Ethical reasoning in technology: using computational approaches to integrate ethics into AI systems

Abstract

Purpose This paper does not concern with the “why” of ethics. Such questions are typically of interest to philosophers and are outside the scope of this work. In the next section, the authors offer a look into “what” of ethics, i.e. various types and subtypes of ethics. Subsequently, the authors explore “how” of ethics, by summarising various computational approaches to ethical reasoning offered by researchers in the field. Design/methodology/approach The approaches are classified based on the application domain, ethical theory, agent type and design paradigm adopted. Moreover, promising research directions towards ethical reasoning are also presented. Findings Since the field is essentially interdisciplinary in nature, collaborative research from such areas as neuroscience, psychology, artificial intelligence, law and social sciences is necessary. It is hoped that this paper offers much needed insight into computational approaches for ethical reasoning paving way for researchers to further engage with the question. Originality/value In this paper, the authors discussed vaious computational approaches proposed by researchers to implement ethics. Although none of the approaches adequately answer the question, it is necessary to engage with the research effort to make a substantial contribution to the emerging research area. Though some effort has been made in the design of logic-based systems, they are largely in stages of infancy and merit considerable research.