FLORIDI, CAPURRO AND ARENDT IN DIALOGUE

ETHICAL AND CONCEPTUAL CONTRIBUTIONS TO EXPLAINABLE ARTIFICIAL INTELLIGENCE

Abstract

Advances in so-called Explainable Artificial Intelligence have been presented as a possible response to growing demands for algorithmic transparency. However, a conceptual and theoretical gap persists, since the assumptions that define what constitutes an adequate explanation are neither culturally neutral nor universally shared, making this a field that remains largely unexplored by Information Science. In this context, this study aims to foster an interdisciplinary dialogue between the works of Luciano Floridi, Rafael Capurro, and Hannah Arendt, with the goal of establishing ethical and conceptual foundations for the development of Explainable Artificial Intelligence systems in culturally heterogeneous contexts. Floridi’s ontocentric macroethics—grounded in the concepts of the infosphere, informational entropy, and moral agency—is articulated with Capurro’s intercultural ethics of information, anchored in the plurality of informational worlds. Complementarily, Arendt’s categories of plurality, action, and judgment are drawn upon to incorporate the political dimension of algorithmic explainability. The research adopts a qualitative theoretical-conceptual approach, combining philosophical conceptual analysis, intellectual genealogy, and hermeneutics. As a result, a framework composed of four guiding principles for the development of XAI systems is proposed: informational integrity, culturally situated readability, promotion of judgment capacity, and reduction of power asymmetries. It is concluded that the articulation between Floridi, Capurro, and Arendt offers an original contribution from Information Science to the development of AI systems that are more transparent, ethically grounded, and sensitive to sociocultural plurality.