An ontology-based approach to engineering ethicality requirements

Abstract In a world where Artificial Intelligence (AI) is pervasive, humans may feel threatened or at risk by giving up control to machines. this context, ethicality becomes major concern prevent AI systems from being biased, making mistakes, going rogue. Requirements Engineering (RE) the research area that can exert great impact in development of ethical design. However, proposing concepts, tools and techniques support incorporation into software processes as explicit requirements remains challenge RE field. paper, we rely on Ontology-based (ObRE) method elicit analyze (‘Ethicality requirements’ adopted name for class studied paper analogy other quality engineering, such usability, reliability, portability, etc. The use term (as opposed ‘ethical requirements’) highlights they represent systems, analogous how ‘trustworthiness trustworthy systems. To put simply: predicates ‘ethical’ ‘trustworthy’ are not meant be predicated over themselves). ObRE applies ontological analysis ontologically unpack terms notions referred elicitation. Moreover, instantiates ontology uses it guide activity. previous presented solution concerning two principles, namely Beneficence Non-maleficence. present extends work targeting important those Explicability Autonomy. For each these new do unpacking relevant elicitation guidelines, well examples context driverless car case. Furthermore, validate our approach analysing made case contrast with similar case, assessing method’s coverage w.r.t European Union guidelines Trustworthy AI.