Making ethics practical: User stories as a way of implementing ethical consideration in Software Engineering

Context: Shortcomings of AI systems have recently brought ethics into the spotlight in Software Engineering (SE) in the form of AI ethics. However, actually implementing ethics into practice remains a challenge in both AI ethics and SE at large. Translating abstract ethical principles into requirements and features is difficult and lacks established processes, as well as practices and methods. Objective: In this study, we explored user stories as a way of implementing ethics in SE. Initially, we simply investigated whether user stories could be utilized for this purpose. After we began to consider this possible, we began to develop the concept of ethical user stories (EUSs) as a specific practice for this purpose. Method: We utilized a design science research (DSR) approach to first explore the use of user stories in implementing ethics, and then to develop the concept of EUS. This process featured three DSR phases through which the concept of EUS was iteratively developed with empirical data. Results: Over three DSR iterations, we studied 689 user stories produced in different contexts including both student and industry settings. Based on the data, we defined the concept of EUS and provided empirical validation for it. Conclusions: The concept of EUS provides a novel way of tackling ethics in SE. This paper presents the concept in-depth, along with practical suggestions for utilizing EUS.