Advances in Human-AI Collaboration

7Human-AI Teaming

Tianyi Yuan, Minqian Yang, Dian Yu, and Pei-Luen Patrick Rau

Department of Industrial Engineering, Tsinghua University, Beijing, China

7.1 Theoretical Foundations

7.1.1 Definitions

The concept of human–AI teaming is an emerging interdisciplinary utility shaped by recent developments in artificial intelligence (AI) and its implementation in different domains. Over the past three decades, scientists have explored the concept of human–AI teaming. At first, AI systems were built to be autonomous, identifying and solving problems without human input. Nevertheless, as AI technologies evolved, it was discovered that there was a glaring need for human expertise and judgment to enhance decision-making processes. This idea began to turn into human-centered collaborative systems. Humans and AI were conjured more as partners who would work and live with each other instead of some distant terrestrial being.

Various academics have suggested various definitions of human–AI teaming. The nearest definition comes from Berretta et al. (2023), who define human–AI teaming as “a process between one or more human(s) and one or more (partially) autonomous AI system(s) serving as team members with distinctive and complementary capabilities, who work interdependently toward a common goal.” Besides, Van Zoelen et al. (2021) present a series of dynamic definitions of human–AI teaming from various perspectives (co-adaptation, co-learning, co-evolution), where humans and AI adapt over ...