Towards a Computational Architecture for Co-Constructive Explainable Systems

Abstract

In this paper we consider the interactive processes by which an explainer and an explainee cooperate to produce an explanation, which we refer to as co-construction. Explainable Artificial Intelligence (XAI) is concerned with the development of intelligent systems and robots that can explain and justify their actions, decisions, recommendations, and so on. However, the cooperative construction of explanations remains a key but under-explored issue. This short paper proposes an architecture for intelligent systems that promotes a co-constructive and interactive approach to explanation generation. By outlining its basic components and their specific roles, we aim to contribute to the advancement of XAI computational frameworks that actively engage users in the explanation process.