The Components of Trust for Collaborating With AI Colleagues

Abstract

AI technologies are capable of improving the performance and productivity of teams in a variety of work contexts. These advantages may be optimized when the AI agent is considered a full team member. A vital component of the agent’s acceptance as a team member or colleague is the degree to which its human coworkers feel they can trust it. To explore what factors affect the perceptions of an AI agent as a trustworthy team member and a legitimate colleague, we interviewed twenty-two professionals representing various work roles. Our results revealed that the following qualities contribute to professionals’ trust in AI as a colleague: a visual presence reflective of coworkers, engagement in feedback loop and team processes through human communication, and the ability for self-development. These findings contribute to the CSCW community by advancing the current understanding of human-AI teaming and informing the design of trustworthy AI agents into the workplace.