Data, Algorithms and Otherness. The Erasure of the Other

In a context increasingly shaped by AI-based technologies to which decision-making and agency are often delegated, this paper examines the potential role of design in preserving alterity. We begin by presenting critical perspectives on the concept of data, essential to AI-based technologies, emphasizing its situated nature and the associated dynamics of power, oppression, and limitations in representing diverse phenomena and social groups. We then provide an overview of AI technologies, followed by a discussion of strategies available for articulating the "non-inclusion" inherent in these systems. Two complementary approaches are identified: "in-vitro", aimed at reconstructing and demystifying the often opaque processes underpinning algorithmic operations, and "in-the-field", focused on observing and documenting examples of anomalies in these technologies as they function in real-world contexts, highlighting elements that elude algorithmic classification. This latter body of works, which can be seen as “catalogs of errors,” serves as a foundation for exploring algorithmic otherness and for advancing more inclusive technologies.