Generative AI Models as Wicked Resources: Integrating Property Rights Theory and Stakeholder Dynamics to Govern Resource Ambiguity and Unpredictability The modern economy increasingly depends on advanced digital resources, such as generative artificial intelligence (AI) models, that create value through complex technological processes in novel and sophisticated ways. We examine these resources through the lens of 'resource wickedness,' characterized by attribution ambiguity (i.e., extent to which value creation processes can be delineated into distinct contributory components) and emergent unpredictability (i.e., degree to which resource capabilities and impacts can be anticipated within existing frameworks). By integrating property rights theory (PRT) with stakeholder resource-based theory using system dynamics modeling, we explore how these characteristics reshape governance challenges. Using generative AI models as an exemplar, our analysis reveals two interconnected feedback loops: one centered on resource creation and value appropriation, and another focused-on stakeholder bargaining and institutional adaptation. This theoretical synthesis extends PRT by revealing how wickedness characteristics affect both the institutional evolution of property rights regimes and the underlying processes of value creation and capture that these rights govern. The resulting framework provides both theoretical tools for analyzing increasingly complex resources and practical guidance for organizations navigating these governance challenges.
