Year: 2025
Type: preprint
Abstract: Humans and machines working together in teams is becoming increasingly common. For human-machine teams to succeed, the humans must have a positive perception of their machine teammate in terms of its capability and social presence. We begin by discussing why a measure of machine teammate perceptions is needed. Then, we develop and validate a 20-item multidimensional scale, which we label as the Human-Machine Teammate Inventory (HMTI), to measure perceptions of a machine teammate quality consisting of task trustworthiness, cognitive understanding, explainability, empathy, and liking dimensions. Evidence from four large population samples provides support that (a) our proposed 5-factor, bifactor structure replicates and fits the data better than alternative models; (b) the HMTI exhibits high reliability and content validity; (c) the HMTI demonstrates evidence of convergent and discriminant validity with other constructs in its nomological network; and (d) the HMTI demonstrates incremental validity in the prediction of human-machine teaming outcomes over and above related measures of machine social perceptions. Overall, results from our study show that the HMTI is a valid and reliable measure, providing a more nuanced approach to evaluating human-machine teamwork. The HMTI provides researchers and practitioners with a scale that can help advance our understanding of how to promote effective human-machine teamwork.
Authors: Ryan Grant, Tyson Jordan, Adam Goodie, Prashant Doshi
Institution: University of Georgia
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: gold
Award: G7452299184
OpenAlex ID: W4409487440