Abstract

This study presents a technical, legal, and philosophical analysis of the intricate relationship between big data, artificial intelligence and explanations. The presence of heterogeneous datasets used as input for machine learning techniques raises questions on a possible broadening of the conceptualisation of algorithmic Explicability to cover Knowability elements that also include data-related features. This paper proposes the inclusion of dynamics elements of explanations that cover the entire workflow of data analysis, from input data to the automated decision, consistently with research and governance trends on the Explicability of artificial intelligence systems