Reciprocal Human Machine Learning (RHML): Human-AI Collaboration based on theories of dyadic learning

Abstract

In this position paper we advocate a Reciprocal Human Machine Learning paradigm based on two theories of human-human learning behavior. Drawing from Jörg's theory of reciprocal learning in dyads and the Jewish tradition of Havruta - pair-based study, we suggest that human-machine collaboration based on these established human-human collaborative forms can achieve a rich and robust human-in-the-learning-loop (HITLL) framework in which both parties experience learning over time.