Elaborateness of Explanations to Understand AI Recommendations

Successful human-AI collaboration requires an understanding of the AI system which can be achieved by human-interpretable explanations.In an experimental study (N = 109), we found more elaborated explanations to foster causability and trust, whereas adaptability of elaborateness provides no further benefit.Although cognitive load was not affected by explanations, its causal role requires further investigation.Future research should integrate learning sciences with explainable AI research to consider system and human aspects of understandable AI.