User speech rates and preferences for system speech rates

Abstract

Prior studies of human communication have demonstrated that prosocial outcomes occur when facets of communication converge between interlocutors—for example, social likeability and perceived competence increase when people adapt their pitch or rate of speech towards one another. Inspired by these findings, we examined whether user preference for fast or slow system speech rate covaries with user speech rate. Experimental participants uploaded samples of their reading speech and spontaneous speech, and then listened to fast and slow system speech before choosing their preferred system speech rate. Results aligned with our hypothesis that fast speakers prefer faster system speech and that slow speakers prefer slow. Design implications are discussed, as well as plans for future studies that explore the potential for systems to automatically measure the user’s speech rate and converge towards it.