Explainable ArtificiaI Intelligence

Abstract

Artificial Intelligence (AI) has become one of the central axes of digitization strategies in the European Union (EU). Despite its great benefits, this group of emerging technologies also presents significant ethical risks and challenges that must be addressed. For this reason, the EU has proposed four fundamental ethical principles for AI: fairness, respect for human autonomy, prevention of harm, and explicability. The central concept of the paper is the explicability of AI (XAI). Besides contextualizing the relevance of XAI, this work aims to tackle some of the central aspects around it. Therefore, it will address what explicability is, why explanations are important (in the field of AI), when it is necessary to give explanations, and the different forms of explanations available to make citizens understand the decisions made by AI systems.