Artificial Intelligence Along the Supply Chain
Abstract
Problem definition: The rapid advancement of artificial intelligence (AI) is revolutionizing production and operational models, yet whether and how AI technologies diffuse across economically linked firms, especially partners along supply chains, remains underexplored. Methodology/results: This paper documents the diffusion of AI technologies along supply chains where principal customers’ AI adoption significantly predicts dependent suppliers’ subsequent adoption. We establish the causal effect of this diffusion through rigorous and comprehensive empirical strategies that address endogeneity concerns, including placebo tests, selection tests, and an identification strategy based on the random H-1B visa lottery outcome. The diffusion mechanism is primarily driven by learning rather than catering, with suppliers acquiring specific AI applications from customers. The learning effect is stronger in relationships with higher relationship-specific investments and more frequent interactions. Suppliers experiencing AI diffusion achieve substantial economic benefits through quality improvement and cost reduction. Managerial implications: For corporate managers, our findings demonstrate that supply chain relationships serve as critical conduits for building AI capabilities, particularly for resource-constrained smaller suppliers, highlighting the strategic value of cultivating partnerships with technologically advanced firms. For policymakers, our results reveal supply chains as natural mechanisms for democratizing AI access, suggesting that policies supporting supply chain integration and knowledge sharing could accelerate the broader diffusion of AI benefits across the economy and society.

