Design patterns for human-AI co-learning: A wizard-of-Oz evaluation in an urban-search-and-rescue task

The rapid advancement of technology empowered by artificial intelligence is believed to intensify the collaboration between humans and AI as team partners. Successful requires partners learn about each other task. This human-AI co-learning can be achieved presenting situations that enable share knowledge experiences. In this paper we describe development implementation a task context procedures for studying co-learning. More specifically, designed specific sequences interactions aim initiate facilitate process. effects these interventions on learning were evaluated in an experiment, using simplified virtual urban-search-and-rescue human-robot team. human participants performed victim rescue- evacuation mission with wizard-of-Oz (i.e., confederate experimenter who executed robot-behavior consistent ontology-based AI-model). interaction sequences, formulated Learning Design Patterns (LDPs), intended bring Results show LDPs support understanding awareness their robot partner teamwork. No found fluency, nor performance. are used discuss importance co-learning, challenges designing tasks research into phenomenon, conditions under which likely successful. study contributes our how from AI-partners, propositions intentional (LDPs) provide directions applications future teams.