Embedded Ethics for Responsible Artificial Intelligence Systems (EE-RAIS) in disaster management: a conceptual model and its deployment

In this paper, we argue that Responsible Artificial Intelligence Systems (RAIS) require a shift toward embedded ethics to address value-based challenges facing AI in disaster management; and we propose a model to achieve it. Disaster management requires Artificial Intelligence Systems (AIS) that would be sensitive to ethical, legal, and multi-dimensional values while being responsive and accountable in complex and acute disruptions that simultaneously call for fair, value-laden, and immediate decisions. Without such a necessary shift, AIS will be incapable of responding properly to major value-based challenges of axiological and hierarchical types, and might leave AIS vulnerable to meta-disasters, such as intelligent digital disasters. This study focuses on RAI in the context of disaster management and proposes a model of Embedded Ethics for Responsible Artificial Intelligence Systems (EE-RAIS), which is empowered by four platforms of embedded ethics—educational, cross-functional, developmental, and algorithmic embedded ethics—as well as four imperative metrics—ethical intelligence, legal intelligence, social-emotional competency, and artificial wisdom. The final section of the paper explores how EE-RAIS can be deployed for the purpose of disaster management and fair crisis informatics.