Human-AI Interaction for Diverse Humans: What Cognitive Style Disaggregation Reveals

Although guidelines for human-AI interaction (HAI) provide important advice on how to help improve user experiences with AI products, little is known about HAI for diverse users' experiences with AI. Without understanding factors that lie behind differences among diverse users' experiences with AI products, designers lack information they need to make AI products more equitable and inclusive. To investigate whether and how diverse users' different cognitive styles might help account for their differences, we used data from 16 experiments on Amershi et al.'s HAI Guidelines, and disaggregated by the participants' cognitive styles. The results of disaggregating revealed 112 phenomena that were not apparent without taking cognitive style diversity into account. We also show how the cognitive style differences can explain demographic differences among genders and among gender-age intersectional groupings, and can point the way toward making HAI experiences more equitable and inclusive.