Data-driven discrimination. How Artificial Intelligence can (re)produce spatial asymmetries in urban space

This work aims to contribute to the debate on the transformative effects generated by digital technologies in urban contexts (Lazzeroni, 2004), by focusing on the studies on platform urbanism and the effects of algorithmic automation and Artificial Intelligence (AI) in the space of places. In particular, the contribution explores the problematized case of Airbnb as an ‘actor of the urban fabric’ in the city of Florence (Italy), which is addressed from an evolutionary point of view. Starting from studies on the dynamics of polarization (Picascia et al., 2017), neighbourhood effect (Gurran & Phibbs, 2017; Celata & Romano, 2022), algorithmic automation (Celata et al., 2020; Romano et al., 2023), the research questions explore a) what the effects of the implementation of AI are in those digitally-mediated processes that already generate transformative effects at the intra-urban scale and b) whether these models contribute to mitigating or amplifying pre-existing socio-spatial asymmetries. The results discuss different levels of data-driven discrimination: the first type of discrimination is between those who produce data and those who do not; the second is between those who can circumvent algorithms to their advantage and those who cannot; the third concerns the spatial polarization of the Airbnb platform at the intra-urban scale and the centre-periphery relation. In this context, the work discusses the potential risks of quickly adopting AI for economic purposes, which could amplify pre-existing spatial asymmetries in urban space.