Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


School of Computing

Committee Chair/Advisor

Nathan J. McNeese

Committee Member

Brian Dean

Committee Member

Eileen Kraemer

Committee Member

Brygg Ullmer

Committee Member

Laine Mears


The introduction of computational systems in the last few decades has enabled humans to cross geographical, cultural, and even societal boundaries. Whether it was the invention of telephones or file sharing, new technologies have enabled humans to continuously work better together. Artificial Intelligence (AI) has one of the highest levels of potential as one of these technologies. Although AI has a multitude of functions within teaming, such as improving information sciences and analysis, one specific application of AI that has become a critical topic in recent years is the creation of AI systems that act as teammates alongside humans, in what is known as a human-AI team.

However, as AI transitions into teammate roles they will garner new responsibilities and abilities, which ultimately gives them a greater influence over teams' shared goals and resources, otherwise known as teaming influence. Moreover, that increase in teaming influence will provide AI teammates with a level of social influence. Unfortunately, while research has observed the impact of teaming influence by examining humans' perception and performance, an explicit and literal understanding of the social influence that facilitates long-term teaming change has yet to be created. This dissertation uses three studies to create a holistic understanding of the underlying social influence that AI teammates possess.

Study 1 identifies the fundamental existence of AI teammate social influence and how it pertains to teaming influence. Qualitative data demonstrates that social influence is naturally created as humans actively adapt around AI teammate teaming influence. Furthermore, mixed-methods results demonstrate that the alignment of AI teammate teaming influence with a human's individual motives is the most critical factor in the acceptance of AI teammate teaming influence in existing teams.

Study 2 further examines the acceptance of AI teammate teaming and social influence and how the design of AI teammates and humans' individual differences can impact this acceptance. The findings of Study 2 show that humans have the greatest levels of acceptance of AI teammate teaming influence that is comparative to their own teaming influence on a single task, but the acceptance of AI teammate teaming influence across multiple tasks generally decreases as teaming influence increases. Additionally, coworker endorsements are shown to increase the acceptance of high levels of AI teammate teaming influence, and humans that perceive the capabilities of technology, in general, to be greater are potentially more likely to accept AI teammate teaming influence.

Finally, Study 3 explores how the teaming and social influence possessed by AI teammates change when presented in a team that also contains teaming influence from multiple human teammates, which means social influence between humans also exists. Results demonstrate that AI teammate social influence can drive humans to prefer and observe their human teammates over their AI teammates, but humans' behavioral adaptations are more centered around their AI teammates than their human teammates. These effects demonstrate that AI teammate social influence, when in the presence of human-human teaming and social influence, retains potency, but its effects are different when impacting either perception or behavior.

The above three studies fill a currently under-served research gap in human-AI teaming, which is both the understanding of AI teammate social influence and humans' acceptance of it. In addition, each study conducted within this dissertation synthesizes its findings and contributions into actionable design recommendations that will serve as foundational design principles to allow the initial acceptance of AI teammates within society. Therefore, not only will the research community benefit from the results discussed throughout this dissertation, but so too will the developers, designers, and human teammates of human-AI teams.

Author ORCID Identifier




To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.