Disentangling Human-AI Hybrids

TLDR

AI holds great potential for organizations, but realizing this requires human‑AI interworking, which is still poorly understood in a holistic sense. This study seeks to clarify the direction of human‑AI interworking by presenting a taxonomy of human‑AI hybrids derived from a literature review and 101 case examples. The authors developed the taxonomy using weak sociomateriality as a theoretical lens and performed cluster analysis on the sample to identify archetypal human‑AI hybrid configurations. The taxonomy establishes a foundation for understanding human‑AI hybrids, while the derived archetypes demonstrate the diverse roles AI systems can assume in interworking scenarios.

Abstract

Abstract Artificial intelligence (AI) offers great potential in organizations. The path to achieving this potential will involve human-AI interworking, as has been confirmed by numerous studies. However, it remains to be explored which direction this interworking of human agents and AI-enabled systems ought to take. To date, research still lacks a holistic understanding of the entangled interworking that characterizes human-AI hybrids, so-called because they form when human agents and AI-enabled systems closely collaborate. To enhance such understanding, this paper presents a taxonomy of human-AI hybrids, developed by reviewing the current literature as well as a sample of 101 human-AI hybrids. Leveraging weak sociomateriality as justificatory knowledge, this study provides a deeper understanding of the entanglement between human agents and AI-enabled systems. Furthermore, a cluster analysis is performed to derive archetypes of human-AI hybrids, identifying ideal–typical occurrences of human-AI hybrids in practice. While the taxonomy creates a solid foundation for the understanding and analysis of human-AI hybrids, the archetypes illustrate the range of roles that AI-enabled systems can play in those interworking scenarios.