Human-autonomy Teaming: Need for a guiding team-based framework?

Whereas high-performance teamwork has been studied empirically for 70 years, a new form of teaming is on the rise. Enabled through the rapid progression of artificial intelligence, a human-autonomy team (HAT) involves one or more autonomous computerized agents collaborating with humans on interdependent tasks toward the achievement of a common goal. Whereas research on HATs is exploding in recent years, that research has not strongly embraced the vast literature, theory, and methods already developed in the all-human teaming literature. Moreover, definitional and construct validity issues, in terms of what constitutes a HAT, persist in the literature. In the current article we offer construct clarity and we integrate the Input-Mediator-Output model from the high-performance teaming literature to help future researchers classify the variables under study, theorize deeper, and consolidate findings across studies. Both the construct clarity we offer and our theoretical integration will serve as a valuable perspective for contextualizing the studies in the current Special Issue as well as in designing and interpreting future research in the HAT area.

Automated summary

While high-performance teamwork has been studied extensively, the rise of human-autonomy teams (HAT) enabled by artificial intelligence is a new research area. However, this field has not adequately incorporated existing literature, theory, and methods from traditional all-human teaming. This article provides construct clarity and integrates the Input-Mediator-Output model to help researchers classify variables and consolidate findings in HAT research. The construct clarity and theoretical integration offered here will serve as a valuable perspective for current and future research in the HAT field.