Responsible and human centric AI-based insurance advisor

Abstract

In this paper we present novel approach, based on AI, to suggest insurance coverage for users and families. It is not yet fully explored the domain on what does it mean to give responsible and human centric recommendation in the context of AI-based insurance. Therefore, in this article we provide an in-depth analysis on i) specifications and requirements for the system from regulation point of view, ii) instructions on how to design such system and on which data it can rely upon, iii) which recommender techniques can be used for developing such advisor, and finally iv) off-the-shelf components for trustworthy, responsible ethical behavior of this Ai-empowered tool. Solution as proposed in our paper will be transparent, trustworthy, and responsible towards the final users, thus we hope, better accepted by customers. After describing such possible system design and architecture, we discuss critically challenges and opportunities for deployment of such systems in insurance companies.