, public discourse, and institutional services. AI is thus approached not merely as a technological tool but as a ‘non-human actor’ (Latour) that actively shapes social imaginaries through algorithmic narratives. Drawing on Pierre Bourdieu’s theory, social class is conceptualised as a multidimensional configuration of economic, social, and cultural capital, of which digital competences have become a crucial contemporary component. Recognised by the European Commission as essential for social and economic life, as well as active civic participation and political agency, digital competences increasingly function as markers of social position. Their absence deepens existing inequalities and reinforces the nexus between digital and social exclusion. Consequently, digital competences should be understood not as a matter of individual choice, but as a structural and political issue embedded in power relations. The paper argues that dominant narratives of individual responsibility contribute to the privatisation of digital exclusion. Lack of competence is framed as a personal failure rather than a product of social conditions. In this context, AI — much like the school system in Bourdieu’s analyses — operates as an instrument of reproduction. AI legitimises the dominant social order by presenting particular linguistic styles, argumentative patterns, and cultural references as universal and neutral. In the Polish context, this process results in the systematic marginalisation of working-class experiences within algorithmic narratives. The absence of these perspectives limits the articulation of social grievances, contributing to alienation and undermining democratic legitimacy. Simultaneously, the ability to critically engage with AI has become a new axis of differentiation within the middle class itself. The paper conceptualises AI as a mechanism of symbolic violence, trained on datasets dominated by middle- and upper-class cultural patterns. Finally, it asks whether AI will remain an apparatus of symbolic domination or whether, under democratic governance, it might support more inclusive forms of participation. The future of digital democracy depends on collective political choices regarding how AI is designed, regulated, and socially embedded.