Augmenting Human Teams with Robots in Knowledge Work Settings: Insights from the Literature

Abstract

Recent developments in large language models open doors for AI and robots to augment knowledge workers and teams in a variety of domains, such as customer service, data science, legal work, and software development. In this essay, we review 317 articles from multiple disciplines and summarize the insights in a theoretical framework linking key robot attributes to human perceptions and behaviors. The robot attributes include embodiment, nonverbal and verbal communication, perceived gender and race, emotions, perceived personality, and competence. The outcomes include human perceptions, acceptance, engagement, compliance, trust, and willingness to help. We identify four differences between one human and one robot settings and team settings and use them as the springboard to generalize insights from the literature review to the design and impact of a robot in assisting humans in knowledge work teams. We report two high-level observations around the interplay among robot attributes and context dependent designs and discuss their implications.