The microfoundations of human-AI knowledge co-creation – an emergent knowledge perspective

Purpose This paper aims to explore the interaction mechanisms between human tacit knowledge and artificial intelligence (AI) codified knowledge in digitally augmented decision-making environments, with a focus on the co-creation of emergent knowledge and the shifting dynamics. It investigates how human-AI knowledge partnerships produce novel epistemic outcomes and challenge traditional notions of authority, control and intellectual property. Design/methodology/approach This is a conceptual paper that develops a multilevel framework on Human-AI knowledge co-creation framework encompassing: (1) human tacit knowledge, (2) AI codified knowledge, (3) emergent cocreated knowledge using socio technical systems and systems thinking theory. Findings The study proposes that human-AI collaborations operate as co-evolutionary epistemic relationships rather than linear tool-user interactions. These recursive partnerships produce emergent knowledge artifacts that transcend the capacities of either humans or AI alone. Such co-creation blurs traditional boundaries of expertise, disrupts established intellectual property regimes and raises risks of epistemic alienation, where humans lose interpretive control over knowledge outputs they are expected to own or validate. Practical implications Organizations need to redesign knowledge management systems, performance metrics and governance architectures to accommodate distributed expertise and co-created knowledge. Managers are encouraged to implement dynamic authority distribution mechanisms and cultivate epistemic transparency in human-AI systems. Social implications They underscore the need to reevaluate how knowledge, skill and accountability are defined and distributed in increasingly automated environments. Originality/value This paper contributes to conceptualization of human-AI knowledge creation as a recursive, emergent process and proposes new frameworks for understanding expertise ownership in hybrid intelligence systems.