Seeing Like a Supply Chain: Understanding Data in Logistics

• In Event: Critical Data Studies: Ethics and Human Contexts of Data Science I

Abstract

Supply chains, as Deborah Cowen and others have shown, are critical to understanding the global economy. As offshoring has become increasingly ubiquitous, supply chains have become head-spinningly complex. Vendors and subcontractors nest inside each other like Russian dolls, and manufacturers claim that they can't keep their own supply chains straight.

Nevertheless, just-in-time production and rapid cycles of obsolescence require that commodities be handed through the supply chain with minimal latency. To manage this complexity and mitigate risk, corporations have turned to complex supply-chain management software, which identifies and "heals" rifts in the network of suppliers.

This presentation investigates the growing prevalence of machine-learning techniques - algorithms that "learn" and evolve from massive pools of data - in supply-chain management. As the global market roils and the market permits ever-smaller margins, supply-chain management software incorporates increasingly sophisticated machine-learning algorithms, turning to neural-networking models and multi-agent systems, among others, in an effort to eliminate supply-chain latency. These algorithms rely on very particular facts even as they ignore others, resulting in the curious circumstance that Apple can deliver an iPhone to a customer with exquisite speed while also claiming (accurately) near-total ignorance about the labor conditions of its subcontractors.

Author

• Miriam Posner, University of California, Los Angeles