Personalized Complementarity in Human-AI Collaboration

Uncertainty and the potential for complementarity between humans and AI have shaped a new kind of interaction, typically referred to as a collaborative relationship. In this paper, we pick up on the discussion of common issues in human-AI collaboration based on a literature review and further discuss ways to and challenges of personalized complementarity in human-AI relations. We hypothesize that a combination of reciprocal, mixed-initiative communication may support up-to-date mental models and therefore strengthen appropriate trust and reliance, ultimately leading to a higher chance of effectively exploiting existing complementarity potentials.