ABSTRACT
Algorithms have advanced in status from supporting human decision-making to making decisions for themselves. The fundamental issue here is the relationship between Big Data and algorithms, or how algorithms empower data with direction and purpose. In this paper, I provide a conceptual framework for analyzing and improving ethical decision-making in Human-AI interaction. On the one hand, I examine the challenges and the limitations facing the field of Machine Ethics and Explainability in its aim to provide and justify ethical decisions. On the other hand, I propose connecting counterfactual explanations with the emotion of regret, as requirements for improving ethical decision-making in novel situations and under uncertainty. To test whether this conceptual framework has empirical value, I analyze the COVID-19 epidemic in terms of "what might have been” to answer the following question: could some of the unintended consequences of this health crisis have been avoided if the available data had been used differently before the crisis happened and as it unfolded? © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.