Evolving framework for building companionship among human and assistive systems

Abstract

The recent progress in artificial intelligence is allowing assistive systems like the voice-based assistant, virtual agents to become more personalized and adaptable. The role of these systems is also shifting from being a mere assistant to a person-al companion. However ‘personal companionship’ being a subjective term lies open to interpretation, thus posing a challenge for the creators of these assistive systems. This study is an attempt to address this challenge with a user-centric approach. Based on insights gathered from an activity based method called forced photo elicitation techniques with 25 users, we evolved Human Machine Companionship Framework, as a reference tool for designing effective personalized connections between assistive systems and its user. We describe each of the essential behavioural traits that a companion should exhibit and their evolution with time and information gained. Lastly, we establish the use of this companionship framework by discussing its application in case of the social robots.