A Scenario-Based Model for Analyzing the Ethical Requirements of Intelligent Autonomous Systems

In this paper, a model for eliciting and analyzing the ethical requirements of an intelligent system is proposed. Considering that intelligent systems are increasingly entering various ethical territories of human life and becoming able to make decisions with ethical consequences, it is suggested that we need to develop guiding models with the perspective that they actively involve requirements engineers in discovering the ethical requirements of such systems during a system's life cycle. In a sense, beyond giving engineers pre-specified ethical standards, engineers themselves, in collaboration with each other and stakeholders, should also be able to iteratively analyze the ethical requirements which are specific to the system under development. A model is better to mainly provide a framework for such a collaborative process. What is suggested in this paper is a scenario-based model of this kind. The effectiveness of scenarios in discovering and specifying a system's requirements has been explored well in the literature. It is argued both how we can use scenarios for ethical purposes in the process of requirements engineering and how philosophical ethical theories can be used in the proposed model by considering, as an example, three influential ethical theories: consequentialism, deontology, and virtue ethics.