Perceived Shared Understanding Between Humans and Artificial Intelligence: Development and Validation of a Self-Report Scale

Abstract

The perception of shared understanding between individuals is key to constructive meaning-making, effective collaboration, and satisfying interaction outcomes. Recent scholarship indicates this operation extends to human–artificial intelligence interactions such that a validated instrument to capture humans’ perceived shared understanding (PSU) with artificial intelligence is key to advancing work in that domain. Building on extant exploratory work, this project develops and initially validates a PSU scale in two studies. Participants shared past large-language model conversations and then reflected on them to respond to a pool of candidate scale items. Exploratory factor analysis yielded a single-factor, eight-item solution interpreted to represent a social-semantic construal of the artificial intelligence’s shared understanding—that is, that they are sharing meaning with someone. The scale demonstrates significant associations with theoretically relevant measures; factor structure and convergent validity are replicated in a separate sample. This novel instrument points to a convergence of sociality and meaning in PSU and serves as a springboard for future research and practical applications.