Unlocking the Potential of Machine Ethics with Explainability

Roughly speaking, the research field of machine ethics deals with devising behavioral constraints on computational systems to ensure restricted, morally acceptable behavior. The potential benefits of researching machine ethics are substantial, encompassing contributions to ethical AI development and the societal impact of computational systems. However, there are genuine concerns and risks associated with this research (e.g., the potential to undermine human autonomy) that must be carefully considered. In this article, we will explore the question of whether it is worthwhile to conduct research in machine ethics, given the potential demerits and challenges involved. Central to our study is the proposition that explainability, such as is being explored in connection with explainable artificial intelligence (XAI), can serve as a powerful tool to augment the advantages of machine ethics research, mitigate its disadvantages, and create unique advantages of its own. Overall, we conclude that the study of machine ethics is worthwhile, especially when it is supported by research on explainability.